Journal of Applied Ec,ology 1998, 35, 821828 Habitat use by shorttoed eagles Circaetus gallicus and their reptilian prey during the breeding season in Dadia Forest (northeastern Greece) D.E. BAKALOUDIS", C.G. VLACHOST and G.J. HOLLOWAY* *School of Animal and Microbial Sciences, University of Reading, Whiteknights PO Box 228, Reading RG6 6AJ, UK; and?department of Forestry and Natural Environment, Aristotelian University of Thessaloniki, PO Box 241, Thessaloniki 54006. Greece Summary 1. The area surrounding and including Dadia Forest, northeastern Greece, is well known for its diversity of breeding raptors, including many species of conservation concern. Dadia Forest has been exploited by humans for many centuries, but more recent social and economic changes have stimulated proposals that the forest should be subject to habitat management to protect the fauna of the region. 2. We examined the distribution of reptiles over nine different habitat types with a view to assessing the importance of these habitats for foraging by shorttoed eagles Circaetus gallicus. In addition, data on the diet of the species were collected from direct observations at nests. 3. The shorttoed eagle relies heavily on snakes for food. The most important prey species was the grass snake Nutrix natrix, although the Montpellier snake Malpolon monspessulunus and large whip snake Coluber jugularis also featured prominently as prey items at certain nests. Montpellier snakes and large whip snakes were distributed across all habitat types, but grass snakes were concentrated in areas of mainly intensive, but also nonintensive, cultivation. 4. Analysis showed that shorttoed eagles concentrated their foraging efforts in three habitat types: intensive and nonintensive cultivation and grasslands. Grass snakes were abundant on cultivated land but relatively scarce on grassland. Forested areas were largely avoided by foraging eagles. The data show that for the shorttoed eagle the distribution and abundance of prey items on the ground does not reflect food availability. 5. The possible effect of changes in habitat management on the shorttoed eagle population in Dadia is discussed, in particular the establishment of exclusion zones that could result in progressive canopy closure. Keywords: birds of prey, foraging, forest management, wildlife management. Ecology (1 998) 35, 821828 0 1998 British Ecological Society Introduction Dadia Forest borders the Evros River in Evros Province, northeastern Greece. It is particularly important for wildlife because it offers the only extensively wooded region along the Evros Valley. The position of Dadia Forest, lying just north of the Evros Delta and close to the eastern end of the Mediterranean, makes it an important flyway for birds moving north Correspondence: G.J. Holloway (fax: 01189310180; e mail: g.j.holloway(4reading.ac.uk). and south during the spring and autumn migrations, respectively. Ornithologically, however, Dadia is perhaps best known for the remarkable diversity of breeding birds of prey (Hallman 1979), including a number of vulnerable and rare species, such as the black vulture Aegypius monachus L., shorttoed eagle Circaetus gallicus (Gmelin), lesser spotted eagle Aquifa pomarina Brehm. and imperial eagle Aguila heliaca Savigny (Tucker & Heath 1994). The area is also important for some isolated populations of species more widespread elsewhere, such as the griffon vulture Gyps fuitlvus (Hablizl). In addition to birds of prey, other thre 821
822 Habitat use by the shorttoed eagle (01998 British Ecology. 35, 821828 atened species, such as the black stork Ciconia nigra (L.), breed in Dadia Forest. As a result, Dadia has become a focus for protection and conservation by international wildlife organizations. One of the most important considerations for any conservation programme for the study area is that humans have exploited Dadia Forest for generations. Traditionally, Dadia has been used extensively by goat and cattle herders for livestock grazing. Ungulate grazing damages vegetation, resulting in the opening up of the forest canopy and the maintenance of open shrublands and grasslands. In oak woodland where livestock has recently been allowed to graze, regeneration of trees through browsing is restricted, causing widespread forest degradation. More recently, timber exploitation has become important, causing areas of forest to be thinned on a rotational basis. The result is that Dadia Forest is a patchwork of habitats consisting of closed pine and oak woodlands, degraded oak forest, open shrubland and rocky areas. The latest, and perhaps most devastating development as far as wildlife is concerned, has been the restraint of the Evros River through banking. Formally, the river flooded extensively annually, so that the flood plain was very marshy and difficult to farm for long periods of the year. With the strengthening and raising of the banks of the river, the flood plains have become available for exploitation and are now farmed intensively. Efforts are being made to introduce a programme to manage and protect the wildlife of the area, but in order to develop such a programme data are required. One of the most important questions concerns the likely impact of human activities (beneficial and harmful) on the plants and animals of the region. This information could be used to determine whether some (more) areas should be designated as exclusion zones, from which certain human activities could be excluded. Due to the Mediterranean climate, reptiles are very common in the region (Helmer & Scholte 1985; Strijbosch, Helmer & Scholte 1989). Reptiles form a vital link in the food chain as predators of a variety of small animals ranging from insects to amphibians and small mammals, and as food for higher order predators such as shrikes, falcons and the two principal reptileeating eagle species of the region, the shorttoed eagle and lesser spotted eagle (excluding the golden eagle Aquila chrysuetos L. which preys extensively on tortoises). These two eagle species have considerable conservation value. The shorttoed eagle has an estimated global population of 590014000 breeding pairs, whilst the world breeding population of the lesser spotted eagle lies between 6700 and 9500 pairs (Tucker & Heath 1994). Taking the lower estimates, the sizes of the breeding populations in Greece are between 5% and 8.5% of the world estimate for the shorttoed eagle, but only about I % for the lesser spotted eagle which, generally, has a more northerly distribution in Europe (Cramp & Simmons 1980; Tucker & Heath 1994). Despite its unfavourable conservation status, hardly anything is known about the feeding and breeding ecology of the shorttoed eagle. Food supplies are a major factor limiting the breeding numbers of many raptors (Newton 1979). In order to take measures to conserve shorttoed eagles in Dadia, information on their preferred prey, foraging behaviour and the extent to which human activity affects their food supplies is required. The implementation of a management plan for the area may result in changes in the way the land is utilized for agriculture and forestry. It is important to know how reptiles respond to land use activities to predict how any changes might affect the abundance and diversity of reptiles, which would have a knockon effect on many groups of animals, including shorttoed eagles. Despite their obvious significance in the food chain in Dadia, little is known about reptile abundance, distribution and diversity. The purpose of the current study was to gather information to establish how reptile populations respond to the different habitat types and land uses of the area. In addition, data on the preferred food of shorttoed eagles were collected from observations at the nest. Finally, to link these two data sets together, a study was carried out to determine the most important habitats for actively foraging eagles. Materials and methods STUDY AREA AND HABITAT CLASSIFICATION The study area is located in the central part of Evros Province, northeastern Greece (400 59'410 15'N, 260 19'260 36'E), and covers approximately 37 200 ha (7250 ha are protected). Situated on the southern edge of the Rhodopi mountain chain, its altitude ranges from 20 to 700m above sea level. The whole area is crisscrossed with steep valleys. The climate is submediterranean with a strong continental character: dry summers and cold winters. The minimum mean monthly temperature is 1.8 "C (January) and the maximum mean monthly temperature is 30.0 "C (July). Northerly winds predominate and the annual precipitation is about 740 mm. The area is characterized by a mosaic of habitats, but dominated by woodlands. The main tree species are: Pinus brutiu, Pinus nigra, Quercus conferta, Q. pubescens, Q. sessilijlora and Q. cerris. Other common plant species present include Phyllirea media, Arbutus andruchne, Juniperus oxycedrus and Erica urborea (see below). For the authorities to the European flora see Halliday & Beadle (1983). For the purpose of the present investigation nine different habitat types were identified. 1. Intensively cultivated areas. These are situated by the Evros river and its tributaries, generally <40m above sea level with a flat terrain. The areas cultivated are generally large with very little vertical structure
823 D.E. Bakaloudis, C.G. Vlachos & G.J. Holloway 0 1998 British Ecology, 35, 821828 (e.g. hedgerows) separating the fields. The crops are grown on an annual cycle and are dominated by cotton and sugarbeet, which are irrigated during the dry months of summer. Irrigation involves redirecting water from the Evros River into dammed reservoirs and then across the cultivated areas in channels. 2. Nonintensively cultivated areas. These are situated by the small villages and generally 40100 m above sea level. These areas are hilly and used to grow cereals, sunflowers and grapes, with many hedges and coppice strips. The fields are smaller than on intensively cultivated land, sometimes only measuring about 30 x 30m. In addition, fields are often left fallow, sometimes for considerable periods of time, so that rough patches of grass occur between some of the fields. 3. Shrublands. These are mainly situated in the southwestern part of the study area where the grazing pressure is relatively low. The maquis vegetation is dominated by Phyllirea media, Arbutus andrachne, Palliurus spinac~r~sti, Carpinus orientalis, Ustrya carp~n~olia, Cistus spp., Erica arborea and Erica manipuliflora, and forms > 40% ground cover. 4. Pine forest (canopy >40% closed). This is situated mainly in the central part of the study area, approximately 100300 m above sea level and is characterized by mature associations of Pinus brutia and Pinus nigra. 5. Mixed pineoak forest (canopy >40% closed). This is principally found in the northwestern part of the study area. approximately 300700 m above sea level and is dominated by Pinus brutia and Quercus spp. associations. 6. Oak forest (canopy >40% closed). This is mainly found in the northwestern part of the study area, approximately 500700 m above sea level and is dominated by Q. conferta, Q. pubescens, Q. sessiliflora, Sorbus torminalis, Fraxinus ornus and Colutea arborescens. 7. Degraded oak forest (canopy <40% closed). This is mainly found in the western part of the study area, approximately 400600 m above sea level and is dominated by old oak trees in the upper layer and Cistus spp. in the herb layer. This habitat is heavily grazed by cattle, goats and sheep. 8. Grasslands with sparse shrubs (shrub coverage 140%). These areas are generally between 0.05 and 1.5 ha with a patchy, but sparse, distribution of shrubs (e.g. Palliurus spinachristi, Juniperus oxycedrus and Curpinus spp.). Grasslands are used extensively for grazing livestock, which probably helps to maintain an open structure. 9. Rocky areas. These are mainly found on the steep southfacing slopes in the southern part of the study area and on mountain tops where the soil has been eroded away. Essentially bare rock covers up to 40% of this habitat. REPTILES Data were collected between the beginning of April and the end of September in 1996. For each habitat type (see above), five randomly selected plots were established (45 plots in all). Each plot measured 10 x 100 m (0.1 ha) and was at least 200 m away from the boundaries of any other plots to avoid edge effects. The shape of each plot facilitated sampling. Two people were able to walk the length of each plot approximately 5 m apart, enabling each reptile to be spotted easily and scored. Identification was carried out following Arnold, Burton & Ovenden (1978); for authorities to the European reptiles see Halliday & Adler (1986). All 45 plots were sampled once each week during the study period. Sampling was carried out between 08:OO and 19:00h, but the order in which the plots were visited was randomized to avoid time of day effects. To normalize variances, the data were first log transformed for statistical analysis (Bartlett 1947; Zar 1996). Analysis of variance and chisquared (x ) tests were carried out using Minitab (version 10.5); Fratios and X values are presented in the text. DIET Information on the diet of shorttoed eagles was collected by direct observations at nests. In 1996 and 1997, three observation posts were erected close to separate shorttoed eagle nests so that the prey items brought to the nest could be scored. The hide in which the observer was secreted was moved gradually closer to the nest over a period of some weeks prior to egg hatch, so as not to alarm the parents unduly and possibly cause desertion. All of the hides in the end were approximately 20 m from the nest. Observations were made with the use of a 30 x telescope trained on the nest to aid prey identification. Prey identification was carried out on the basis of features described in Arnold, Burton & Ovenden (1978). FORAGING SITES Throughout 1996 and 1997, data were collected on the habitat type over which eagles were seen hunting. A predetermined route was followed on a weekly basis during the breeding season. From high vantage points along the route, hunting eagles were noted and the habitat type over which they were hunting was scored. As sightings were made from points well above the surrounding landscape, potential bias from being able to see hunting eagles more easily over open rather than closed habitats was effectively eliminated. Eagles engaged in hunting are easy to recognize as they frequently hover and search the ground with lowered head and, of course, drop to the ground periodically during attempts to capture prey. The observation points were sufficiently far apart to be reasonably confident that hunting birds were not noted twice. Hunting eagles tend to fly slowly inspecting the ground for prey. If an eagle was seen twice, it would have moved a long way in a short period of time and
824 Habitat use by the shorttoed eagle would almost certainly have ceased hunting. The data were analysed following the method of Randall Byers, Steinhorst & Krausman (1984), which enables a statistical test of which habitats are preferred (or avoided) for hunting based on the number of observations and the total areas of the various habitat types. The area of each habitat type was measured from largescale maps of Dadia and surrounding regions. Results DISTRIBUTION OF REPTILES A total of 1703 sightings of reptiles was made during the study period, represented by eight lizard and 10 snake species. Table1 shows the mean numbers of sightings of each species in each habitat type. There was a 10fold variation in the numbers of recordings made in each habitat (F8.,026 = 34.9, P < 0.001), with rocky areas (habitat 9) being least favoured and pineoak forest (habitat 5) having the highest number of sightings. All of the forest containing oak trees had relatively high reptile counts (habitats 5, 6 and 7), but the numbers were dominated by green lizard Lacerta uiridis and snakeeyed skink Ablepharus kitaibelii. Interestingly, intensively cultivated areas (habitat 1) also had a high count (and low species number, 9) but in this case the numbers were dominated by two snake species, grass snake Nutrix natrix and dice snake N. tessellata, which accounted for nearly 80% of the records. Only one habitat, rocky areas, was associated with fewer species than intensive cultivation, and here only seven species were recorded. Between 11 and 16 species of reptiles were recorded in the other habitats, the highest number of species being recorded in oak forest. By Far the most common species of reptile recorded was green lizard. Generally, between 30% and 50% of the sightings in each habitat were of this species, the exception being intensively cultivated areas where it constituted only 6% of records. The only other reptile species to be recorded in all habitat types was the Balkan green lizard L. trilineata. Among the snakes, two species, grass and dice snakes, shared 65% of the observations, but nearly all of these came from just one habitat, intensively cultivated areas. Three snake species were noted in all habitat types: Montpellier snake Malpolon monspessulanus, large whip snake Coluberjugularis and nosehorned viper Vipera ammodytes. Table 1. Mean numbers of lizard and snake species visit' hectare' of nine habitat types (1 = intensively cultivated areas, 2 = nonintensively cultivated areas, 3 = shrublands, 4 = pine forest, 5 = mixed oakpine forest, 6 = oak forest, 7 = degraded oak forest, 8 = grasslands, 9 = rocky areas) in the region of Dadia Forest in northeastern Greece in 1996. The species involved include green lizard Lacerta oiridis, Balkan green lizard L. trilineata, snakeeyed lizard Ophisops elegans, snakeeyed skink Ablepharus kitaibelii, Erhard's wall lizard Podarcis erhardii, Balkan wall lizard P. taurica, common wall lizard P. muralis, European glass lizard Ophisaurus apodus, grass snake Natrix nutrix, dice snake N. tessellata, Montpellier snake Malpolon monspessulanus, large whip snake Coluber jugularis, Dahl's whip snake C. najadum, Aesculapian snake EIaphe longissima, fourlined snake E. quatuorlineata, smooth snake Coronella austriaca, cat snake Telescopus,fallax and nosehorned viper Vipera ummodytes Habitat Species I 2 3 4 5 6 7 8 9 Grand mean Lizards L. uiridis 7.5 16.2 25 30 52.1 55.4 36.7 9.6 9.6 26.9 L. trilineata 9.2 7.1 4.2 2.9 1.7 2.5 7.9 10 1.2 5.2 0. elegans 0 7.5 12.9 2.5 22.5 2.1 0 2.5 1.7 5.7 A. kitaibelii 0 0 2.1 9.2 43.3 35.8 31.2 5 0.4 14.1 P. erhardii 0 2.1 1.7 2.5 23 8 2.5 0 0 4.4 P. raurica 0 4.2 2.9 0 6.7 0.4 4.6 0 0 2.1 P. muralis 0 0 0 0 2.1 2.1 1.2 0 0 0.6 0. apodus 1.2 3.7 2.1 1.2 2.1 1.2 2.1 1.7 0 1.7 Snakes N. natrix 47.9 5.4 1.7 0 2.1 1.7 1.2 0.8 0 6.8 N. tesselluta 51.7 0 0 0.8 1.2 1.2 0 0 0 6.1 M. monspessulanus 2.5 1.2 2.9 1.2 1.2 2.5 3.3 2.1 0 1.9 C. jugularis 4.6 1.2 3.3 1.7 1.7 0.8 2.1 2.5 1.2 2.1 C. nujadum 0 0 0.8 0.4 0.8 1.2 0.8 0.8 0.4 0.6 E. longissima 1.2 0.8 0.4 0.8 0.4 0.4 0 1.7 0 0.6 E. quatuorlineata 0 0.4 0 0 0 0 0 0 0 0.04 C. austriaca 0 0 0 0 0 0.4 0.4 0 0 0.09 01998 British T. fallax 0 0.4 0 0 0 0 0 0 0 0.04 v. ammodytes 1.7 0.4 1.7 0.4 1.7 2.5 1.7 3.3 0.8 I.6 Habitat means 7.1 2.8 3.4 3.0 9.0 6.6 5.5 2.2 0.8 Ecologj. 35, Species number 9 13 13 12 15 16 13 11 7 821828
825 EAGLE DIET FORAGING D.E. Bakaloudis, C.G. Vlachos & G.J. Hollowuy Table 2 shows the numbers of different types of prey brought to each of three nests. The data show that snakes dominated the diet of the shorttoed eagle and comprised nearly 80% of the prey items taken. Of the snakes taken to the nests, grass snakes were the most numerous, constituting over 55% of the snakes taken. If the data are split into Natrix spp. and nonnatrix spp., there was no significant difference among the nests (x2 = 1.3, 2 d.f., NS). However, it is clear that there was some variation among nests in the non Nutrix spp. taken. For example, the adults associated with nest 1 captured a high proportion of Montpellier snakes, while the adults from nest 2 caught a higher proportion of large whip snakes than the other eagles, although the latter observation was based on a small sample. It is possible that amongnest variation in the snake species caught depends on precisely where the adults concentrate their hunting efforts, since Table 1 shows that the reptile community present was heavily dependent on habitat type. Lizards were less frequently taken than snakes and only two species, European glass lizard Ophisuurus upodus and green lizard, were captured. Very few animals other than reptiles were observed being brought to the nest. Table 3 shows the habitats over which shorttoed eagles were seen foraging in 1996 and 1997 and the percentage of time that eagles were noted foraging over the various habitats. In addition, a corrected percentage is provided in Table 3 that takes into account the total area of each habitat in the study area. It can be seen from Table 3 that shorttoed eagles spent a lot of their time foraging over open habitats (habitats 1, 2 and 8). Of particular importance were the cultivated areas (both intensive and nonintensive) and grasslands. Degraded oak forest also appeared to be quite important, followed by shrublands and rocky areas. An analysis of the utilization of each habitat (observed frequency) in relation to the percentage of land covered by each habitat (which determines expected frequency) was carried out following the procedure described by Randall Byers, Steinhorst & Krausman (1984). Since there was no difference between years in the distribution of sightings across the various habitat types (1 = 9.76, 7 d.f., NS), the data were combined for analysis. Overall, there was a highly significant difference between the expected and observed level of usage of each habitat type (x = 162.2, 8 d.f., P < 0.001). Having made the appropriate Bonferroni corrections (Randall Byers, Table 2. The numbers (n) and percentages of different prey types brought to each of three shorttoed eagle Circaetus gallicus nests Nest 1 2 3 Species n % n % n Y 0 1998 British Ecolog.v, 35, 821828 Snakes Natrix natrix N. tessellaru Malpolon monspessu~anu~ Elaphe longissirna Coluber jugularis C. najadum Unknown Lizards Ophisaurus upodus Lacerta viridis Tortoises Testudo sp. Mammals Apodemus syltraticus Amphibians Bufo bufo Grand total 19 39.5 11 57.9 13 1 5.3 3 10 20.8 1 5.3 3 2 4.2 2 2 4.2 4 21.0 2 I 2.1 1 2.1 35 12.9 17 89.5 23 4 8.3 1 7 14.6 2 10.5 2 11 22.9 2 10.5 3 1 2.1 I 2.1 48 100 19 100 21 48.2 11.1 11.1 7.4 7.4 85.2 3.7 7.4 11.1 3.7 100
826 Hubitat use by the shorttoed eagle Table3. Numbers of observations (n) and percentages of year totals of shorttoed eagles Circaetus gallicus in 1996 and 1997 hunting over different habitat types: 1 = intensively cultivated areas; 2 = nonintensively cultivated areas; 3 = shrublands; 4 = pine forest; 5 = mixed oak pine forest; 6 = oak forest; 7 = degraded oak forest; 8 = grasslands; 9 = rocky areas (see text for habitat descriptions). An estimate of the number of observations of hunting eagles per hectare of each habitat is also given (% ha ') Habitat Year I 2 3 4 5 6 7 8 9 1996 n 10 53 5 2 5 0 6 26 2 109 Yo 9.2 48.7 4.6 1.8 4.6 0 5.5 23.8 1.8 100 % ha' 19.3 23.3 11.0 0.5 2.2 0 17.4 19.4 6.9 100 1997 n 24 40 5 1 5 0 6 34 5 120 % 20.0 33.3 4.2 0.8 4.2 0 5.0 28.3 4.2 100 O/o ha I 33.2 13.6 7.9 0.2 1.6 0 12.5 18.2 12.8 100 Q 1998 British of Applied.. Ecologr, 35. 82 1828 Steinhorst & Krausman 19841, intensive cultivation, nonintensive cultivation and grasslands were utilized more than expected, while pine, mixed and oak forests were all utilized less than expected (P < 0.05 at least in all cases). The degree of utilization of the remaining three habitat types (shrublands, degraded oak forest and rocky areas) did not deviate significantly from expected. Discussion Table 2 shows that the most important prey species for shorttoed eagles in the study area were grass snakes. Some other snake species, such as Montpellier snake and large whip snake, formed an important component of the diet of the young eagle in one or more of the nests. The next most important category of prey included lizards, but only two species were taken: green lizard and European glass lizard. This confirms findings from other studies that snakes rather than lizards are the main prey of shorttoed eagles (Thiollay 1968; Laszlo 1971; Cramp & Simmons 1980; Petretti 1988; Ivanovsky 1992). The majority of the prey species taken grow to a similar size, around 1 m in length. This includes both of the Natrix species, Dahl's whip snake C. najadum and the European glass lizard. The leopard snake Eluphe situlu is rare around Dadia, but pellet analysis (D.E. Bakaloudis, unpublished data) has shown that this species is occasionally taken. Leopard snakes also grow to around 1 m in length. Montpellier snakes can grow to 2 m in length, but individuals of this size were not seen as prey items. Madon (1933) also noted that most of the prey taken in Europe are about 1 m long, while in Italy most snakes taken measure from 60cm to l00cm (Petretti 1988). In the Indian subcontinent, the snakes taken are considerably larger, 1.514 m long (Ah & Ripley 1968). The size of the snake taken may be related to the optimization of energy and time budgets (Begon, Harper & Townsend 1996). Smaller prey items take the same amount of time to capture and transport as larger prey but offer fewer calories. Snakes much larger than 1 m in length may be more difficult or hazardous to carry. This may also account for why lizards are not taken more frequently as they tend to be considerably smaller than snakes. The hornnosed viper V. ummodytes is the most venomous snake in the region and this species was not recorded as a prey item (Table 2). However, shorttoed eagles do not appear to avoid tackling venomous snakes, as the Montpellier snake is also venomous, although rearfanged (Arnold, Burton & Ovenden 1978). Vipera ammodytes is probably largely excluded from the diet of the shorttoed eagles in Dadia as a result of its small size (usually under 65cm), along with certain other species recorded in Table 1, such as Dahl's whip snake. Most of the grass snakes were recorded in habitat 1, intensive cultivation, which is the most open habitat of all those studied (Table I), at least prior to crop growth. Habitat 1 is crisscrossed with irrigation channels that are frequently used, especially later in the season when Dadia Forest and its environs become very dry. Because of the regular irrigation, there are many amphibians present. With their reliance on amphibians as food, it is perhaps not surprising that so many grass snakes should be found in habitat 1 (Table 1). However, a large number of dice snakes were also found in habitat 1 (Table l), but dice snakes only appeared infrequently as an item of prey for shorttoed eagles (Table2). Dice snakes spend a lot more time in and even under water than grass snakes (Arnold, Burton & Ovenden 1978), which may reduce the susceptibility of dice snakes to capture by hunting shorttoed eagles. Table 3 summarizes the data on the habitats used by shorttoed eagles for hunting in Dadia Forest. Open habitats were favoured, as noted by other authors (Thiollay 1968; Choussy 1973; Petretti 1988; Bocca 1989). Indeed, many species of diurnal raptors rely on open areas for hunting (e.g. Wakeley 1978; Newton,
827 D.E. Bukuloudis, C.C. Vluchos & G.J. Hollowuy 0 1998 British Ecology, 35, 821828 Davis & Moss 1996). According to Table 1, habitat 1 (intensive cultivation) contained many grass snakes, the favoured prey species of shorttoed eagles in Dadia Forest, so it is perhaps not surprising that shorttoed eagles hunted over intensive cultivation. The other two habitats significantly favoured for hunting were nonintensive cultivation (habitat 2) and grasslands (habitat 8). Nonintensive cultivation was associated with a relatively high grass snake count, but grassland was not. Also, neither of these habitats contained particularly high counts of any other snake species. It is clear, therefore, that the distribution of potential food items on the ground cannot be used as a good guide to food availability for shorttoed eagles in Dadia Forest. The Dadia Forest region is a patchwork of different habitats (see Study area and habitat classification). Large areas remain wooded and, although the wood is actively logged, the sizes of these forested areas do not appear to be declining. Wooded areas, particularly consisting of Calabrian pine Pinus brutia and black pine P. nigra on south facing slopes, are used by shorttoed eagles for nesting (D. E. Bakaloudis, C. G. Vlachos and G. J. Holloway, unpublished data). Where the trees have been removed, a variety of open habitats exist, including cultivated areas, grasslands and shrublands. These sections are maintained open by human activities, either for crop growing or through the grazing of animals. The present study has shown that some of these open areas, especially nonintensive cultivation and grasslands, are vitally important for foraging shorttoed eagles. In fact, open grasslands in particular are generally very important for foraging in a range of raptor species (e.g. red kite Milvus milvus L., Newton, Davis & Moss 1996; ferruginous hawk Buteo regalis, Wakeley 1978; and golden eagles, Tjernberg 1985). Therefore, shorttoed eagles, and possibly also other raptor species, are benefiting from the effect that people have on the landscape of Dadia Forest. Because of the remarkable diversity of birds of prey in Dadia, efforts have been made to protect some areas to encourage the growth of raptor populations. For example, two large exclusion zones were established in 1980 from which human activities, such as logging and agriculture, have been halted or restricted. The consequence of these protection measures has been gradual canopy closure resulting in lower numbers of raptors breeding in these exclusion zones currently than when the protected areas were first established. However, the population sizes of some species, such as black vulture, have increased in the protected areas although, since a feeding station was also established in Dadia in 1987 to increase winter survival, the precise cause of this increase is not clear (C.G. Vlachos, D.E. Bakaloudis and G.J. Holloway, unpublished data). It has been suggested that further exclusion zones should be established for the benefit of the birds, although such proposals are often tabled without supporting data. From the results of the cur rent study, it is likely that the restriction of traditional human activities in the forest would not help species such as shorttoed eagles, and that the establishment of more exclusion zones might be premature. It would appear that the shorttoed eagle population, in Dadia at least, is not adversely affected by humans. Consequently, it is likely that the retention of activities that maintain the Dadia region as a patchwork of open and wooded habitats is the best way to maintain a healthy population of shorttoed eagles, and possibly certain other raptor species as well. The fact that shorttoed eagles frequently hunt over intensively cultivated land might be a cause for concern if the birds are being exposed to pesticides. However, this has yet to be studied. Acknowledgements We are most grateful to E. Bakaloudis, P. Bakaloudis, S. Tasiopoulos, P. Goudiakas and P. Darisaplis for assistance, and to the Ministry of Agriculture of Greece, General Secretariat of Forests and Natural Environment, for permission to carry out this study. We are also grateful to two anonymous referees for making suggestions for improvements to the text. References Ali, S. & Ripley, S.D. (1968) Hunrlbook of the Birds of India and Paklwzn, Vol. 1. Oxford University Press, London. UK. Arnold, E.N., Burton, J.A. & Ovenden, D.W. (1978) A Field Guide to ihe Reptiles and Amphibians qfbritain and Europe. Collins, London. UK. Bartlett, M.S. (1947) The use of transformations. Biometrics, 3, 3952. Begon, M., Harper, J.L. & Townsend, C.R. (1996) Ecoiogy. Individuals, Populations umd Communities, 3rd edn. Blackwell Scientific Publications, Oxford, UK. Bocca, M. (1989) Status del Biancone (Circaetus gallicus), dell Aquila reale (Aquila chrysac,ros)e del Pellegrino (Falco peregrinus) in Valle d Aosta. Bolletina Museo Regionale de Scien:e Naturali di Torino, I, 163183. Choussy, D. (1973) Observations sur le Circaete Jeanle Blanc. Nos Oisraux, 32, 8289. Cramp, S. & Simmons, K.E.L. (1980) Handbook of the Birds of Europe, the Middle East and North Africa. II. Hawks to Bustards. Oxford University Press, London, UK. Halliday, T. & Adler, K. (1986) The Encyclopaedia of Reptiles and Amphihians. Allen and Unwin, London, UK. Halliday, G. & Beadle, M. (1983) Consolidated Index to European Flora. Cambridge University Press, Cambridge, UK. Hallman, B.C.G. (1979) Guidelines for the Conserration of Birds of Prey in Ewos: A Report of IUCN/WWF Project 1684. IUCN, Gland. Helmer, W. & Scholte, P. (1985) Herpetological Research in Ewos, Greece: Proposal for a Biogenetic Resenw. S.E. H. Consertution ComniitteeCouncil of Europe Study on Critical Bioropes for Threatened Amphibians and Reptiles. Report by Research Institute for Nature Management, Arnhem and Department of Animal Ecology, Catholic University, Toenooiveld, Nijmegen. Ivanovsky, V.V. (1992) Ecology of shorttoed eagle nesting in Byelorussian Poozerje. Modern Ornithology (ed. E. N.
828 Habitat use by the shorttoed eagle Kurochkin) pp. 6977. Nauka Publishers, Moscow, Russia. Laszlo, B. (1971) Data on the food of the Shorttoed Eagle (Circaetus gallicus). Allatani Kozlemenyek, 58, 166. Madon, P. (1933) Les Rapaces d Europe. Toulon. Newton, I. (1979) Population Ecology of Raptors. T. & A. D. Poyser, London. Newton, I., Davis, P.E. & Moss, D. (1996) Distribution and breeding of red kites Milvus mihus in relation to afforestation and other land use in Wales. Journal of Applied Ecology, 33,21 G224. Petretti, F. (1988) Notes on the behaviour and ecology of the shorttoed eagle in Italy. Le Gerfuut, 78, 261286. Randall Byers, C., Steinhorst, R.K. & Krausman, P.R. (1984) Clarification of a technique for analysis of utilizationavailability data. Journul qf Wildlife Management, 48, 10.501053. Strijbosch, H., Helmer, W. & Scholte, P. (1989) Distribution and ecology of lizards in the Greek Province of Evros. AmphibiuReptilia, 10, 151174. Thiollay, J.M. (1968) Essai sur les rapaces du Midi de la France: distribution, ecologie. Circaete JeanleBlanc (Circaetus gallicus). Alauda, 336, 179189. Tjernberg, M. (1985) Spacing of golden eagle Aquila chrysaetos nests in relation to nest site and food availability. Ibis, 127,250255. Tucker, G.M. & Heath, M.F. (1994) Birds in Europe: Their Conservation Status. Bird Life International, Cambridge, UK. Wakeley, J.S. (1978) Factors affecting the use of hunting sites by ferruginous hawks. Condor, 80, 3 16326. Zar, J.H. (1996) Biosfutistical Analysis, 3rd edn. Prentice Hall, London, UK. Received 16 April 1998: revision receiiled 20 August 1998 G 1998 British Ec olog,v, 35, 821828