Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

The feeding habits of wolves in relation to large prey availability in northern Italy

The feeding habits of wolves in relation to large prey availability in northern Italy Meriggi, A.. Brangi, A.. Matteucci, C. and Sacchi, 0. 1996. The feeding habits of wolves in relation to large prey availability in northern Italy. - Ecography 1 9 287-295. We investigated wolf feeding habits in relation to the abundance of wild and domestic ungulates to test the hypothesis that large prey are preferred and that their abundance affects the use of other food categories and diet breadth. We determined diet composition by scat analysis from December 1987 to December 1992. The research was carried out in three study areas located in northern Italy and characterised by marked differences in wild and domestic ungulate abundance. In study area A (low wild and domestic ungulate availability) fruits. livestock. other vertebrttes and wild ungulates made up the bulk of the diet (71% in volume). In area B (high availability of livestock) wolf diet was mtinly based on sheep and wild boars (80% in volume). In study area C (high availability of wild ungulates) wild ungulates were the main food of wolves ( W h in volume). Significant differences were found among study areas in the mean percentage volume of all food categories and in particular for wild ungulates, livestock, other vertebrates and fruits (p < 0.0oOI in all cases). Diet breadth decreased in areas with high availability of large wild and domestic herbivores. The use of livestock species was lower where there was high abundance, richness and diversity of the wild ungulate guild. Selection for wild ungulate species was partially affected by their abundance: however other factors as prey social behaviour. adaptability to the habitat (for introduced species), and body size could have an important role in species selection by wolves. In particular in area C wild boars were selected for, roe and red d m avoided, and fallow deen and mouflons used as available. Livestock species were used in relation to their abundance and accessibility. in particular sheep were selected for and cattle avoided; but if calves born in the pastures were considered as the only available cattle, they were selected for and sheep were used as available. Large and in particular wild herbivores were found to be of great importance for the wolf population maintenance in northern Italy, one of the most important recovery areas of Mediterranean wolves. A . Meriggi (correspondence). A . Brangi and 0. Sacchi, Diparrimento di Biologia AnimaIe, Pia::a Borta, 9. 1-27100 Pavia, Italy. - C. Matreucci, Coop, S.T.E.R.N.A., Via Pedriali 12, 1-47100 Forli, Italy. Over the last decade wolf Canis lupus L. populations have shown an evident recovery in most of their European range even though the species should still be considered as endangered because of population fragmentation and low genetic variability (Bjarvall 1983, Voskir 1983, Pulliainen 1985, Okarma 1989, Blanco et a]. 1990, Genov 1991, Meriggi et al. 1991, Reig et al. Accepted 16 October 1995 Copyright Q ECOGRAPHY 1996 ISSN 0906-7590 Printed in Ireland - all rights reserved ECOGRAPHY 1 9 3 (1996) 1991, Boitani 1992, Okarma 1993, Randi 1993). The expansion of the wolf population seems to be linked to protection by law or reduced wolf hunting seasons and to increased wild ungulate populations. This is particularly true in their Mediterranean range where in recent years wolves have shown marked changes in feeding ecology, most of all in recolonized areas of northern Fig. 1. Location of the study areas. Italy (Matteucci 1991. Mattioli et al. in press, Meriggi et al. in press). In these areas wolves seem to prey intensively on large wild herbivores and they exert a high pressure on free-grazing and unguarded livestock on mountain pastures (Ragni et at. 1985, Meriggi et at. 19911, wherever availability of wild ungulates is low. This evidence is in sharp contrast with previous findings (Macdonald et al. 1980. Boitani 1982). Consequently, wolf feeding behaviour in its southern ranges seems to be very similar to that of northern ones (North America and North Europe), where wolves usually select few large prey species in relation to their suitability (Huggard 1993. Jedrzejewska et al. 1994). A number of studies carried out in the Meditkrcanean range have shown important differences among study areas and periods in wolf feeding ecology, which are probably due to the overall availability of suitable prey (Meriggi et at. 1991. Boitani 1992. Patalano and Lovari 1993, Ragni et al. in press). However, only simultaneous studies on the feeding ecology of wolves in different areas could point out the influence of large prey availability (wild and domestic ungulates) on the use of different food sources. For this purpose. study areas should show marked diflerences in abundance, diversity and richness of both wild ungulates and livestock. This paper describes the variations in wolf diet and resource exploitation in areas with marked differences in large prey availability. The study was carried out with the following aims: 1) to assess the importance 288 of wild ungulate species as a prey for wolves; 2) to investigate whether livestock is an alternative food used by wolves when wild ungulates are Scarce or absent and whether the presence of large wild prey can reduce predation on domestic prey; 3) to investigate whether the overall wolf diet can be conditioned by (wild and domestic) ungulate presence. Materials and methods Study areas The research was carried out in three different mountain areas located in the western, central and eastern parts of the northern Apennines (N. Italy) (Fig. I). The first area (area A. 500 km’) was between 800 and 1700 m a.s.1.; mixed woods were 24% of the whole area. beech woods 32X. conifer reafforestation 3‘%, pastures 26% and scrub 8%. Cultivated fields were restricted to the immediate areas of the sparse small villages. The wolf was the only species of large carnivore present in the area. In winter 1990. I5 wolves distributed in 3 main packs were censused by foot-prints in the snow. The second study area (area B, 200 km’) was between 500 and 1650 m a.s.1. Woods were 64% of the whole area. mainly with dominant beech and chestnuts. pastures scrub t9’%1.and cultivated fields (mainly hay-fields) c 1%. In this area also. wolves were the only species of large carnivores present. About five wolves were recorded and no evidence of reproduction was ECOGRAPHY 193 llW6l found. The third study area (area C. 130 km') was located on the northern slope of the eastern part of northern Apennines between 550 and 1660 m a.s.1. Forest cover was 81% (58% mixed woods, 26% beech and 16% conifer woods), pastures 14'%1, scrub 3% and cultivated areas only 2%. During the study period c. 15 wolves ( 2 packs and some single individuals) were present in this area. In all study areas the human population has decreased by 80% over the last thirty years. The climate of the three study areas was submediterranean with rainfall concentrated in spring and autumn, mild winters and reduced snowfall. located in each study area was calculated assuming an equal hunting effort since the hunting season duration and the number of hunters were the same in the three of them. All predation events on livestock were recorded during the study period in area A, distinguishing by species and age class. No data on predation were collected in area B and C because of the poor collaboration of livestock owners. Statistical analyses Changes in diet among years, study areas and seasons (winter = December-February; spring = March-May; summer = June-August; autumn = September-November) were analysed by Kruskall-Wallis (multiple comparisons) and Mann-Whitney (painvise comparisons) tests on mean percentage volumes. and by xz test on frequencies of occurrence (Siege1 1956). We also calculated the dietary overlap index C between seasons and study areas (Horn 1966, Goszczynski 1986). Diet breadth in the different study areas was estimated using the B index (Feisinger et al. 1981). Selection by wolves for the different species of wild ungulates was analysed in area C by comparing the expected usage proportion (EUP= proportion of availability) with the observed usage proportion (OUP). by means of the Bonferroni confidence interval method (Manly et al. 1993). The same analysis was carried out in areas A and B for livestock species. No analyses of selection for livestock species were performed in area C because of the low number of scats containing domestic ungulates (low sample sue). Mean percentage volumes of different species in the scats were used as the best available estimate of usage. Diet analysis Wolf diet was studied by scat analysis; scats were collected from December 1987 to May 1992 in study area A. from January to December 1992 in area B, and from December 1990 to October 1992 in area C. In study areas A and B scats were collected on transects repeated each season (30 transects for a total of 181 km in A and 22 for a total of 148 km in B). In area C scats were collected seasonally at known marking sites. Scat analysis was carried out by Kruuk and Parish (1981) method (Meriggi et al. 1991). We did not use the method proposed by Floyd et al. (1978) to assess prey biomass because this method has only been tested for a small number of North-American species and cannot be used for other prey without significant bias (Weaver 1993). Identified prey were grouped into 8 food categories (Appendix 1); frequency of occurrence and the mean percentage volume. including those scats in which the category was not present, were calculated for each category and for large prey species. Large prey availability The abundance of wild ungulates was assessed in area A by tracks in the transects, calculating an abundance index (number of tracks per km) for each species and then ranking the species in order of abundance. In area B the index was not calculated because only one species (wild boar) was present. In area C abundance estimates of the different species were done from observation points and by transect counts on sample areas. Domestic ungulate species were directly censused by surveys in all three study areas in the grazing period and by interviews with livestock owners. The abundance of wild and domestic ungulates, with the exception of wild boar, were not compared among study areas because of the different methods of assessing the availability used. In order to compare the relative abundance of wild boar in the three study areas we used the number of harvested animals per km from 1987 to 1992. The average value of harvested animals per hunting area ECWRAPHY 193 (lY96) Results Diet We collected a total of 519 scats, 292 in study area A, 71 in B and 156 in C. In study area A fruits, livestock, other vertebrates and wild ungulates were the bulk of the diet, amounting to 71.2'!/0 by volume (Table 1) and 112.3% by frequency of occurrence (Fig. 2). The diet breadth was very high (B = 0.79). In study area B wolf diet was mainly based on livestock and wild ungulates (sheep and wild boar). These two pooled categories accounted for 8 Y in Oo volume and 90% in terms of frequency of occurrence. Small mammals were the second main food source (Table 1; Fig. 2). Diet breadth was lower than in area A (B = 0.35). Wild ungulates were the main food of wolves in study area C, where they amounted to 90% in both volume and frequency of occurrence; the other food 289 Table I. Mean percentagc volumes (fSE) of food categories of wolf diet in the three study areas and significance or the ditkrences among areas (Kruskall-Wallistest). Food categories Wild ungulates Study areas (N= 292) B (N = 71) 3035. 15) 28.6(4.99) I .9(1.47) -(-) C (N= 156) ~ ~~ Significance levels p < 0.0001 p < o.Oo01 p<o.Oo01 p < 0.000I Wild boar Roe deer Fallow deer Red deer Mouflon Livestock Sheep Goats Cattle Horses Small mammals Other vertebrates Invertebrates Fruits Other veg. matter Garbage 1 3 . 3 I .89) I I .( 9 1.79) 0.7(0.47) 0.9(0.52) 4-1 4-) 19.1(2.16) 9.8( I .62) I S(0.68) 6.5(1.37) I .3(0.62) 1 I .5( 1.43) I3.q I .79) 7.6( I .23) 24.7(2.00) 2.7(0.75) 5.8( I .25) 4-) 5l.W.53) 3545.39) 10.q3.39) 5.q2.50) 0.a -) 6 4 2 . 4I ) 3.8(2.06) 0.02(0.01) 2.9(1.36) 2.3(1.06) 0:q -) -(-) 90.3( I .95) 49.6(3.76) 14.3(2.74) I8.9I2.95) 2.5(1.24) 4.9(1.68) 2.q1 .I 5 ) 0.7(0.42) 4-) 1.3(0.88) 0.01-) 2.2(0.96) 0.9(0.67) O.S(O.38) 0.02(0.02) I.( 9 1.08) o.q-1 p < o.Oo01 p<O.OoOI p<O.OoOI p < 0.05 p = 0.373 p < o.Oo01 p < o.Oo01 p<o.Oo01 p < o.Oo01 p<O.OoOl p < 0.005 categories were quite unimportant (Table I; Fig. 2). In this area we recorded the lowest diet breadth ( B = 0.15). Because of the lack of significant differences among study years in each area, the whole scat samples were used to compare the wolf diets among study areas. Significant differences were found both in the mean percentage volumes and frequency of occurrence of all food categories and in particular for wild ungulates, livestock, other vertebrates and fruits (p < 0.001 in all cases) (Table I , Fig. 2). The greatest overlap was be- tween areas A and B (C = 0.59) and the lowest between areas A and C (C = 0.27). In study area A. 4 categories showed significant seasonal changes (fruits. livestock, small mammals and invertebrates; p e 0.05 in all cases both for ‘%I vol. and F.O.) In areas B and C only other vegetable matter and wild ungulates differed among Seasons (p < 0.05 both for ‘%I vol. and F.O., and p < 0.05 for F.O.respectively). The greatest homogeneity of diet throghout the year was recorded in area C (C min. = 0.93, C max. = 0.99). FO% Large prey availability On the basis of tracks recorded on the 132 transects covered in 1991 and 1992 in study area A, we calculated an abundance index of 1.34 signs km-‘ for wild boar and 0.11 for fallow deer. Roe deer tracks were never encountered on the transects but some individuals were occasionally observed in the western part of the study area. We therefore considered roe deer as present at lower densities than fallow deer. With regard to livestock species we censused 1700 sheep and goats (3.4 km-*), 3300 head of cattle (6.6 km-2) and 400 horses (0.8 km-I). About 500 calves ( I km-’) were born each year on the pastures from April to September. The grazing period was from April to October for cattle, from March to December for sheep and goats and from May to November for horses. In area B the only wild ungulate species present was wild boar plus some fallow deer individuals that had escaped from an enclosure. The yearly presence of livestock amounted to 5550 head: 3000 sheep and goats ( I S km-’), 2 150 cows (10.7 km-’) and 400 horses (2 km-’). No calves were born on the pastures and sheep and goats were free grazing all year round. ECOGRAPHY 193 (1996) 100- WU LIV SM OV INV FRUOVM GA -AREA A AREA I) 0 AREA Fig. 2. Percentage frequency of occurrence of food categories in the three study areas. Food categories: WU =Wild ungulates; LIV = Livestock; SM = Small mammals; OV = Other vertebrates: INV = Invertebrates; FRU = Fruits; OVM = Other vegetable matter; GA = Garbage. 50i November (64.2% of predation cases) (Fig. 3). Of the 48 preyed calves, whose birth date was known. 77.1'% were between 1 and 10 days old and 22.9% > 10 days old: this difference was significant (Fisher exact test p < 0.01). Use and selection for large prey species In all three study areas the wild boar was the wild ungulate species with the highest mean percentage volumes followed by Fallow deer; roe deer occurred in scats collected in study areas A and C and not in B, while mouflon and red deer were only preyed upon in area C (Table 1). The differences among study areas were significant for all species. Seasonal differences were significant for fallow deers in area B. which were used only in autumn (x2=9.97, p<O.O5), and for wild boars in area C. which were used more in spring and in autumn (x2 = 20.49, p c 0.001). In areas A and B the use of wild ungulate species was in accord with Fig. 3. Monthly distribution of livestock predation cases in the availability, while in area wolves selected wild study area A (pooled data 1987-1992). boars, avoided roe and red deers and used fallow deers and mouflons in proportion to the availability (Table 2). The sheep was the livestock species with the highest while for cattle and horses the grazing season was mean percentage volumes in study areas A and B, from April to October. In study area C five species of wild ungulates were while in area C cows were more used. Goats were the present. The abundance estimates were 850 roe deers second species in order of importance in area B while they did not occur in scats collected in area C (Table (6.5 km-'1, 400 wild boars (3.1 km-'). 380 fallow deers (2.9 km-'), 250 red deers (1.9 km-') and 50 (0.4 1). The differences among study areas were significant km -') mouflons (introduced). Domestic ungulates for all species except for the horse. Seasonally signifiwere 1540 head of cattle (11.8 km-') and 200 sheep cant differences in mean percentage volumes were and goats (1.5 km-?), no horses were reared in the recorded for sheep in areas A (x'= 29.22, p c 0.001) and B (xz = 7.68, p = 0.05) where they were most used area. Calves born on pastures were 680 (5.2 km-'), in area A in autumn and in area B in summer, and for and the grazing period for all species was from April cows in area A (x2=9.85, p<O.O5), which were to October. We recorded the highest number of harvested wild boars in area B (R = 1.1; SD = 0.59; mostly eaten in summer. Among livestock species, in area A wolves selected range = 0.4-2.0); wild boars were less abundant in sheep and goats, avoided cattle and used horses as area A (X =0.4: SD =0.27; range=0.03-0.9) and in available. In study area B the same patterns were area C (R = 0.1; SD = 0.08; range = 0.08-0.2). The found but horses were not used at all (Table 3). In difference between the study areas was significant area C the usage proportions were greater than the (one-way ANOVA F'.,, = 9.23. p c 0.005). and acavailability for sheep and goats and lower for cattle, cording to the multiple range test (Duncan procedure) but the low sample size precludes test for differences.. there were significant differences (p < 0.05) between arConsidering the calves born on the pastures as the eas A and B and between areas B and c. only available cattle in area A, calves were selected for (EUP = 0.192; OUP = 0.34; p < 0.05) and horses avoided (EUP = 0.154; OUP = 0.068; p c 0.05). while sheep and goats were used as available (EUP=0.654; Predation on livestock OUP = 0.592; p > 0.05). During the study period (1987-1992) we recorded a Considering all species (wild and domestic) pooled total of 195 cases of predation on livestock in study together no significant correlations were found bearea A; of these 34.4% were sheep, 63.6% calves and tween availability of ungulates and mean percentage 2.0% foals; no adult cows or horses were preyed upon. volumes in the diet (N = 14; r = 0.373; p = 0.189), but Calves were preyed upon mostly from April to July for livestock species only a borderline significance was (68.5% of predation cases). sheep from July to found (N =9: r=0.650 p=O.O58). E C W R A P H Y 193 (IYY6l Table 2. Comparison between expected (EUP = availability proportions) and observed usage proportions (OUP) ol' wild ungulate species in study area C and significance of the ditferences (Bonferroni interval analysis; N = 147 scats with wild ungulates). Wild ungulate species Roe deer Wild boar Fallow deer Red deer EUP OUP 0. I59 0.546 0.209 0.029 Significance levels p < 0.01 p < 0.01 Mouflon 0.440 0.207 0. I97 0.129 0.026 NS p < 0.01 NS et al. 1984). These explanations for wild boar selection by wolves were supported by the greater usage of this species recorded in spring and autumn in area C and by Diet the extremely low occurrence of piglets in fecal samples The patterns of food resource exploitation by wolves (3.3% of total sample and 11.8% of the scats with wild were very different in the three study areas. The feeding boar). On the other hand, roe deer could be more behaviour of wolves varied from being extremely spedifficult to detect by wolves because it is a solitary and cialized in large wild prey (area c)to extremely generalterritorial species, particularly in forest habitats (Hugist (area A). It seems, however, that where large prey gard 1993). With regard to the other wild ungulate were present these were more heavily exploited. The species present in area C, the greater usage recorded for results of other studies throughout the global range of the wolf have shown that the specialisation in large wild fallow deer and mouflon (used as available) in respect prey is most common. In fact diets based exclusively on to the red deer (avoided) could be due, for fallow deer, a few wild ungulate species have been found in North to the abundance (the third species in order of America. in some areas of Iberia, in eastern Europe and availability with a density comparable to that of the recently in India (Carbyn 1983, Salvador and Abad wild boar) and for mouflon to the small body size and 1987. Reig and Jedrzejewski 1988, Fuller 1989. Jhala to the fact that it is an introduced species in area C, 1993, Smietana and Klimek 1993). The three study where there is no availability of rocky habitats useful areas also differed greatly with regard to seasonal diets. for escaping from predators. Livestock seemed to be intensively preyed upon, in The variations from one Season to another decreased with feeding specialization by wolves; i.e. with the particular in areas A and B. in relation to the overall increase in the occurrence of large wild herbivores in availability and accessibility of prey. Sheep seemed to be selected among livestock species, probably because the diet. of their relatively small body size and, in area B. because they were free-grazing all year round. Nevertheless in study area A sheep were more preyed upon in Selection for large prey species late summer and autumn: at the beginning of the grazing period the wolves preyed almost exclusively on Wolves have been found to prey selectively upon large prey in several studies, but selection patterns are usually cattle and in particular on calves < 10 days old. In late very different from one area to another (Cowan 1947, summer calf availability and accessibility decreased beMech 1970, Kolenosky 1972, Carbyn 1983, Peterson et cause of the decrease in births and because of the al. 1984. Ballard et al. 1987, Jedrzejewski et al. 1992, inclusion after their first two weeks of life of the calves Huggard 1993). Several factors can affect selectivity by in herds, which exert active defence against predators wolves for large prey species, such as the availability (Meriggi et al. 1991). The importance of prey accessibiland profitability of prey types, the degree of habitat ity in wolf prey selection was shown by the changes in overlap between predator and prey, risk of injury, selection analysis results when only the calves were probability of successful capture and the prey encoun- considered as available cattle. As a matter of fact, cattle shifted from avoided species to selected ones, while ter rate (Curio 1976, Endler 1991, Huggard 1993). The high use of wild boar by wolves, recorded in the sheep shifted from selected to used as available and three study areas, could be explained as follow: I) horses from used as available to avoided. Predation on individuals belonging to this species live in large sized livestock has already been shown by several studies on groups easily detectable by a predator; 2) births are wolf feeding habits (Pulliainen 1965, Castroviejo et al. spread throughout the year, though concentrating par- 1975, Fritts and Mech 1981, Gunson 1983, Bjorge and ticularly in spring and in autumn (Meriggi et al. 1988); Gunson 1983. Tompa 1983, Reig et al. 1985, Meriggi et 3) subadults are forced to leave matriarchal groups in al. 1991, Patalano and Lovari 1993), and it seems that coincidence with the new births, resulting of course in domestic ungulates are an alternative prey exploited their easier predation (Heck and Raschke 1980, Mauget when wild prey are less available. The data collected in Discussion ECOGRAPHY 1 9 3 (1996) Table 3. Comparison between expected (EUP = availability proportions) and observed usage proportions (OUP) of livestock species in study areas A and B and significance of the ditferences (Bonferroni interval analysis; N=number of scats with livestock). Study area Livestock species ~~ EUP OUP Significance levels p <00 .1 p < 0.01 NS p <00 .1 p < 00 .1 ( N = 67) Sheep and goats Cattle Horses B (N=40) Sheep and goats Cattle 0.315 06 1 .1 0.074 0.580 0.4I 7 study area C confirm this hypothesis since in this area, in the presence of a rich and abundant wild ungulate guild, wolf diet was mainly based on these prey despite the high availability of livestock. According to optimal foraging and optimal diet theories (Stephen and Krebs 1986, Hughes 1993), wolves should select domestic rather than wild ungulates beciluse of: I ) their highly clumped distribution in a few known areas of pastures (reduction of search time and greater opportunities for prey choice); 2) their low ability to recognize and avoid predators as a result of domestication (greater probability to reach and successfully attack prey); 3) their low escape effectiveness (reduction of probability of failure). However the disturbance by man may make domestic prey less profitable than wild ones. In particular wolves are subjected to the risk of being killed and they have a high probability of being unable to exploit the kill entirely. and domestic large prey species s e to follow the em optimal foraging and optimal diet theories, but some factors such as birth date distribution, population structure (for prey species), adaptability to the habitat (for wild prey) and rearing methods and human disturbance (for domestic prey) can affect wolf feeding behaviour and the selectivity pattern. Acknowledgemenrs - We are grateful to M. Apollonio, L. Boitani, P. Galeotti, S. Lovari, C.Mason and E. Okarma for helpful comments on an early draft of the manuscript. S. Boutin and W. Jedrzejewski provided useful criticisms to the paper. The study was supported by the provinces of Alessandria, Genova, Piacenza e Spezia. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Ecography Wiley

The feeding habits of wolves in relation to large prey availability in northern Italy

Download PDF
9 pages

Loading next page...
 
/lp/wiley/the-feeding-habits-of-wolves-in-relation-to-large-prey-availability-in-DZJWDirqex

References (41)

Publisher
Wiley
Copyright
Copyright © 1996 Wiley Subscription Services, Inc., A Wiley Company
ISSN
0906-7590
eISSN
1600-0587
DOI
10.1111/j.1600-0587.1996.tb01256.x
Publisher site
See Article on Publisher Site

Abstract

Meriggi, A.. Brangi, A.. Matteucci, C. and Sacchi, 0. 1996. The feeding habits of wolves in relation to large prey availability in northern Italy. - Ecography 1 9 287-295. We investigated wolf feeding habits in relation to the abundance of wild and domestic ungulates to test the hypothesis that large prey are preferred and that their abundance affects the use of other food categories and diet breadth. We determined diet composition by scat analysis from December 1987 to December 1992. The research was carried out in three study areas located in northern Italy and characterised by marked differences in wild and domestic ungulate abundance. In study area A (low wild and domestic ungulate availability) fruits. livestock. other vertebrttes and wild ungulates made up the bulk of the diet (71% in volume). In area B (high availability of livestock) wolf diet was mtinly based on sheep and wild boars (80% in volume). In study area C (high availability of wild ungulates) wild ungulates were the main food of wolves ( W h in volume). Significant differences were found among study areas in the mean percentage volume of all food categories and in particular for wild ungulates, livestock, other vertebrates and fruits (p < 0.0oOI in all cases). Diet breadth decreased in areas with high availability of large wild and domestic herbivores. The use of livestock species was lower where there was high abundance, richness and diversity of the wild ungulate guild. Selection for wild ungulate species was partially affected by their abundance: however other factors as prey social behaviour. adaptability to the habitat (for introduced species), and body size could have an important role in species selection by wolves. In particular in area C wild boars were selected for, roe and red d m avoided, and fallow deen and mouflons used as available. Livestock species were used in relation to their abundance and accessibility. in particular sheep were selected for and cattle avoided; but if calves born in the pastures were considered as the only available cattle, they were selected for and sheep were used as available. Large and in particular wild herbivores were found to be of great importance for the wolf population maintenance in northern Italy, one of the most important recovery areas of Mediterranean wolves. A . Meriggi (correspondence). A . Brangi and 0. Sacchi, Diparrimento di Biologia AnimaIe, Pia::a Borta, 9. 1-27100 Pavia, Italy. - C. Matreucci, Coop, S.T.E.R.N.A., Via Pedriali 12, 1-47100 Forli, Italy. Over the last decade wolf Canis lupus L. populations have shown an evident recovery in most of their European range even though the species should still be considered as endangered because of population fragmentation and low genetic variability (Bjarvall 1983, Voskir 1983, Pulliainen 1985, Okarma 1989, Blanco et a]. 1990, Genov 1991, Meriggi et al. 1991, Reig et al. Accepted 16 October 1995 Copyright Q ECOGRAPHY 1996 ISSN 0906-7590 Printed in Ireland - all rights reserved ECOGRAPHY 1 9 3 (1996) 1991, Boitani 1992, Okarma 1993, Randi 1993). The expansion of the wolf population seems to be linked to protection by law or reduced wolf hunting seasons and to increased wild ungulate populations. This is particularly true in their Mediterranean range where in recent years wolves have shown marked changes in feeding ecology, most of all in recolonized areas of northern Fig. 1. Location of the study areas. Italy (Matteucci 1991. Mattioli et al. in press, Meriggi et al. in press). In these areas wolves seem to prey intensively on large wild herbivores and they exert a high pressure on free-grazing and unguarded livestock on mountain pastures (Ragni et at. 1985, Meriggi et at. 19911, wherever availability of wild ungulates is low. This evidence is in sharp contrast with previous findings (Macdonald et al. 1980. Boitani 1982). Consequently, wolf feeding behaviour in its southern ranges seems to be very similar to that of northern ones (North America and North Europe), where wolves usually select few large prey species in relation to their suitability (Huggard 1993. Jedrzejewska et al. 1994). A number of studies carried out in the Meditkrcanean range have shown important differences among study areas and periods in wolf feeding ecology, which are probably due to the overall availability of suitable prey (Meriggi et at. 1991. Boitani 1992. Patalano and Lovari 1993, Ragni et al. in press). However, only simultaneous studies on the feeding ecology of wolves in different areas could point out the influence of large prey availability (wild and domestic ungulates) on the use of different food sources. For this purpose. study areas should show marked diflerences in abundance, diversity and richness of both wild ungulates and livestock. This paper describes the variations in wolf diet and resource exploitation in areas with marked differences in large prey availability. The study was carried out with the following aims: 1) to assess the importance 288 of wild ungulate species as a prey for wolves; 2) to investigate whether livestock is an alternative food used by wolves when wild ungulates are Scarce or absent and whether the presence of large wild prey can reduce predation on domestic prey; 3) to investigate whether the overall wolf diet can be conditioned by (wild and domestic) ungulate presence. Materials and methods Study areas The research was carried out in three different mountain areas located in the western, central and eastern parts of the northern Apennines (N. Italy) (Fig. I). The first area (area A. 500 km’) was between 800 and 1700 m a.s.1.; mixed woods were 24% of the whole area. beech woods 32X. conifer reafforestation 3‘%, pastures 26% and scrub 8%. Cultivated fields were restricted to the immediate areas of the sparse small villages. The wolf was the only species of large carnivore present in the area. In winter 1990. I5 wolves distributed in 3 main packs were censused by foot-prints in the snow. The second study area (area B, 200 km’) was between 500 and 1650 m a.s.1. Woods were 64% of the whole area. mainly with dominant beech and chestnuts. pastures scrub t9’%1.and cultivated fields (mainly hay-fields) c 1%. In this area also. wolves were the only species of large carnivores present. About five wolves were recorded and no evidence of reproduction was ECOGRAPHY 193 llW6l found. The third study area (area C. 130 km') was located on the northern slope of the eastern part of northern Apennines between 550 and 1660 m a.s.1. Forest cover was 81% (58% mixed woods, 26% beech and 16% conifer woods), pastures 14'%1, scrub 3% and cultivated areas only 2%. During the study period c. 15 wolves ( 2 packs and some single individuals) were present in this area. In all study areas the human population has decreased by 80% over the last thirty years. The climate of the three study areas was submediterranean with rainfall concentrated in spring and autumn, mild winters and reduced snowfall. located in each study area was calculated assuming an equal hunting effort since the hunting season duration and the number of hunters were the same in the three of them. All predation events on livestock were recorded during the study period in area A, distinguishing by species and age class. No data on predation were collected in area B and C because of the poor collaboration of livestock owners. Statistical analyses Changes in diet among years, study areas and seasons (winter = December-February; spring = March-May; summer = June-August; autumn = September-November) were analysed by Kruskall-Wallis (multiple comparisons) and Mann-Whitney (painvise comparisons) tests on mean percentage volumes. and by xz test on frequencies of occurrence (Siege1 1956). We also calculated the dietary overlap index C between seasons and study areas (Horn 1966, Goszczynski 1986). Diet breadth in the different study areas was estimated using the B index (Feisinger et al. 1981). Selection by wolves for the different species of wild ungulates was analysed in area C by comparing the expected usage proportion (EUP= proportion of availability) with the observed usage proportion (OUP). by means of the Bonferroni confidence interval method (Manly et al. 1993). The same analysis was carried out in areas A and B for livestock species. No analyses of selection for livestock species were performed in area C because of the low number of scats containing domestic ungulates (low sample sue). Mean percentage volumes of different species in the scats were used as the best available estimate of usage. Diet analysis Wolf diet was studied by scat analysis; scats were collected from December 1987 to May 1992 in study area A. from January to December 1992 in area B, and from December 1990 to October 1992 in area C. In study areas A and B scats were collected on transects repeated each season (30 transects for a total of 181 km in A and 22 for a total of 148 km in B). In area C scats were collected seasonally at known marking sites. Scat analysis was carried out by Kruuk and Parish (1981) method (Meriggi et al. 1991). We did not use the method proposed by Floyd et al. (1978) to assess prey biomass because this method has only been tested for a small number of North-American species and cannot be used for other prey without significant bias (Weaver 1993). Identified prey were grouped into 8 food categories (Appendix 1); frequency of occurrence and the mean percentage volume. including those scats in which the category was not present, were calculated for each category and for large prey species. Large prey availability The abundance of wild ungulates was assessed in area A by tracks in the transects, calculating an abundance index (number of tracks per km) for each species and then ranking the species in order of abundance. In area B the index was not calculated because only one species (wild boar) was present. In area C abundance estimates of the different species were done from observation points and by transect counts on sample areas. Domestic ungulate species were directly censused by surveys in all three study areas in the grazing period and by interviews with livestock owners. The abundance of wild and domestic ungulates, with the exception of wild boar, were not compared among study areas because of the different methods of assessing the availability used. In order to compare the relative abundance of wild boar in the three study areas we used the number of harvested animals per km from 1987 to 1992. The average value of harvested animals per hunting area ECWRAPHY 193 (lY96) Results Diet We collected a total of 519 scats, 292 in study area A, 71 in B and 156 in C. In study area A fruits, livestock, other vertebrates and wild ungulates were the bulk of the diet, amounting to 71.2'!/0 by volume (Table 1) and 112.3% by frequency of occurrence (Fig. 2). The diet breadth was very high (B = 0.79). In study area B wolf diet was mainly based on livestock and wild ungulates (sheep and wild boar). These two pooled categories accounted for 8 Y in Oo volume and 90% in terms of frequency of occurrence. Small mammals were the second main food source (Table 1; Fig. 2). Diet breadth was lower than in area A (B = 0.35). Wild ungulates were the main food of wolves in study area C, where they amounted to 90% in both volume and frequency of occurrence; the other food 289 Table I. Mean percentagc volumes (fSE) of food categories of wolf diet in the three study areas and significance or the ditkrences among areas (Kruskall-Wallistest). Food categories Wild ungulates Study areas (N= 292) B (N = 71) 3035. 15) 28.6(4.99) I .9(1.47) -(-) C (N= 156) ~ ~~ Significance levels p < 0.0001 p < o.Oo01 p<o.Oo01 p < 0.000I Wild boar Roe deer Fallow deer Red deer Mouflon Livestock Sheep Goats Cattle Horses Small mammals Other vertebrates Invertebrates Fruits Other veg. matter Garbage 1 3 . 3 I .89) I I .( 9 1.79) 0.7(0.47) 0.9(0.52) 4-1 4-) 19.1(2.16) 9.8( I .62) I S(0.68) 6.5(1.37) I .3(0.62) 1 I .5( 1.43) I3.q I .79) 7.6( I .23) 24.7(2.00) 2.7(0.75) 5.8( I .25) 4-) 5l.W.53) 3545.39) 10.q3.39) 5.q2.50) 0.a -) 6 4 2 . 4I ) 3.8(2.06) 0.02(0.01) 2.9(1.36) 2.3(1.06) 0:q -) -(-) 90.3( I .95) 49.6(3.76) 14.3(2.74) I8.9I2.95) 2.5(1.24) 4.9(1.68) 2.q1 .I 5 ) 0.7(0.42) 4-) 1.3(0.88) 0.01-) 2.2(0.96) 0.9(0.67) O.S(O.38) 0.02(0.02) I.( 9 1.08) o.q-1 p < o.Oo01 p<O.OoOI p<O.OoOI p < 0.05 p = 0.373 p < o.Oo01 p < o.Oo01 p<o.Oo01 p < o.Oo01 p<O.OoOl p < 0.005 categories were quite unimportant (Table I; Fig. 2). In this area we recorded the lowest diet breadth ( B = 0.15). Because of the lack of significant differences among study years in each area, the whole scat samples were used to compare the wolf diets among study areas. Significant differences were found both in the mean percentage volumes and frequency of occurrence of all food categories and in particular for wild ungulates, livestock, other vertebrates and fruits (p < 0.001 in all cases) (Table I , Fig. 2). The greatest overlap was be- tween areas A and B (C = 0.59) and the lowest between areas A and C (C = 0.27). In study area A. 4 categories showed significant seasonal changes (fruits. livestock, small mammals and invertebrates; p e 0.05 in all cases both for ‘%I vol. and F.O.) In areas B and C only other vegetable matter and wild ungulates differed among Seasons (p < 0.05 both for ‘%I vol. and F.O., and p < 0.05 for F.O.respectively). The greatest homogeneity of diet throghout the year was recorded in area C (C min. = 0.93, C max. = 0.99). FO% Large prey availability On the basis of tracks recorded on the 132 transects covered in 1991 and 1992 in study area A, we calculated an abundance index of 1.34 signs km-‘ for wild boar and 0.11 for fallow deer. Roe deer tracks were never encountered on the transects but some individuals were occasionally observed in the western part of the study area. We therefore considered roe deer as present at lower densities than fallow deer. With regard to livestock species we censused 1700 sheep and goats (3.4 km-*), 3300 head of cattle (6.6 km-2) and 400 horses (0.8 km-I). About 500 calves ( I km-’) were born each year on the pastures from April to September. The grazing period was from April to October for cattle, from March to December for sheep and goats and from May to November for horses. In area B the only wild ungulate species present was wild boar plus some fallow deer individuals that had escaped from an enclosure. The yearly presence of livestock amounted to 5550 head: 3000 sheep and goats ( I S km-’), 2 150 cows (10.7 km-’) and 400 horses (2 km-’). No calves were born on the pastures and sheep and goats were free grazing all year round. ECOGRAPHY 193 (1996) 100- WU LIV SM OV INV FRUOVM GA -AREA A AREA I) 0 AREA Fig. 2. Percentage frequency of occurrence of food categories in the three study areas. Food categories: WU =Wild ungulates; LIV = Livestock; SM = Small mammals; OV = Other vertebrates: INV = Invertebrates; FRU = Fruits; OVM = Other vegetable matter; GA = Garbage. 50i November (64.2% of predation cases) (Fig. 3). Of the 48 preyed calves, whose birth date was known. 77.1'% were between 1 and 10 days old and 22.9% > 10 days old: this difference was significant (Fisher exact test p < 0.01). Use and selection for large prey species In all three study areas the wild boar was the wild ungulate species with the highest mean percentage volumes followed by Fallow deer; roe deer occurred in scats collected in study areas A and C and not in B, while mouflon and red deer were only preyed upon in area C (Table 1). The differences among study areas were significant for all species. Seasonal differences were significant for fallow deers in area B. which were used only in autumn (x2=9.97, p<O.O5), and for wild boars in area C. which were used more in spring and in autumn (x2 = 20.49, p c 0.001). In areas A and B the use of wild ungulate species was in accord with Fig. 3. Monthly distribution of livestock predation cases in the availability, while in area wolves selected wild study area A (pooled data 1987-1992). boars, avoided roe and red deers and used fallow deers and mouflons in proportion to the availability (Table 2). The sheep was the livestock species with the highest while for cattle and horses the grazing season was mean percentage volumes in study areas A and B, from April to October. In study area C five species of wild ungulates were while in area C cows were more used. Goats were the present. The abundance estimates were 850 roe deers second species in order of importance in area B while they did not occur in scats collected in area C (Table (6.5 km-'1, 400 wild boars (3.1 km-'). 380 fallow deers (2.9 km-'), 250 red deers (1.9 km-') and 50 (0.4 1). The differences among study areas were significant km -') mouflons (introduced). Domestic ungulates for all species except for the horse. Seasonally signifiwere 1540 head of cattle (11.8 km-') and 200 sheep cant differences in mean percentage volumes were and goats (1.5 km-?), no horses were reared in the recorded for sheep in areas A (x'= 29.22, p c 0.001) and B (xz = 7.68, p = 0.05) where they were most used area. Calves born on pastures were 680 (5.2 km-'), in area A in autumn and in area B in summer, and for and the grazing period for all species was from April cows in area A (x2=9.85, p<O.O5), which were to October. We recorded the highest number of harvested wild boars in area B (R = 1.1; SD = 0.59; mostly eaten in summer. Among livestock species, in area A wolves selected range = 0.4-2.0); wild boars were less abundant in sheep and goats, avoided cattle and used horses as area A (X =0.4: SD =0.27; range=0.03-0.9) and in available. In study area B the same patterns were area C (R = 0.1; SD = 0.08; range = 0.08-0.2). The found but horses were not used at all (Table 3). In difference between the study areas was significant area C the usage proportions were greater than the (one-way ANOVA F'.,, = 9.23. p c 0.005). and acavailability for sheep and goats and lower for cattle, cording to the multiple range test (Duncan procedure) but the low sample size precludes test for differences.. there were significant differences (p < 0.05) between arConsidering the calves born on the pastures as the eas A and B and between areas B and c. only available cattle in area A, calves were selected for (EUP = 0.192; OUP = 0.34; p < 0.05) and horses avoided (EUP = 0.154; OUP = 0.068; p c 0.05). while sheep and goats were used as available (EUP=0.654; Predation on livestock OUP = 0.592; p > 0.05). During the study period (1987-1992) we recorded a Considering all species (wild and domestic) pooled total of 195 cases of predation on livestock in study together no significant correlations were found bearea A; of these 34.4% were sheep, 63.6% calves and tween availability of ungulates and mean percentage 2.0% foals; no adult cows or horses were preyed upon. volumes in the diet (N = 14; r = 0.373; p = 0.189), but Calves were preyed upon mostly from April to July for livestock species only a borderline significance was (68.5% of predation cases). sheep from July to found (N =9: r=0.650 p=O.O58). E C W R A P H Y 193 (IYY6l Table 2. Comparison between expected (EUP = availability proportions) and observed usage proportions (OUP) ol' wild ungulate species in study area C and significance of the ditferences (Bonferroni interval analysis; N = 147 scats with wild ungulates). Wild ungulate species Roe deer Wild boar Fallow deer Red deer EUP OUP 0. I59 0.546 0.209 0.029 Significance levels p < 0.01 p < 0.01 Mouflon 0.440 0.207 0. I97 0.129 0.026 NS p < 0.01 NS et al. 1984). These explanations for wild boar selection by wolves were supported by the greater usage of this species recorded in spring and autumn in area C and by Diet the extremely low occurrence of piglets in fecal samples The patterns of food resource exploitation by wolves (3.3% of total sample and 11.8% of the scats with wild were very different in the three study areas. The feeding boar). On the other hand, roe deer could be more behaviour of wolves varied from being extremely spedifficult to detect by wolves because it is a solitary and cialized in large wild prey (area c)to extremely generalterritorial species, particularly in forest habitats (Hugist (area A). It seems, however, that where large prey gard 1993). With regard to the other wild ungulate were present these were more heavily exploited. The species present in area C, the greater usage recorded for results of other studies throughout the global range of the wolf have shown that the specialisation in large wild fallow deer and mouflon (used as available) in respect prey is most common. In fact diets based exclusively on to the red deer (avoided) could be due, for fallow deer, a few wild ungulate species have been found in North to the abundance (the third species in order of America. in some areas of Iberia, in eastern Europe and availability with a density comparable to that of the recently in India (Carbyn 1983, Salvador and Abad wild boar) and for mouflon to the small body size and 1987. Reig and Jedrzejewski 1988, Fuller 1989. Jhala to the fact that it is an introduced species in area C, 1993, Smietana and Klimek 1993). The three study where there is no availability of rocky habitats useful areas also differed greatly with regard to seasonal diets. for escaping from predators. Livestock seemed to be intensively preyed upon, in The variations from one Season to another decreased with feeding specialization by wolves; i.e. with the particular in areas A and B. in relation to the overall increase in the occurrence of large wild herbivores in availability and accessibility of prey. Sheep seemed to be selected among livestock species, probably because the diet. of their relatively small body size and, in area B. because they were free-grazing all year round. Nevertheless in study area A sheep were more preyed upon in Selection for large prey species late summer and autumn: at the beginning of the grazing period the wolves preyed almost exclusively on Wolves have been found to prey selectively upon large prey in several studies, but selection patterns are usually cattle and in particular on calves < 10 days old. In late very different from one area to another (Cowan 1947, summer calf availability and accessibility decreased beMech 1970, Kolenosky 1972, Carbyn 1983, Peterson et cause of the decrease in births and because of the al. 1984. Ballard et al. 1987, Jedrzejewski et al. 1992, inclusion after their first two weeks of life of the calves Huggard 1993). Several factors can affect selectivity by in herds, which exert active defence against predators wolves for large prey species, such as the availability (Meriggi et al. 1991). The importance of prey accessibiland profitability of prey types, the degree of habitat ity in wolf prey selection was shown by the changes in overlap between predator and prey, risk of injury, selection analysis results when only the calves were probability of successful capture and the prey encoun- considered as available cattle. As a matter of fact, cattle shifted from avoided species to selected ones, while ter rate (Curio 1976, Endler 1991, Huggard 1993). The high use of wild boar by wolves, recorded in the sheep shifted from selected to used as available and three study areas, could be explained as follow: I) horses from used as available to avoided. Predation on individuals belonging to this species live in large sized livestock has already been shown by several studies on groups easily detectable by a predator; 2) births are wolf feeding habits (Pulliainen 1965, Castroviejo et al. spread throughout the year, though concentrating par- 1975, Fritts and Mech 1981, Gunson 1983, Bjorge and ticularly in spring and in autumn (Meriggi et al. 1988); Gunson 1983. Tompa 1983, Reig et al. 1985, Meriggi et 3) subadults are forced to leave matriarchal groups in al. 1991, Patalano and Lovari 1993), and it seems that coincidence with the new births, resulting of course in domestic ungulates are an alternative prey exploited their easier predation (Heck and Raschke 1980, Mauget when wild prey are less available. The data collected in Discussion ECOGRAPHY 1 9 3 (1996) Table 3. Comparison between expected (EUP = availability proportions) and observed usage proportions (OUP) of livestock species in study areas A and B and significance of the ditferences (Bonferroni interval analysis; N=number of scats with livestock). Study area Livestock species ~~ EUP OUP Significance levels p <00 .1 p < 0.01 NS p <00 .1 p < 00 .1 ( N = 67) Sheep and goats Cattle Horses B (N=40) Sheep and goats Cattle 0.315 06 1 .1 0.074 0.580 0.4I 7 study area C confirm this hypothesis since in this area, in the presence of a rich and abundant wild ungulate guild, wolf diet was mainly based on these prey despite the high availability of livestock. According to optimal foraging and optimal diet theories (Stephen and Krebs 1986, Hughes 1993), wolves should select domestic rather than wild ungulates beciluse of: I ) their highly clumped distribution in a few known areas of pastures (reduction of search time and greater opportunities for prey choice); 2) their low ability to recognize and avoid predators as a result of domestication (greater probability to reach and successfully attack prey); 3) their low escape effectiveness (reduction of probability of failure). However the disturbance by man may make domestic prey less profitable than wild ones. In particular wolves are subjected to the risk of being killed and they have a high probability of being unable to exploit the kill entirely. and domestic large prey species s e to follow the em optimal foraging and optimal diet theories, but some factors such as birth date distribution, population structure (for prey species), adaptability to the habitat (for wild prey) and rearing methods and human disturbance (for domestic prey) can affect wolf feeding behaviour and the selectivity pattern. Acknowledgemenrs - We are grateful to M. Apollonio, L. Boitani, P. Galeotti, S. Lovari, C.Mason and E. Okarma for helpful comments on an early draft of the manuscript. S. Boutin and W. Jedrzejewski provided useful criticisms to the paper. The study was supported by the provinces of Alessandria, Genova, Piacenza e Spezia.

Journal

Ecography – Wiley

Published: Jan 1, 1996

There are no references for this article.