Barking in Roe Deer (Capreolus Capreolus): Seasonal Trends and Possible Functions

Hystrix, (n.s.) 13 (1-2) (2002): 13-18

BARKING IN ROE DEER (CAPREOLUS CAPREOLUS): SEASONAL TRENDS AND POSSIBLE FUNCTIONS

IVA ROSSI”, LORENZA MAURI*, SIMONA LAFICARA” AND MARCO APOLLONIO”

Dipartimento di Etologia, Ecologia ed Evoluzione, Universita degli Studi di Pisa, Via Volta 6, 56126 Pisa, Italy * Dipartimento di Zoologia ed Antropologia Biologica, via Muroni 25, 07100 Sassari, Italy; [email protected]

ABSTRACT - We studied barking behaviour of roe deer, Capreolus capreolus, in an area when wolves actively prey on this species, to determine the possible functions of this signal. Our data showed that males barked more frequently than females and that males barked more often during the territorial period than outside it. Undisturbed deer that barked spontaneously, before the arrival of the observer, were significantly more likely to be male than female, while both males and females showed the same likelihood to bark in response to the presence of a human observer. When barking occurs as a result of disturbance, it may act as a “pursuit-deterrent signal” rather than to warn conspecifics of potential danger. We suggest that counterbarking also acts as a pursuit-deterrent call, since audibility is reinforced by duetting, and signals to the predator that it has been detected by the pair. Key words: barking, roe deer, alarm signal, pursuit - deterrent signal, territoriality

INTRODUCTION ing in muntjacs (Muntiacus reevesi), Tilson and Norton (1981) for alarm duetting in klip- Among artiodactyls, the use of alarm signals springer (Oreotragus oreotragus), LaGory in response to predators is widespread (1987) for snorting in white-tailed deer (Guthrie, 1971). Alarm signals usually in- (Odocoileus virginianus)and Car0 (1994) for volve the use of a white rump patch or tail, snorting in African bovid. some form of a bounding gait and sometimes Vocal communication may be particularly an auditory signal such as snorting, stamping important for species in closed, forested or barking. Such antipredator behaviour may habitats where dense vegetation reduces the be considered a warning signal to con- effectiveness of visual signals. specifics about the presence of a predator, as The roe deer (Capreolus capreolus) is a in Spanish ibex (Capra pyrenaica) that give small cervid that inhabits many different en- alarm calls to alert group members which vironments, predominantly wooded habitats, then flee together (Alados and Escos, 1988). in Europe. Antipredator behaviour may also signal in- Barking is the most typical and intensive formation to the predator and thus act as a call of roe deer. Typically, roe deer bark in “pursuit-deterrent’’ signal, by which the response to some form of disturbance and predator is informed that it has been detected barking is then associated with postures that and that further stalking is unlikely to lead to are characteristic of a state of alarm in cervi- capture. This latter hypothesis was proposed dae: flared target (erect caudal hair), jumps, by Woodland et al. (1980) for tail flicking in stilted gait, and brusque vertical head move- swamphen (Porphyrio porphyrio). Other au- ments (Hewison et al., 1998). Roe deer, thors have proposed a similar function for un- however, can also bark spontaneously dur- gulate alarm signals: Yahner (1980) for barking agonistic interactions or while marking 14 I. Rossi et al. a territory. Thus, barking could play an im- ulation in the Casentinesi Forest has not portant role in the territorial system of this been hunted since World War 11, but is species (Danilkin, 1996; Reby 'et al., preyed upon by the wolf. Roe deer in this 1999a,b). Males seem to bark more than fe- area live are sympatric with red deer males and barking increases during the ter- (Cewus elaphus), fallow deer (Duma duma) ritorial period (Danilkin, 1996; Reby et al., and wild boar (Sus scrofa). Density of roe 1998). While visual communication is inef- deer was estimated to be about 10 deer/100 ficient in closed habitats and scent marking ha in 1998 by drive counts (Arezzo advertises the long-term presence of territo- Province, unpublished) and the population rial individuals, barking signals the immedi- structure was 33.5% adult males, 6.2% ate location of the emitter and plays an im- young males, 49% adult females and 12.3% portant role in territory maintenance (Reby fawns (Laficara S. et al., unpublished). et al., 1999 a and b). Few roe deer studies on barking behaviour were present in litera- Methods ture and our study in an a mountainous habi- From May 1998 to April 1999, we collect- tat is the first in Italy. Every month we col- ed data on foot along eight fixed transects lected data along fixed transects and vantage of an overall length of about 40 Km, and points to investigate in which time period from two fixed vantage points far away barking occurs more frequently. We ob- from the transects, within the study area. served deer barking to determine the differ- Each transect and vantage point were regu- ence between sexes in the barking frequen- larly walked and visited once a month at cy and in the response to disturbance by hu- dawn or dusk. The average time spent by mans throughout the year. We also investi- observers along transects and at vantage gated if solitary individuals were more like- points each month was 30 hours. We always ly to bark than individuals in a group. recorded all direct (visual) observations of roe deer that barked, determining whenever possible the sex, age class and behaviour of METHODS the emitter (the distance to the observer was Study Area lesser than 50m). We considered the follow- We carried out the study in the Casentinesi ing age classes for the two sexes: fawn (0-1 Forest National Park, in the northern Apen- year old), young (1-2 years old), adult (>2 nines, in the Arezzo province, between Tus- years old) for males; fawn (0-1 year old), cany and Romagna, Italy. The area is de- adult (>1 year old) for females. We also not- limited by the Monte Falterona (NW) and ed whether observed animal(s) barked and if the Passo dei Mandrioli (SE). Our study other deer in the vicinity responded by area consisted of 5,000 ha on the Tuscany counterbarking. In addition, we recorded inside of the Casentinesi Forest (43"48'N, direct (auditive) evidence of all other barks 11"49'E). This area ranges in elevation from heard, without seeing roe deer that barked. 400 to 1,520 m, with a mean annual tem- The total number of barks recorded was 96, perature of 8" C and a mean annual snow of which 74 were direct observations of deer cover of 94 days at 1,100 m. The region is barking and 22 were only acoustic contacts. mountainous and only small parts (4%) of the study area are flat. Most of the area is Statistical Analyses wooded (97%) and above 1,000 m eleva- We used chi-square tests to look for variation tion, forests are dominated by Fagus and in barking frequency (a binary variable: Abies and, under 1,000 m elevation, by barkho-bark) between two periods (territor- Quercus, Tiliu and Acer. The roe deer pop- ial: from March to August, and non-territor- Barking in roe deer 15

20 +dusk 18 - +down 16 -

14 -

12 - ln

WY JUN JUL AUG SEP OCT NOV EC JAN FEE M4R A!+?

Figure 1 - Number of barks recorded while walking fixed transects and from fixed vantage points for roe deer in the Casentinesi Forest from May 1998 to April 1999. ial: from September to February), and sexes. more males barked than females (included We also looked for variation in barking fre- both disturbed and undisturbed deer) during quency between the sexes and two individual the territorial period (x2=28.521, df= 1, p< situations (“disturbed” and “undisturbed”). 0.001), but not during the non-territorial pe- All roe deer that were observed barking were riod (Fig. 2). divided into two groups on the basis of their The percentages of barking males in com- behaviour: “disturbed” deer barked in re- parison to all males observed during the sponse to the presence of a human observer transects, reached a peak in April and May (the observer saw the deer, then saw it bark- (Fig. 3). In contrast, the percentages of barking and going away), while the “undisturbed” ing females in comparison to all transect fe- deer barked spontaneously, before the arrival males were essentially constant over the of the observer (the observer herd barking be- year. The analysis of barking roe deer fore to see the deer, then saw it barking and showed that the 9% of all observed roe deer staying where it was). barked and the 53% of barking deer were disturbed Among the percentages of barking roe deer in comparison to all observed RESULTS roe deer, disturbance by humans generated During the study year, the number of a different response between females but not recorded barks increased across the territor- males (x2=6.376, df= 1, p< 0.02). ial period, reaching a peak in April and May, Although we never find cases of several an- and diminished from October through imals barking inside the same group, in March (Fig. 1). The higher values were da- some cases (28%), barking provoked coun- ta collected at dusk (Fig. 1). Although both terbarking in neighbouring animals. Such sexes of roe deer can bark, significantly reciprocal barking sessions included up to 16 I. Rossi et al

eQ) z 20

0 No Territoriality Territoriality

Figure 2 - Number of male and female roe deer that barked during the territorial and no- territorial period in the Casentinesi Forest from May 1998 to April 1999.

Figure 3 - Percentages of barking male and female roe deer in comparison to all observed deer in the Casentinesi Forest from May 1998 to April 1999. Barking in roe deer 17

barking +counterbarking 20

m .-5 15 eE nI 0 10 2

0 MAY JUN JUL AUG SEP OCT NOV DEC JAN FEB MAR APR

Figure 4 - Number of individual barks and counterbarks recorded for roe deer in the Cdsentinesi For- est from May 1998 to April 1999. three deer from distinct locations. The fre- rial function of this call. More undisturbed quency of counterbarking did not vary males barked spontaneously before the ob- across the year, unlike barking which in- server arrived than did undisturbed females creased during the territorial period (Fisher and, in this case, barking was not a response test, p=0.048; Fig. 4). to the presence of the observer. However, males and females were equally likely to DISCUSSION bark in response to the presence of a human We found that the rate of barking increased observer. When barking occurs as a result of across the territorial period and reached a disturbance, it could be an alarm call de- peak in April and May, coincident with in- signed to warn conspecifics of imminent creasing intolerance between males and in- danger. As Tilson and Norton pointed out tensive marking activity (Johansson and (1981) for alarm duetting in klipspringer, Liberg, 1996). This result confirms that barking in roe deer is louder than necessary barking can have also a territorial function to warn immediate group members and con- among roe deer (Danilkin, 1996; Reby et veys information on the location of the al., 1999a,b). Although both sexes of roe barker. These qualities of calls could be deer can bark, significantly more males properties to facilitate communication of the barked than females, during the territorial caller’s alertness to potential predators and period (from March to August) and the act as pursuit-deterrent signals, as Woodland number of observed bucks barking was very et al. proposed (1980) for tail flicks of the high in April and May. Thus, barking may swamphen. In fact, while roe deer that were act as a mode of communication with con- surprised at close quarters fled immediately specifics (Hewison et al., 1998) signalling without barking (Hewison et al., 1998), deer the presence, spatial position and identity of standing at a distance from the observer the emitter. Also the behavioural analysis of barked repeatedly as if to inform a potential observed roe deer barking showed a territo- predator that the deer was aware of its pres- 18 I. Rossi et al.

ence (Reby et al., 1999a). Similarly, coun- Hewison, A.J.M., Vincent, J.P. and Reby, D., terbarking may also act as a pursuit-deter- 1998. Social organisation of European rent call (Reby et al., 1999a). Indeed, the roe deer. In: Andersen R., Duncan P., and frequency of counterbarking did not vary Linnell J.D.C. (eds.), The European roe across our study, unlike barking that reached deer: the biology of success, Scandina- a peak during the territorial period. As vian University Press, Oslo: 189-219. Tilson and Norton proposed (1981) for Johansson, A. and Liberg, O., 1996. Func- alarm duetting in klipspringer, the audibili- tional aspects of marking behaviour by ty of the alarm signal is reinforced by duet- male roe deer (Capreolus capreolus). J. ting, which is a signal to the predator that Mammal., 77: 558-567. the pair has detected it. LaGory, K.E., 1987. The influence of habi- tat and group characteristic on the alarm and flight response of white-tailed deer. ACKNOWLEDGEMENTS Anim. Behav., 35: 20-25. Reby, D., Cargnelutti, B. and Hewison, The State Forest Administration of Pratovecchio A.J.M., 1999a. Context and possible for his collaboration. This work was supported functions of barking in roe deer. Anim. by the Tuscany Region and by the 40% Pro- Behav., 57: 1121-1128. gramme of the Italian Ministry of University and Reby, D., Cargnelutti, B., Joachim, J. and Scientific Research. We also thank Wendy King Aulagnier, S., 1999b. Spectral acoustic for comments on the text and for the revision of structure of barking in roe deer (Capreo- the English language. lus capreolus). Sex-, age- and individual- related variations. C.R. Acad. Sci. Paris, Sciences de la vie, 322: 271-279. REFERENCES Reby, D., Hewison, A.J.M., Carnelutti, B., Angibault, J.M. and Vincent, J.P., 1998. Alados, C.L. and Escos, J., 1988. Alarm Use of vocalizations to estimate popula- calls and flight behaviour in Spanish ibex tion size of roe deer. J. Wildl. Manage., (Capra pyrenaica). Biol. Behav., 13: 11- 62: 1342-1348. 21. Tilson, N.L. and Norton, P.M., 1981. Alarm Caro, T. M., 1994. Ungulate antipredator be- duetting and pursuit deterrence in an haviour. Preliminary and comparative da- African antelope. Amer. Nat., 118: 455- ta from African bovids. Behaviour, 128: 462. 189-228. Woodland, D.J., Jaafar, Z. and Knight, M.L., Danilkin, A., 1996. Behavioural ecology of 1980. The ‘pursuit deterrent’ function of Siberian and European roe deer. Chap- alarm signals. Amer. Nat., 115: 748-753. man and Hall, London. Yahner, R.H., 1980. Barking in a primitive Guthrie, R.D., 1971. A new theory of mam- ungulate Muntiacus reevesi: function malian rump patch evolution. Behaviour, and adaptativeness. Amer. Nat., 116: 38: 132-145. 157-177.