Mixed Roosts in the Montagu's Harrier Circus Pygargus During Courtship

Biologia 65/2: 338—343, 2010 Section Zoology DOI: 10.2478/s11756-010-0021-y

Mixed roosts in the Montagu’s harrier Circus pygargus during courtship

Jaroslaw Wi˛acek

Department of Nature Conservation, Biology Institute, Curie-Sklodowska University, Akademicka 19 Street, 20-033 Lublin, Poland; e-mail: [email protected]

Abstract: Mixed communal roosting of Montagu’s harrier Circus pygargus in the pre-laying period was observed on Calcareous Marshes in Eastern Poland from 1992 to 1995. To my knowledge, this behaviour was described in literature for the first time. The communal roosting in Montagu’s harrier during courtship can help in estimation of mate attraction and finally in mate choice. Harriers from communal roosts start egg laying earlier when compared to the outside roosts. Communal roosting as anti-predator behaviour can help with predator detection and provides benefits to all members of the group. The pair formation process has led to disintegration of communal roosting. Males were more common in the roosting places than females. The time of roosting was correlated with the photoperiod. The weather and predators impact delayed the formation of mixed roosting places. Key words: pre-laying period; mixed roosting; Circus pygargus

Introduction up to two thousand individuals of Montagu’s harrier were observed in India in winter (Rahmani & Man- Communal roosting behaviour is relatively common in akadan 1987). Similarly, in Senegal, in the winter ag- wildlife. Different size groups of animals spend nights gregation, nearly one thousand harriers were counted together, including mammals (Anderson 1998; Lewis (Cormier & Baillon 1991). During the time of migra- 1995; Wilkinson 1995) or birds (Eiserer 1984). During tion, up to 400 Montagu’s harriers were observed near different periods of their life cycle, large aggregations Gibraltar (Bernis 1980). During the breeding season the of birds are observed, sometimes up to a few thou- number of roosting birds is significantly lower. Obser- sand individuals (Campbell & Lack 1985). Ornitho- vations from eastern Poland showed that size of male logical literature shows many arguments for the ben- communal roosts in the post-fledging period fluctuated efits and high adaptive value of communal roosting be- from 4 to 15 harriers (Kitowski 2004). Most of the au- haviour (Beauchamp 1999). One of them is presented as thors have underlined the fact that females and males “thermoregulation benefits” (Yom-Tov et al. 1977; Du roost separately during the breeding season (reviewed Plessis et al. 1994). The presence of nearby compan- in Clarke 1996). Comparatively less is known about the ions in the roosts reduces energetic costs by providing adaptive function of communal roosts in the pre-laying protection from weather or predator attack. Commu- period. In “harrier literature” we have no detailed de- nal roosting increases predator detection and decreases scription of the formation and roosting behaviour dur- individual predation risk (Eiserer 1984). Weatherhead ing the courtship period. Some authors have underlined (1983) found that birds on the edge of a roosting place that communal roosting can be useful in mate choice, create a buffer from predation to birds in the central but our knowledge about this aspect is still incomplete part of the roosting place. Ward & Zahavi (1973) first (Gurr 1968; Watson 1977). proposed that roosts can be an “information centre” The aim of this study was to describe harrier be- for unsuccessful birds in detecting good foraging areas. haviour in mixed roosting places during the pre-laying This hypothesis was applied in describing foraging be- period in the context of mate choice, predation risk and haviour in some raptor species as red kite Milvus mi- the impact of weather conditions. To my knowledge, grans L., 1758 (Hiraldo et al 1993). Other authors un- mixed roosting in Montagu’s harrier Circus pygargus derline that foragers can recruit new companions at the (L., 1758) is described for the first time in ornithologi- roosts to increase their group size (Evans 1982; Richner cal literature. & Hebb 1995). Roosting behaviour is common in all species of Circus genus (Clarke 1996; Simmons 2000). In Mon- Material and methods tagu’s harrier, this behaviour is mainly observed out- From 1992 to 1995 communal roosting behaviour of Mon- side the breeding period. Large roosting groups, with tagu’s Harrier was observed on Calcareous Marshes near

c 2010 Institute of Zoology, Slovak Academy of Sciences Roosts in Circus pygargus 339

harriers spent nights and built their nests. The surrounding areas as meadows, crops and bushes were used by harriers as foraging areas. For a more detailed description of the study area see Wi˛acek (2008, 2009). Field observations were conducted during 4 seasons after the arrival of birds at the breeding places in mid April until the start of egg laying at the beginning of May. Com- munal roosting behaviour was observed from 4.30 p.m. at pre-roosts meadows to the time when birds drop down for the night near sunset and ﬁnished at 8.30 p.m. The observations were performed near communal roosting places from distances of about 250 m, using 10 × 50 binoculars and 20 × 77 telescope. During all seasons, 29 observation sessions were conducted, each lasting for 4 hours. The total observation time was 116 hours. 82 Montagu’s harriers of both sexes were observed. Nearly half of them spent nights at communal roosts (Table 1). Some of them (n = 13) were caught, ringed and individually marked (11 females and 2 males). The birds were caught in special ornithological nets using the eagle owl Bubo bubo (L., 1758) as a decoy (Busse 2000). The marked birds were easy to identify at sunset. The colour markers did not modify the harriers’ behaviour (Kochert at al. 1983). The surface of the roosting place was estimated by the surface area of the circle. The diameter was estimated with a measuring tape as the maximum distance between two birds at roosts with an accuracy to 1 m. The localization of harrier’s pre-roosting places and roosts were marked on the ﬁeld map (scale 1:10 000). Every year, all nests in the study area were found. In the following days, nests were monitored twice a week. The day of the 1st egg laying was recorded. Analyses were made with nonparametric statistics (Mann-Whitney test). All analyses were performed with Statistica 6.1.

Results

General pattern Communal roosting of Montagu’s harrier was observed between 1992 and 1994. In 1995, no communal roosting was observed. Birds spent nights in dispersion or inside nesting territory. The communal roosts observed Fig. 1. Map of the study area. 1 – marsh, 2 – meadows, small in the pre-laying period were formed by birds of both circle – pre-roosting place, big circle – roosting place. sexes. Mixed roosts were observed in all seasons of observations. In following years of observations roosting places were changed but pre-roosting places situated on surrounding meadows were stable (Fig. 1). In 1992, Chelm in Eastern Poland (Fig. 1). The observations were 19 birds were presented on roosting, in 1993 7 birds, conducted within the 376 ha nature reserve “Bagno Sere- bryskie” (51◦10 N, 23◦37 E). The study area is a part of and 14 in 1994 (Table 1). In all years of observations, a the Special Protection Area for birds within the Natura 2000 higher number of males (from 2 to 5 males) in compar- network (Buczek & Buczek 1996; Sidlo et al. 2004). The ison with females was observed (Fig. 2). From the ﬁrst sedge Cladietum marisci is the dominant vegetation type day to the end of the second week after their arrival with Cladium mariscus as the dominant species in which to the breeding place, the number of birds increased.

Table 1. The number of the Montagu’s harriers in the breeding and roosting areas during the study.

Seasons Breeding pairs Birds on communal roosting Birds outside communal roosting

1992 11 19 5 1993 12 7 18 1994 9 14 6 1995 6 0 13

Total 38 40 42 340 J. Wi˛acek

1992 1992 20 20 males 18 roosting 15 females 16 breeding 14 12 10 10

Number 8 Number 5 6 4 2 0 0 1 3 5 7 9 111315171921 1 3 5 7 9 111315171921 Days after arrival of the 1st bird Days after arrival of the 1st bird

1993 1993 9 12 8 males roosting 10 7 females breeding 6 8 5 4 6 Number 3 Number 4 2 1 2 0 0 12345678910 123456789101112131415 Days after arrival of the 1st bird Days after arrival of the 1st bird

1994 1994 16 14 14 males 12 roosting 12 females breeding 10 10 8 8 6

Number 6 Number 4 4 2 2 0 0 1 3 5 7 9 111315171921 1 3 5 7 9 11 13 15 17 19 21 Days after arrival of the 1st bird Days after arrival of the 1st bird

Fig. 2. The numbers of birds in communal roosting (left graphs) and the numbers of birds roosting and starting to breed (right graphs) in 1992–1994.

Table 2. The habitat parameters measured at roosting places.

1992 1992 1932 1994 Parameters Main roost Satellite roost

Vegetation height (mean, cm) 89.2 90.5 95.5 88 Water depth (mean, cm) 3.9 1 3.5 3.1 Area (ha) 1.8 0.5 0.6 1.5 Distance to meadows (m) 210 250 235 270 Day of arrival April 16 April 17 April 19 April 20

After the 14th day the number of birds at the roosting after sunrise (near 4.30 a.m.). First flights to hunting place systematically decreased (Fig. 2). The process of areas started between 5.00–6.00 a.m. pair formation led to disintegration of communal roosting (Fig. 2). In general, after a pair was formed, the Habitat harriers spent the night inside their territory. In 1992, Every year, roosts were situated in the sedge Cladium at the beginning of the third week, communal roost- mariscus, the dominant species in the study area, with ing divided into two groups. These two groups were an admixture the reed Phragmites australis. Annually, separated by a distance of 150 m. In 1992 and 1994, thesizeoftheroostschanged(Table2).Thesizeof communal roosting behaviour lasted 21 days, while in the roosting place depended on the number of birds 1993, only 10 days (Fig. 2). This early disintegration in the roosting group. A larger group of birds needed of the roost was caused by fire on the marshes. After a bigger area. Roosting places were situated far from the fire, birds did not return to communal roosts. They the edge of the meadows (Table 2). Roosting places spent the night at places of daily activity in the breed- in 1992 and 1993 were surrounded by drainage ditches ing area. During every season of observation, different (width of 2–3 m) while roosts in 1994 were placed in sized groups of birds were observed outside communal the centre of marshes near the border of the area with roosts (Table 1). Harriers departed from roosting places ditches. Roosts in Circus pygargus 341

24 100 preroosting 20 80 roosting 16 60

12 40 Minutes Hours 8 20

4 0 123 0 1 3 5 7 9 111315171921 Weeks after arrival of the 1st bird Days after arrival Fig. 5. Time spent by harriers in pre-roosting places (means for 1992–1994, n = 12). Fig. 3. The time of arrival to preroosting and roosting places (all seasons).

storm, low temperature) disturb harrier roosting be- 45 haviour. The presence of an intruder in the proximity 40 of the roosts increased the time of flight before landing 35 at the final roosting place (up to 15 min.). Inclement 30 25 weather decreased the time of choosing the place of 20 communal roosting (from 15 to 35 min.). During field

Minutes 15 studies, these disturbing factors were observed rarely 10 (27.04.1992, 30.04.1992, 23.04.1993), therefore, statisti- 5 0 cal estimation of these parameters was impossible. The 123 harriers from communal roosting places started eggs Weeks after arrival of the 1st bird laying earlier (Mann-Whitney test: Z = –2.93, P = 0.003, n = 22). Fig. 4. Total time needed for birds to drop into vegetation roosts in the following weeks of courtship (means for 1992–1994, n = 12). Discussion

Most authors described communal roosts in Montagu’s Behaviour harriers separately in the breeding season (Cramp & Two hours before sunset, harriers started to occupy pre- Simmons 1980; Clarke 1996; Kitowski 2004). Their de- roosting meadows around marshes. These birds were scriptions report that males tend to remain separated observed in seven pre-roosts places situated near the from females. Exceptionally, males spent nights to- edges of marshes (Fig. 1). These meadows were used by gether with young males, but the presence of both sexes harriers every year. The earliest time of arrival to pre- in the breeding season was unreported (Clarke 1996). roosts places was recorded at 4.38 p.m. The latest time Harriers described here are reported for the first time of landing at pre-roosts was observed at 7.50 p.m. – the in ornithological literature. Some authors suggest that twenty first day after arrival. As the season progressed, harriers’ communal roosting during the breeding season birds delayed their arrival to the pre-roosting (r = 0.92, is connected with mate choice (Watson 1977; Clarke P = 0.001, n = 13) and roosting sites (r = 0.93, P = 1996). Communal roosting in the pre-laying period is 0.0001, n = 13) (Fig. 3). Similarly, the time spent by an integral part of pair-formation in the Montagu’s har- harriers in pre-roosting places decreased (r = –0.62, P rier (Wi˛acek 2006). In pre-roosts meadows and roosts = 0.05, n = 12). The cumulated data for three follow- places, harriers spent a lot of time during courtship. ing weeks after arrival are presented as mean values in Therefore, preliminary estimation of mate quality be- Fig. 5 (n = number of observation sessions). Time of fore final pair formation based on male sky-dancing be- passage from pre-roosting places to final roosting place haviour is possible in communal roosting places (Sim- decreased too (r = –0.58, P = 0.04, n = 12). The cu- mons 1988; Pandolfi & Barocci 1994; Wi˛acek 2004). Ob- mulated data for three following weeks after arrival are servations of male age, physical conditions and aggres- presented as mean values in Fig. 4 (n = number of ob- sive behaviour before roosting can help females in mate servation sessions). The earliest time of arrival to final choice. Therefore, birds roosting communally formed roosts was recorded at 6.15 p.m. – sixth day after ar- pairs earlier and they also started earlier egg laying. rival. The latest time of landing at communal roosts The number of birds in roosting places was strongly de- was observed at 8.15 p.m. – the twentieth day after pended on the number of pairs that started nest build- arrival. ing and egg laying (Fig. 2). Disintegration of communal Birds spent nights in vegetation at a distance of a roosts was connected with the beginning of mating and few meters between each other. The intruders, mainly incubation in harrier’s population. Observations from the red fox Vulpes vulpes (L., 1758) or wild boar Sus later stages of the breeding season in Montagu’s Har- scrofa L., 1758 and roe-deer Capreolus capreolus (L., rier confirmed that thesis (Wi˛acek 2004). 1758) and inclement weather before sunset (rainfall, Opposite to the first three seasons, in 1995, birds 342 J. Wi˛acek spent nights separately. A low number of birds at breed- silent and quick (Clarke 1996). Harriers settle in one ing places was probably the reason that communal place usually a few minutes after sunset. Generally, har- roosting behaviour was not observed that year. Com- rier’s behaviour observed in the study area was similar paratively small numbers of breeding pairs were noted to the description cited by many authors in literature at the study area during that season (Table 1). Addi- (Watson 1977; Cramp & Simmons 1980; Hamerstrom tionally, the day of arrival to the breeding place near 1986). The time of flight before final settling decreased Chelm was the latest (20 April) in comparison with for- in the following days of observations, but the strong im- mer seasons of observations from 1991 to 1994 (Wi˛acek pact of weather conditions was observed. Strong wind, 1998). Therefore, the harriers started roosting immedi- storm, rainfall or low temperatures decreased the time ately inside their nesting territories after arrival with- of flights above the place of roosting. The long pre- out communal roosts forming. roosting behaviour in bad weather conditions increased Every season, communal roosts were situated in the energetic costs of this behaviour. Therefore, birds the sedge Cladium mariscus with an admixture of the landed faster in roosting places when the weather was reed Phragmites communis. Harriers choose communal inclement. This problem was described in Montagu’s roosting in a safe place surrounded by ditches in the harrier by Clarke (1996) and other communally roost- centre of a swamp. In the later stages of the breeding ing species like magpie Pica pica (L., 1758) (Mugaas season, nests were built there. In this study, the non- & King 1981; Reebs 1986; Czechowski et al. 2005) or random choice of roosting sites was discovered. Simi- the common raven Corvus corax L., 1758 (Janicke & larly as in nest site selection, birds choose safe places Chakarov 2007). with high vegetation in wet places, usually situated Fire on the marshes in 1993 caused the disinte- far from the edge of meadows (Liminiana et al. 2006; gration of communal roosting. Stress for birds and the Wi˛acek 2009). Later, these places were used as a breed- habitat destruction that followed caused the complete ing colony by harriers (Wi˛acek 2008). Similar conclu- dispersion of harriers. Long term observation at study sions were shown by Clarke (1996). The same places in areas showed that illegal burning has a strong negative winter are used as communal roosts by hen harriers Cir- impact on habitat quality and conditions for breeding cus cyaneus (L., 1766) (Clarke 1996; Kitowski 2004). In harriers in the next five years (Wi˛acek 2009). The num- the context of harrier species using communal roosting ber of breeding pairs is significantly lower after burnings in different seasons, this behaviour was anti-predator (Wi˛acek 2006a). In the same way, the optimal places for in meaning (Watson 1977; Clarke 1996; Arroyo et al communal roosts are destroyed by fire. 2001). The last conclusion is confirmed by the fact that Predators detected at the time of pre-roosting be- most authors did not observe successful predator at- haviour increased the time of place choice for roost- tacks during the time of roosts despite the presence of ing. Disturbed harriers spend a longer time in the air predators in the roosting neighbourhood. above the roosting place. Roosting behaviour was con- Harrier roosting behaviour depends on the pho- tinued when the predator was far away. Predator im- toperiod. In the following days, harriers spent more pact during the time of roosting and birds’ behaviour time in hunting territories and appeared in the pre- were described in literature. These accidents increased roosts later. In Clarke’s (1996) opinion, harriers forag- the time of choosing the final place of roosts in winter ing closer to the place of roosts during the day per- areas and during migration (Cramp & Simmons 1980; formed foraging activity which took place first in pre- Clarke 1996; Simmons 2000; Arroyo et al. 2004). Similar roosting meadows. Birds from foraging areas situated data were presented in post-fledging period by Kitowski far away arrived later. This coordination was observed (2004). in the study area. The time of arrival from foraging ar- Communal roosting of Montagu’s harrier is a eas to pre-roosts meadows was connected with the time strong protection against ground predators. Many eyes of sunset but decreased together with the increasing of can help in faster detection of predators and roosters birds’ experience during the following days of the pre- are successful with mobbing against intruders. These laying period. Females usually arrived in roosting places anti-predator behaviours were described in literature in first. In the time of courtship, females spent more time different stages of the breeding cycles in harriers (Pan- inside territory than males (Wi˛acek 2008). Their forag- dolfi & Pino D’Astore 1992; Simmons 2000; Arroyo et ing areas are situated in surrounding marshes near daily al. 2001). In the context of anti-predatory behaviour, activity territories, therefore, it took place at roosting “predation avoidance hypothesis” is presented as use- areas first. ful (Weaterhead 1983; Eiserer 1984). The presence of During the breeding season, small groups (3–5 indi- companions at roosting places (similarly with breeding viduals) of harriers were observed on pre-roosts mead- in a semi-colony) increases the possibility of predator ows around the swamp. Outside the breeding season, detection and decreases individual risk predation (Ar- Clarke (1996) described a big pre-roosts concentration royo et al. 2001; Wi˛acek 2008). (125 ind.) of wintering harriers in India. However, dur- Summarizing, communal roosting in Montagu’s ing the breeding season, the number of roosting birds harrier during courtship can help in estimation of mate oscillated near the number of breeding pairs (Wi˛acek attraction, mate choice and pair formation. But the 1998). primary meaning of this behaviour may be the anti- The settling process in harriers’ roosting places is predatory function. Roosts in Circus pygargus 343

Acknowledgements Kochert M.N., Stenhoof K. & Mortish M. 1983. Evaluation of patagial markers for raptors and ravens. Wildl. Soc. Bull. I’m grateful to M. Polak and J. Kloskowski and an anony- 11: 271–281. mous referee for their helpful comments on the manuscript. Lewis S.E. 1995. Roost fidelity of bats – a review. J. Mammal. 76: 481–496. Liminiana R., Soutullo A. & Urios V. 2006. Vegetation height selectionin Montagu’s Harrier Circus pygargus breeding in a References natural habitat. Ardea 94: 280–284. Mugaas J.N. & King J.R. 1981. Annual variation of daily energy Anderson J.R.1998. Sleep, sleeping sites, and sleep-related ac- expenditure by the Black-billed Magpie: a study of thermal tivites; awakening to their significance. Am. J. Primatol. 46: and behavioural energetics. Stud. Avian Biol. 5: 78. 63–75. Pandolfi M. & Barocci A.1994. Description of courtship patterns Arroyo B., Garcia J. & Bretagnolle V. 2004. The Montagu’s Har- in Montagu’s Harrier Circus pygargus. Ethol. Ecol. Evol. 6: rier. BWP Update 6: 41–55. 439–440. Arroyo B., Mougeot F. & Bretagnolle V. 2001. Colonial breeding Pandolfi M. & Pino D’Astore P.R. 1992. Aggressive behaviour and nest defence in Montagu’s Harrier. Behav. Ecol. Socio- in Montagu’s Harrier Circus pygargus during the breeding biol. 50: 109–115. DOI 10.1007/s002650100342 season. Boll. Zool. 59: 57–61. Beauchamp G. 1999. The evolution of communal roosting in Rahmani A.R. & Manakadan R. 1987. A large roost of harriers birds: origin and secondary losses. Behav. Ecol. 10: 675–687. inAndhraPradesh,India.J.BombayNat.Hist.Soc.83: DOI 10.1093/beheco/10.6.675 203–204. Bernis F. 1980. La Migracion de las Aves en el Estrecho de Gibral- Reebs S.G.1986. Sleeping behaviour of black-billed magpies under tar. Vol. 1. Aves planeadoras. Univ. Complutense, Madrid. a wide range of temperatures. The Condor 88: 524–526. Buczek A. & Buczek T. 1996. The Calcareous Marshes near Richner H & Hebb P. 1995. Is the information center hypothesis Chelm Region. IUCN – Poland, Lublin, 40 pp. a flop? Adv. Stud. Behav. 24: 1–45. Busse P. 2000. Bird Station Manual. SE European Bird Migration Simmons R. 1988. Honest advertising sexual selection, courtship Network. University of Gdansk, 264 pp. displays and body conditions of polygynous male harriers. Campbell B. & Lack E. (eds) 1985. A Dictionary of Birds. T & The Auk 105: 303–307. AD Poyser, Calton, UK, 670 pp. Simmons R. 2000. Harriers of the World. Oxford University Press, Clarke R. 1996. Montagu’s Harrier. Arlequin Press, Chelmsford, New York, United States, 384 pp. 208 pp. Sidlo P., Blaszkowska B. & Chylarecki P. 2004. Important Birds Cormier J. & Baillon F. 1991. Concentration de Busards cendres Areas of European Union importance in Poland. RSPB – Circus pygargus dans la region de M’bour (Senegal) durant OTOP, Warsaw, 673 pp. l’hiver 1988–89. Utilisation du milieu et regime alimentaire. Ward P. & Zahavi A. 1973. The importance of certain assem- Alauda 59: 163–168. blages of birds as “information-centres” for food-finding. Ibis Cramp S. & Simmons K.E.L. (eds) 1980. The Birds of Western 115: 517–534. Palearctic, Vol. II. Oxford University Press, Oxford, 695 pp. Watson D. 1977. The Hen Harrier. T & AD Poyser, Berkhamsted, Czechowski P., Jerzak L., Zduniak P. & Bochenski M. 2005. Is UK, 307 pp. the communal roosting behaviour of the Magpie Pica pica Weatherhead P.J. 1983. Two principal strategies in avian com- wind dependent? An example from an isolated population in munal roosts. Am. Nat. 121: 237–243. DOI 10.1086/284053 western Poland. Ardeola 52 (2): 333–339. Wi˛acek J. 1998. Ecology of the pre-laying period in the Mon- Du Plessis M.A., Weathers W.W. & Koenig W.D. 1994. Energetic tagu’s Harrier Circus pygargus on the Calcareous marshes benefits of communal roosting by acorn woodpeckers during near Chelm. Ph. D Thesis. Curie Sklodowska University, the non-breeding season. Condor 96: 631–637. Lublin, Poland, 99 pp. Eiserer L.A.1984. Communal roosting in birds. Bird Behav. 5: Wi˛acek J. 2004. Sky-dancing as an honest criterion of mate 61–80. choice in Montagu’s HarrierCircus pygargus, pp. 845—848. Evans R.M. 1982. Foraging-flock recruitment at a black-billed In: Chancellor R.D. & Meyburg B.U. (eds), Raptors World- gull colony: implications for the information center hypothe- wide, WWBPG, Budapest. sis. The Auk 99: 24–30. Wi˛acek J. 2006. Pair formation in the Montagu’s Harrier. Berkut Gurr L.1968. Communal roosting behaviour of Australasian Har- 15 (1–2): 151–158. rier Circus approximans in New Zeland. Ibis 110: 332–337. Wi˛acek J. 2006a. Aggressive behaviour in Montagu’s Harrier Cir- Hakkarainen H., Ilmonen P. & Koivunen V. 1998. Blood parasites cus pygargus during the courtship period. Biologia 61: 593– and nest defense behaviour of Tengmalm’s Owls. Oecologia 595. DOI 10.2478/s11756–006–0095–8 114: 574–577. DOI 10.1007/s004420050482 Wi˛acek J. 2008. Benefits and costs of semi-colonial breeding in Hamerstrom F. 1986. Harrier, Hawk of the Marshes. Smithsonian the Montagu’s Harrier Circus pygargus. Belg. J. Zool. 138: Institution Press, Washington, 171 pp. 36–40. Hamilton W.D. & Zuk M. 1982. Heritable true fitness and Wi˛acek J. 2009. Nest site selection of Montagu’s Harrier Cir- bright birds: a role of parasites. Science 218: 384–387. DOI cus pygargus breeding in natural habitats in eastern Poland. 10.1126/science.7123238 Ardea 97:117–119. Hiraldo F., Heredia B. & Alonso J.C. 1993. Communal roosting of Wilkinson G.S.1995. Information transfer in bats. Symp. Zool. wintering Red Kites Milvus migrans: social feedings strategies Soc. Lond. 67: 345–360. for the exploitation of food resources. Ethology 93: 117–124. Yom-Tov Y., Imber A. & Otterman J. 1977. The microclimat of Janicke T. & Chakarov N. 2007. Effect of weather conditions on winter roosts of the starling Sturnus vulgaris in Israel. Ibis the communal roosting behaviour of common ravens Corvus 119: 366–368. corax with unlimited food resources. J. Ethol. 25: 71–78. Jerzak L., Kavanagh B. & Tryjanowski P. (eds) 2005. Corvids of Received April 1, 2009 Poland. Bogucki Wyd. Nauk, Poznan, 556 pp. Accepted December 5, 2009 Kitowski I. 2004. Observations on the behaviour of male Mon- tagu’s Harrier Circus pygargus at communal roosts during post-fledging period: a case study, pp. 829—834. In: Chancellor R.D. & Meyburg B-U. (eds), Raptors Worldwide, WWGBP, Budapest.