NESTING SUCCESS OF GREY PARTRIDGES (PERDIX PERDIX) ON AGRICULTURAL LAND IN il NORTH-GENTRAL FRANCE: RELATION To Nssnne 'f'

Transcription

1 ' Socal behavour of hazel grouse n Chna 7 Game ano vvuume o~..e..ce, `..,,..,-. Q.V ISSN l 1 I l manfesté un comportement terrtoral. Leurs domanes vtaux se chevauchaent à 82 % et ls étaent ensemble (à mons de 25 m l'un de l'autre) dans 83 % des localsatons effectuées au cours dela pérode pré-ncubatore (du 2 avrl au 4 ma). ll semble donc que le len qu unt les membres d'un couple sot fort et stable. Les deux mâles non-apparés ont présenté des comportements dflérents en avrl. L'un a occupé un pett terrtore de 2,1 ha (n = 31), alors que l'autre avat un grand domane vtal de 31,3 ha (n = 33), apparemment non-défendu et qu recouvrat en parte pluseurs terrtores d'autres mâles. Au début du mos de ma, les mâles apparés ont gardé les mêmes terrtores qu'en avrl, alors que les mâles non-apparés se sont élognés de 300 à 400 m des domanes vtaux qu 'ls occupaent en avrl. Nos données ne provennent que de quelques gélnottes équpées d'émetteurs. Cependant, elles suggèrent que, dans les montagnes de Changba, les mâles apparés sont terrtoraux et qu'ls font parte de couples apparemment auss étrotement uns qu'en Suède. Ce comportement des gélnottes chnoses est dfférent de celu des gélnottes de la régon extrême orentale russe plus proche de la Chne. Dans cette régon, au prntemps, les gelnottes ont probablement des défcences nutrtonnelles dues aux hvers rgoureux. Cependant, les deux membres de notre couple marqué ont été trouvés ensemble plus souvent que les membres des couples suédos. En Suède, les couples les plus uns étaent ceux qu étaent stués dans les habtats avec les taux de prédaton les plus élevés. La relaton étrote que nous avons observée entre les membres de notre couple pourrat donc être lée auss aux habtats de velles forêts de feullus des montagnes de Changba, relatvement plus ouverts et donc propces à la prédaton. Trad. par E. Taran Fleceved 27 September 1999, accepted 03 Aprl 2000, fnal verson receved 07 September NESTING SUCCESS F GREY PARTRIDGES (PERDIX PERDIX) N AGRICULTURAL LAND IN l NRTH-GENTRAL FRANCE: RELATIN To Nssnne 'f' CVER AND PREDATR ABUNDANCE E. BR (* ), F. REITZ (**), J. CLBERT (*) and F! MAYT (**) (*) Laboratore d'ecologe UMR 7625, Unversté Perre el Mare Cure, BP 237, 7 qua Sant-Bernard, F Pars Cedex 05. ( ) Present address: fce natonal de la chasse et de la faune sauvage, GNEFIA Pette Faune Sédentare de Plane, Sant-Benost, F Aultargs. E mal: e.bro@onc.gouv.tr (**) fce natonal de la chasse et de la faune sauvage, CNEHA Pette Faune Sédentare de Plane,. Sant-Benost, F Auffargs. KEY WRDS: Grey partrdge, Perdx perdx, clutoh fate, predaton, farmng practce, predator abundance, nestng cover, management, agrcultural land, France. ABSTRACT We montored 1,009 radotagged wld female grey partrdges, Perdx perdx, from March *to Septemberon 10 contrastng study areas to dentfy and quantty the causes of nest falure. Werecorded the fate of 407 frst clutches and 141 replacement clutches. Smutaneously, we estmated the abundance of red foxes, Vulpes vulpes, and mustelds, Musteldae, and recorded nestng habtat characterstcs. Success rate vared sgnfcantly across study areas from 31% (n = 16) to 73% (n = 46) for frst clutches (p = 0.013), but not for replacement clutches (range : 0%, n = 6,' 53%, n = 16). Cereals were the best nestng cover for frst clutches (66% success rate, n = 232), whereas replacement clutches n cereals suffered from han/estng n July (29% success rate, n = 45). Nevertheless, the varaton n success rate of frst clutches across study areas was not correlated wth the varaton n the amount of cereals. The man cause of falure of frst clutches (n = 150 clutches faled for dentfed causes) was predaton (70%), followed by farmng practces (22%). Replacement clutches (n = 79 clutches faled for dentfed causes) suffered from predaton (51%) and farmng practces (43%). Ground carnvores (red foxes, mustelds and domestc cats) were responsble for 66% of dentfed ncubatng-female predaton cases (n = 76) and for 36% of dentfed egg-predaton cases (n = 69). The predaton rate on frst clutches was postvely correlatecl wth the abundance of mustelds (p = 0.008) but not of red foxes. Predaton rate was especally hgh (46%) n lnear landscape features (n = 46). The results justfy the local control ofsome musteld speces to allevate clutch losses, together wth approprate feld margn management that could be ncluded as a partrdge conservaton measure n the French scheme Contrat Terrtoral d'explotaton (applcaton of the new CAP drectves).

2 I 200 * Ecologcal correlates or nestng success of gm,,,...-,, I.* INTRDUCTIN * Many brd speces are currently reported as endangered or declnng n Europe (n HAGEMEIJER and BLAIR, 1997). Among them, many nhabt agrcultural land (TUCKER, 1997), lke the grey partrdge, Perdx perdx, whch s a representatve example (e.g. BlFtKAN and JACB, 1988; PTTS and AEBI- SCHER, 1995). PTTS (1997) estmated thatthe number of pars has declned by 80% n Europe snce the As a result, the grey partrdge has been lsted as a speces wth an unfavourable conservaton status n Europe (AEBl- SCHER and KAVANAGH, 1997), and has become an mportant management concern. Wthn Europe, the French grey partrdgepopulaton s of great mportance n terms of both breedng numbers and denstes. Numbers have been estmated at 900,000 ndvduals, rankng France frst by number n western Europe (AEBlSCHER and KAVANAGH, 1997). Denstes may stll average*30 parslkmton large areas (RElTZ, 1997; BR et al., 2000a), and even exceed 60 pars/km2 on some ntensvely managed prvate land (Fédératons départementales des chasseurs, unpublshed data). These denstes are hgher than the values of 5-20 pars commonly reported on study areas n the Unted Kngdom (e.g. PTTS and AEBISCHER, 1995; TAPPER et al., 1996), Germany (e.g. KAI- SER, 1998) and Poland (e.g..panek and KAMlENlAFlZ, 1998). As an mportant game brd n Europe, the grey partrdge has been much studed. Losses occur at dfferent crtcal stages: breedng females, clutches and chcks (e.g. BIRKAN and JACB, 1988). Ecologcal. studes have mostly focused on the last stage of breedng, brood-rearng (e.g. GREEN, 1984; STHER- TN and RBERTSN, 1990; PANEK, 1997), because several demographc analyses have shown that chck survval rate had the greatest effect on populaton fluctuatons (BLANK et al., 1967; PDLER and RGERS, 1975; PTTS, 1980), and determned populaton status (PTTS and AEBISCHER, 1995). Moreover, the shootng bag s postvely correlated wth chck survval rate (PTTS and AEBlSCHER, 1991), so ths factor s of key mportance to shootng nterests. However, a supplementary hypothess has recently arsen. PTTS and AEBISCHER (1995) reported that the declne of partrdge numbers observed n the Sussex study area snce the late was correlated wth a decrease n brood-producton rate, reflectng an ncreased mortalty of breedng females and a hgh predaton rate on clutches. A smlar concern about female sunrval durng breedng had been voced n France n the early 1990s (REITZ, 1990; REITZ and BERGER, 1995). To nvestgate ths aspect, we conducted a large feld study by radotrackng 1,009 wld breedng females n on ten study areas n North-Central France. Ths survey confrmed that females could sufer hgh mortalty rates durng the breedng season (BR et al., 2001). Smulaton modellng supported the hypothess that the declne n adult survval rate could explan the depleton of grey partrdge n some regons of France (BR et al., 2000b). As part of ths study, we nvestgated the determnants of grey partrdge nestng success. ur objectves were to (1) dentfy the causes of nest falure, (2) quantty the relatve mportance of losses to agrcultural farmng or predaton dependng upon nestng cover, and (3) test whether the predaton rate on clutches was correlated wth the abundance of predators. These,results wll provde management gudelnes to ncrease breedng success by mprovng nestng success., ` ` II. MATERIAL AND METHDS li.1. STUDY AREAS We montored grey partrdge nests on ten study areas (rangng from 20 to 150 kmz) n contrastng farmng regons of North-Central France (see Fgure 1 n BR et al., 2000a). These areas were chosen to reflect farmland dversty: land use was tradtonal (mxed arable and lvestock farms) n the study areas A (located n the admnstratve department Nord"), B ( Pas-de-Calas"), C ("Somme ), D ( Sene martme ), and H ( Sarthe ). ther study stes were stuated n open farmland of ntensve cereal producton (E: "Marne", F and G: "Aube", I and J: Loret ). A detaled descrpton of farmland characterstcs s gven n BR et al. (2000a). No partcular habtat management or envronmentally frendly farmng practces were appled n the study areas. Partrdge *densty n sprng ranged from 3 to 28 parslkm2 (BR et al., 2000a,b). Weather condtons were those of a mld temperate clmate. The topography was flat to gently undulatng. II.2. CLUTCH SURVEY We radotracked breedng females to fnd nests. We captured 1,009 wld grey partrdge females from md-march to early Aprl , after breedng pars had formed and settled (see PTTS, 1980; BIRKAN and JACB, 1988). Brds were captured at nght wth a hand-held net and usng* a strong lght. Females were rado-tagged wth 10-g necklace rado-transmtters. (BR et al.,1999), and then montored daly untl September. The rado-transmtters were equpped wth ether a mortalty sensor or an actvty sensor (see BR et al., 1999). The former mechansm ndcated that brds were alve and the latter that brds were movng. ln thsÿlast case, when necessary,.the sgnal was recorded several tmes, durng the day» 'untl t was postve. Therefore, ncubatng females could not be confused wth dead brds. Clutches could be located when ncubaton began because, then, females left ther clutches only once or twce a day to feed (see PTTS, 1980;, BlRKAN and JACB, 1988). To confrm that most clutches were actually found durng ncubaton, we calculated the tme between the date of detecton of a clutch and ts date of hatchng for 197 successful clutches. We found that 86.8% of clutches were dscovered wthn 25 days before hatchng (.e., durng the ncubaton perod, consdered to last 25 days; PTTS, 1980), and about 10% wthn more than 27 days before hatchng (.e., durng the layng perod). The reproductve hstory of each female was known through radotrackng. Thus, frst and replacement clutches had been dentfed. However, a frst clutch lost durng layng or early ncubaton could have gone undetected. lndeed, some clutches reported as frst clutches were thought to have hatched too late (after md-july) to be-actual frst clutches. n these cases, we assumed that clutches thathatched, or were expected to hatch, after md-july were replacement clutches unless they had more than 15 eggs. Conversely, clutches wth less than 11 eggs hatched, or expected to hatch, before md-july were assumed to be replacement clutches (see BIRKAN and JACB, 1988). When clutch data were less precse, we classfed them as undentfed clutch type and dd not nclude them n the analyses (n = 7, Appendx). * We recorded 407 frst clutches and 141 replacement clutches (Appendx). The death of many females wthn the frst week after release (BR et al., 1999)

3 202 Ecologcal correlates of nestng success of grey partrdge E. Bro et al. 203 or before ncubaton (BH et al., 2001), and rado falures (BR et al., 1999) manly explaned the dscrepancy between the number of rado-tagged females and the number of clutches found. Clutches were mostly vsted after females had ceased ncubaton (.e., after hatchng, deserton or clutch destructon) to mnmze predaton and deserton rsks nduced by the observer (see NICHLS et al., 1984). Clutch fate was determned accordng to clutch condton (wthn 24 hr after falure, through the daly montorng of females). Hatchng was recognzed by small regular breaks n the egg-shells, often wth one half of the shell wthn the other one (BlFlKAN and JACB, 1988). The fndng of ntact cold eggs whlst the female was known to be alve,was nterpreted as nest deserton. Clutch destructon was attrbuted to farmng practces when eggs were found broken and compressed nto the nest, and to predaton when large empty fragments of shell were found scattered around the nest. Putatve predators were dentfed by the speces- or group-specfc breaks and punctures n the eggshells caused by the teeth or beaks of predators (BIFIKAN and JACB, 1988). These crtera dd not allow to dentfy a predator speces wth the same accuracy and relablty. For nstance the presence of solated small holes was a postve crteron to dentfy mustelds, Musteldae, whereas such precse crtera were not avalable for red foxes, Vulpes vulpes, and con/ds, Corvdae, whch were probably classfed as undentfled predators". Thus, t s lkely that the nfluence of mustelds was overestmated. V We recorded the habtat characterstcs of nest stes, n partcular the type of nestng cover' and the dstance from nest to feld margn. Nestng cover was classfed nto sx habtat types: cereals (wheat, barley, oats, rye), other crops (maze, sunflowers, olseed rape, sugar beet, peas, potatoes, lnseed), fodder crops (ray-grass, lucerne, clover), meadows (.e., pastures), set-asde (all dfferent types pooled), and lnear landscape features (.e., hedgerow, bank, roadsde, dtch). ll.3. PREDATR ABUNDANCE Red fox abundance (klometrc ndex of abundance) was estmated n late February by drvng slowly by car at nght and countng the number of ndvduals detected along a road transect (wth a length of 0.6 km per km* of study area) by usng a hand-held lght (STAHL, 1990; STAHL and MIGT, 1990). We estmated the* ndex of abundance n late wnter because ths was the most sutable perod n terms of vsblty (STAHL, 1990). Musteld abundance (potentally ncludng the stone marten, Martes fona, the polecat, Putorus putorus, the pne marten, Martes martes, the weasel, Mustela nvals and the stoat, Mustela ermlnea) was estmated n August-September through the presence of faeces along 30 1-km transects covered on foot on each study area. To sample the whole study area homogeneously, we cross-ruled the map nto 30 equal squares. Wthn each square, one 1-km lnear feature was selected as a sample (.e., a hedge, wocdedge, feld boundary, lane). The ndex of abundance used was the proporton of.transects wth faeces (hereafter postve transects"). Data were analysed througha presence-absence varablevrather than usng the number of faeces found along a gven transectbecause ths latter varable does not reflect the abundance of anmals but a specal behavour. Two r, =..- 11: Ham nnart tn test,for the repeatablty of the sa* r E rg; * 1. l mnmum value assumng that a transect was postve only when both surveys were postve. The fnal ndex used n the statstcal analyss was the average value of the maxmum and mnmum. li.4. STATISTICAL ANALYSES Success rate of the clutches was defned as the proporton of successful clutches (.e., clutches for whch at least one egg hatched) n a gven study area. We dstngushed between predaton rate on ncubatng females and predaton rate on clutches because predators of eggs and partrdges supposedly were dfferent ones. The analyss of predaton on ncubatng females s relevant to a study of clutch success because the death of females durng ncubaton leads to falure of the clutch. The predaton rate on clutches was defned as the proporton of clutches that faled because of predaton (.e. number of clutches faled by predaton dvded by the total number of clutches recorded). The defnton of the predaton rate on ncubatng females was smlar. The predaton rates (and more generally the falure rates) estmated wth our data corresponded to mnmum values because falures durng layng or early ncubaton were lkely to go undetected. g ' We excluded from analyses the clutches whose fate was unknown (mssng values: n = 12for frst clutches, n = 0 for replacement clutches), and those that were deserted mmedately after beng vsted by an observer because abandonment could drectly be related to dsturbance (n = 12 for frst clutches, n = 6 for replacement clutches). To take the sample sze of the category undentfed cause of clutch falure nto account, we consdered that the *proporton of clutches that faled because of predaton wthn ths category was the same as the proporton computed wth the cases of dentfed causes of falure (.e., the pro-rata rule"). Ths procedure assumed that non-dentfed causes of falure were not based. Dfferences n sample sze n the.resuits are due to mssng values for other varables such as nestng cover or dstance to the feld margn. Because a relatvely small number of frst and replacement clutches were descrbed per study area and per year (at least for some study areas, Appendx), we focused our analyses on spatal and not on temporal varatons, and pooled all data. Smlarly, for sample-sze reasons, we dd not analyse the data by ncludng both all sx categores of the nestng cover varable and all ten stes of the study area varable, and ther nteracton, but used the followng analyss procedure. We examned the varaton n the causes of clutch falure across study areas frst, to correlate spatal dfferences wth envronmental condtons such as habtat avalablty and predator abundance. Then, we examned the dfferences n the pattern of clutch fate across nestng cover (by poolng data across study areas) to further nvestgate the nfluence of habtat. Frst and replacement clutch data are a pror not ndependent because replacement clutches are lad by females that have already produced a frst clutch whch faled. Comparng the fate of frst and replacement clutches by usng pared analyses was not possble snce all frst clutches were unsuccessful. Thus, there s some pseudoreplcaton n the data used for the tests of comparrson betweennestng attempts. «Dfferences n clutch fate across study areas and across nestng cover types were nvestgated by a log-lnear model (MGCULLAGH and NELDER, 1983) rm.-fr- r:l*:r\l nn lkelhnnd rato ch-square. empty cells were forcbly consd-

4 204 Ecologcal correlates of nestng success of grey partrdge E. Bro et al ) to examne the relatonshp between 1) varablty of the predaton rate across study areas and the abundance of red foxes and mustelds, 2) varablty of success rate and the amount of cereals (proc GENMD, bnomal dstrbuton, logt-lnk functon, controllng for overdsperson, weghted analyses to take account of the number of clutches recorded on each study area). We used a logstc analyss at an ndvdual scale to test whether predaton rsk was hgher for clutches lad near feld boundares. ne-taled tests were performed when testng relatonshps for whch we nvestgated only one devaton of the dfference (.e., whether the predaton rate ncreased wth predator abundance, or whether t was hgher near feld margns) (SKAL and RLHF, 1981). III. RESULTS III.1. SPATIAL VARIATIN IN CLUTCH SUCCESS RATE Clutch success ratedffered between frst and replacement clutches (X2, = 13.15, p < 0.001). Success rate offrst clutches vared sgnfcantly across study areas (X2 = 20.87, pv = 0.013) between 31% (area H, n = 16) and 72% (area D, n = 46) (Appendx). 'Success rate *of replacement clutches ranged from 0% (area F, n = 6) to 53% (area D, n = 16), but dd not vary sgnfcantly across study areas (X2, = 10.46, n.s.). Ths could result from the smaller sample sze per study area. V III.2-. CAUSES' F CLUTCH FAILURE The pattern of falure (.e., the categores death of ncubatng females", clutch destructon and clutch deserton ) dd not vary smultaneously across study areas and clutch type (.e., frst type and replacement type) (nteracton pattern of falure * study area * clutch type": X218 = 22.69, n.s.). but dffered sgnfcantly both across study areas (X215 = 36.44, p = 0.006) and between clutch types (X22 = 9.35, p = 0.009). Faled frst clutches (n = 181) suffered from the death of ncubatng females (40%), clutch destructon (34%) and clutch deserton (26%) n roughly smlar proportons (Fgure 1, Appendx), whereas clutch destructon (51%) was the man type of falure for faled replacement clutches (n = 89), before clutch deserton (27%) and death of the ncubatng females (22%) (Fgure 1). However, n gven study areas, the death of ncubatng females could reach 69% (area F, n = 13 faled frst clutches), the deserton of frst clutches 53% (area A, n = 15 faled clutches) and clutch destructon 75% (area C, n = 16 faled frst clutches) (Appendx). The causes of clutch falure (.e., the categores predaton - on both ncubatng females and eggs -, farmng practces", other causes - such as weather condtons) dffered sgnfcantly between nestng attempts (X2, = 9.14, p = 0.010) and moderately acrossstudy areas (121, = 28.46, p = 0.055), the nteracton "cause of falure * study area * clutch type beng non sgnfcant (fu, = 25.35, n.s.). Frst clutches (n = 150 clutches faled from dentfed causes) experenced hgh predaton rates on both ncubatng females and eggs (70%), whereas replacement clutches (n = 79 clutches faled from dentfed causes) suffered more equal-..; ' :~.-. n:1o/\ and -Farmnr nrflntrtfs f43 /-l ffigljl' S 1 ãl'icii2i. I'IW6V l', J, l ' *.r,. I j*î.. Là -`. '/'~,' 1 l tv ««F m l * L. ÎΚ!'Î~Îî' 5' ',.1- l' t3;2,iášf IS! r.: *-; 2. l * ---«-T,-,-IMF-F Pattern of clutch falure Frst clutches T /o ffã\le «ou -4 (fl = 50 -> 50 Causes of clutch falure =- - É Death of Clutch' Clutch ncubatng females destructon deserton Replacement clutches 100 I * 15 so > so ã Ê %of fa'ure 25 'z" ";;. '25 ' I Ê. 0 0.JF Death of. Clutch - E Clutch ncubatng females destructon asse;-[on 1 Death ofancubang females 1 Pfefllatofl lîl Clutch destructon Farmng practces '., ther causes Cuœh d se'1' " É-Undentfed causes Fgure 1 : Pattern of clutch falure and cause of clutch falure ( other causes" ncludes weather condton) of grey partrdge, Perdx perdx, n relaton to nestng attempt. Causes of clutch falure are detaled for each pattern of clutch falure. Mean 1 standard error across ten study areas n North-Central France, Total n = 407 frst clutches and 141 replacement clutches. Number of clutch falures: 181 frst clutches and 89 replacement clutches. Fgure 1 : Modaltés de perte (par mort de la poule couveuse - en nor -, destructon des œufs - en blanc -, abandon du nd par a poule - en hachuré) et causes d'échec des pontes (prédaton de la poule et/ou des œufs - en nor -, pratques agrcoles - en blanc -, autres causes dont les condtons météorologques - en hachuré -, et causes non dentfées - en hachures horzontales) de perdrx grse, Perdx perdx,, enfoncton.du type de ponte (premères pontes ou pontes de remplacement). Les causes d echecasont détallees par modalte de perte. Moyenne 1 erreur standard pour dx stes d'etude du Centre-Nord de la France, Effectf total n = e: u 4.. nn.. t `

5 206 Ecologcal correlates of nestng success of grey partrdge E. Bro et al. 207 n = 181 1B% Frst clutches 25% n =46 2% AW """ mr, ~w"'w^^^~.à`fw f(\/v` ï 47% w Wwwšš. '~'*"'~.~^~a,. V*"`.-'?v^'Q-"*'\ wwvrvãä Replacement clutches n = 89 Causes of falure 11% IPredatun on munarng females H 5/F* ;:; --; 5 5;; I9% H Predaton on eggs I ï. c Farmng practces ` j;; jïš;; : ;;j `-; ; ljitl'ar causes 25% sav a unaanrrra GGIISGS Egg predators nasa fox = 23 M: ~" ".«-«~*^ 25% IMslelíds ~^^ " š" URIVN uunaenfa.ss % carnvore 4% EC1Vr!s råî ss. *Rk- \S\\ v. ~ ;':I WD EI Undsrtfl ed 22% 7% speces Predators of ncubatng females (n = 76 predaton cases) were red foxes (34%), mustelds (28%), domestc cats (4%), and some raptors (13%) - supposed to be hen harrers, Crcus cyaneus (see BR et al., 2001 ). Undentfed mammalan carnvores accounted for 17% of female losses and undentfed predators for 4%. The man predators of clutches (n = 69 predaton cases) were carnvores: mustelds (33.3%), red fox (1.5%), domestc cats (1.5%) and undentfed carnvores (4.3%). Predaton attrbuted to corvds (carron crow, Corvus corone corone) represented 11.6% and to hedgehogs, Ernaceus europaeus, 7.2%. Predator speces could not be dentfed n 40.6% of predaton cases (Fgure 2). The man cause of clutch falure due to farmng practces was harvestng whereby many clutches were destroyed (Fgure 2). To a far lesser extent, rrgaton caused clutch deserton. The mpact of farmng practces on clutch success depended upon the crops chosen by partrdges to nest n (see below). Most other causes of clutch falure were poor weather condtons (10 out of 17 cases). lndeed, afterheavy rans some clutches were found deserted. We assumed that these clutches had been flooded and were then deserted by the females (Fgure 2). Ths cause of clutch loss globally represented a small proporton of the losses but locally (.e., where storms occurred) mght lead to hgh losses and decreased clutch success. lii.3. CLUTCH FATE IN RELATIN T NESTING HABITAT n = 33 Farmng practces n = 34 D l 16% lwaterng-rrgaton 3% 6% 9% *.. 1s% sn f. : Elsprayng \È 5% 3' sš 3%, vyÿf Unldonlflefmachns :,. ' uμeraton -j-: 12% *VIII 18% Elûuttng set-asde '-Â-I- - ' ' 'äfl» *-:-:-:----.-:-I-:- 34%., * 3% Etcultlng fodder crops '--EIÎÎ-I-'-" El Harveslng cereals Havestlngother crops Ecmls Fgure 2: Dstrbuton (%) of causes of falure of frst and replacement clutches of grey partrdge, Perdx perdíx. Putatve predators of eggs (.e., not ncludng predators of ncubatng females; nformaton about predators of female partrdges s gven n BR et al., 2001) and farmng practces (whereby ncubatng females are klled and/or clutches destroyed) are detaled. Ten study areas n North-Central France, Total n = 407 frst clutches and 141 replacement clutches. Fgure 2 : Dstrbuton (%) des causes d'échec des premères pontes et des pontes de remplacement de perdrx grses, Perdx perdx. Les prédateurs supposés des œufs (ceux des poules perdrx non nclus; vor BR et al., 2001) et les pratques agrcoles (tuant les poules couveuses et/ou détrusant les œufs) sont détallés. Dx stes d'étude dans le Centre-Nord de la France, , effectf total n = 407 premères pontes et 141 :pontes de remplacement. Causes de perte : prédaton des poules couveuses, prédaton des œufs, pratques agrcoles, autres causes, causes non dentfées. Prédateurs des œufs 1 renard, mustéldés, carnvore non dentfé, corvdés, prédateur-non dentfé. Pratques agrcoles: arrosage-rrgaton, tratement /2 46% r The pattern of clutch fate (success v falure due to predaton - on ncubatng females and clutches - and farmng practces) vared smultaneously between clutch type and across nestng cover (nteracton fate * clutch type * nestng cover": X210 > 50, p < , poolng study areas for sample sze reasons; Fgure 3). Clutch fate was strongly dependent upon nestng cover for both frst (fu, = 46.13, p < 0.001) and replacement (fu, = 24.50, p = 0.006) clutches. Success rate of frst clutches (Fgure 3) was hghest n cereals (66%, n = 232) and lowest n fodder crops (23%, n = 13) and meadows (18%, n = 11); t was ntermedate n other crops (50%, n = 13), lnear features (46%, n = 46) and set-asdes (45%, n = 20). Although cereals appeared as the best cover, the varaton n the amount of cereals across study areas was not sgnfcantly correlated wth the varaton n the success of frst clutches (X2, = 0.57, n.s., onetaled weghted analyss). Contrarly to frst clutches, the success rate of replacement clutches (Fgure 3) was hghest n the other crops (52%, n = 27; manly n sugar beet and pea; BH et al., 2000a) and set-asde (45%, n = 11) whereas t was poor n cereals (29%, n = 45) and lnear features (29%, n = 24). All replacement clutches faled n meadows (n = 3) and fodder crops (n = 7). Predaton rate on both ncubatng females and clutches ranged from 23% (n = 13, fodder crops) to 45% (n = 11, meadows) across nestng cover types for frst clutches and from 22% (r = 27, other crops ) to 57% (n = 7, fodder crops) for replacement clutches (Fgure 3). The falure rate of frst clutches due to predaton was hgh (40-45%) n lnear landscape features (n = 46), other crops (n = 10) and meadows (n = 11) and moderate (25%) n cereals (n = 232). ln lnear features (n = 46), predaton occurred on clutches and ncubatng females n equal proporton whereas n cereals (n = 232) females were more senstve to predaton than clutches (Fgure-3). The stuaton was dfferent tor replacement clutches: predaton rate wasthgh (45-60%) n fodder crops (n = *n : I.. -HM nn-ml nn* nerln In - 'I'I\ $thi'i QTIII I'YIf('IF!fâTfl (?4 Â1I In

6 208 Ecologcal correlates of nestng SUCCQSS f 9(9)/ Pflfffdge Lneaffeawåg) Frst clutches ff Sel-*de (20,. Meadflwf (11) -* M) F =1d f f ff ) "ea'f lêå (232) l '. Replacement clutches ' :=;*=.f= -íwl S Î"as(Îfl')* I f* 'F MGEUWS (3) :sf-:i:=:=:=:==š=š=î=elèí*"' ' 'K ššš ""~`:ÎÃ* '.š**? l' =?' %'ê`%éî~:3'*'rã ~3 thffflfïgš rader mpg,e "E " 25. (7) casss Êsqîšë: ššï5 >Im Ill. - - (45).1-0% 20% 40% 60% 80% 100% Clutch fate (%) Q Success liipredaton on ncubatng females li Predaton on eggs El Farmng practces III ther causes Fgure 3: Fate (% success v % falures due to predaton on ncubatng females, predaton on eggs, farmng practce, and other causgs) f ÎITSÎ and feplacemeff ã Ut "Î5 of grey partrdge, Perdx perdx, dependng upon nestng cover. Fgureå nh ÊC Ê; refer to sample sze, poolng study.areas and years. Ten stud/ afeas "`l 0: î en V France, Total n = 407 frst clutches and 141 replacement clutc es. Fgure 3 : Devenr (écloson,'échecs par prédatondes poules couveåses, pandtresš tructon des œufs, par pratques agrcoles, par d autres causes) df-es Pffmlîãn pontes et des pontes de remplacement de DGFCWX 9f'*"=`je efdlffpef '3 ' du couvert de ndfcaton (de haut en bas : ééments lneares, jachefeä. PfÎUf<'=199$, autres cultures, fourrages, céréales). Les chffres entre parentheses corrlelspronddelï nv nffnntfe fne nnntne nn rnflrnnnnt Im: 10 etm: rl átrln rl flnntrn- nr le ị 1l 1 -e! ` f- l.:. E. Bro et al. 209 Replacement clutches were slghtly more senstve to predaton than females (n cereals, lnear features; Fgure 3). g Clutches lad closer to feld margns (see Fgure 3 n BH eta., 2000a) were not more lkely to be depredated (X2, < 0.01, p = 0.462, one-taled test; test performed wth clutches lad n cereals for sample sze reason, poolng frst and replacement clutches). The predaton rate on clutches was roughly constant wth regard to the nearest dstance to feld margn: 5 1m: 29% (n = 31), ]1-5m]: 32.7% (n = 49), ]5-10m]: 24.6% (n = 61), ]10-15m]: 25% (n = 40), 2 15 m: 21.9% (n = 96). The global mpact of farmng practces on frst clutches (n = 407) was moderate, less than 10% of frst clutches were destroyed drectly and ndrectly (through the death of ncubatng females). The most rsky covers were fodder crops, meadows and CAP set-asde (Fgure 3). Replacement clutches (n = 141) suffered more (29%) from farmng practces, especally n cereals, other crops and fodder crops (Fgure 3). ln cereals, hanfestng n July was the man cause of destructon of clutches, nvolvng manly replacement clutches (42% of replacement clutches, n = 45) and a few late frst clutches (3% of frst clutches, n = 232). rrgaton n May-June n ntensve cereal producton regons caused clutch deserton (4 cases of falure of frst clutches out of 10 cases due to farmng practces). ln sugar beets (one of the other crops ),- farmng dsturbance was due to rrgaton and caused deserton of some replacement clutches. ln fodder crops, clutches manly suffered from hay and slage cuttng n June that affected both frst (7 out of A18 clutches) and replacement clutches (3 out of 7 clutches). As for fodder crops, clutches lad n set-asde suffered from cuttng from late May to md-july. The few clutches n pastures were manly destroyed by cattle. The other causes of nest falure were hghway mantenance and road traffc. Some clutches lad on banks or roadsdes were destroyed by hghway mantenance (mowng tall Weeds n late June s early July for safety reasons), and some hens were klled by cars. III.4. PREDATIN RATE IN RELATIN T PREDATR ABUNDANCE For frst clutches the falure rate due to predaton ranged from 19% (A, n = 41) to 54% (C, n = 26) across study areas, and for replacement clutches from 20% (D, n =15) to 67% (F, n = 6). Falure of clutches due to predaton was caused drectly by destructon of eggs or ndrectly by death of ncubatng females. The predaton on ncubatng females v clutches dd not dffer wth clutch type (X2, = 1.96, n.s.; Fgure 2)., The predaton rate on ncubatng females dd not vary sgnfcantly between clutch type and across study areas (nteracton "c utch type * study area": X2, = 5.08, n.s.; clutch type: X2,= 2.7, n.s.; study area: X2 = 14.42, n.s.; range 8-27% for frst clutches and 7-21 % for replacement clutches). The predaton rate on clutches vared across study areas dependng upon clutch type (nteracton clutch type * study area": 129 = 26.06, p *= 0.002). The predaton rate on frst clutches vared across study areas [X2 = 44.18, p < 0.001, range between 0% (area F,- n = 26) and 46% (area C, n = 26)], whereas the predaton rate on replacement clutches dd not vary sgnfcantly across study areas (7529 = 13.72, n.s., rsxnfln R/. flnû/\

7 I 210 Ecologca correlates of nestng success of grey partrdge rate across study areas was sgnfcantly correlated wth the abundance of mustelds (X2, = 5.76, p = 0.008, one-taled weghted test; Fgure 4), whereas the abundance of red foxes was not a sgnfcant correlate (X2, = 0.27, n.s.). The abundance of mustelds was correlated to habtat structure (presence of pastures, dversty of crops and feld sze), beng more abundant n tradtonal farmland than n open plans of cereals (BR, 1998). However, the relatonshps were lnked to one partcular study area, H, a landscape crss-crossed by hedges. tcutches t' Predat rs on rate Ul -B oo N - DC) 0 * Jr I - I I l A70 80 ndex of musteld abundance Fgure 4: Correlaton between the predaton rate on frst clutches of grey partrdge, Perdx perdx, and musteld ndex of abundance. Letters refer to study areas, and crcle dameters to the weght of each observaton (recorded sample sze of clutches, for nstance, I crcle dameter refers to 52 clutches). Musteld abundance ndex s calculated by the number of postve transects. Ten study areas n North-Central France, , total n = 407 frst clutches. Fgure 4 : Corrélaton entre le taux de prédaton sur les premères pontes de perdrx grses, Perdx perdx, et l'ndce d'abondance des mustéldés. Les lettres se réfèrent aux stes d'étude, les damètres des cercles aux pods relatfs des dfférentes observatons (effectfs des nds suvs, par exemple le damètre du cercle I se réfère à l'effectf 52). L'ndce d'abondance des mustéldés correspond au pourcentage de transects postfs. Dx stes d'étude dans le Centre-Nord de la France, Effectf total n = 407 premères pontes. IV. DISCUSSIN C9, Ths work s part of an extensve study of the grey partrdge, and s based on recent data collected n contrastngtypes of arable farmland n North-Central France. lt s an update of our knowledge of nestng ecology by dentfyng and quantfyng the causes of clutch falure under dfferent farmng regmes and pred , ut.. =.....m..+rm. s nf m" mnnrmnm n ranrllv evolvnd svstems E. Bro et al. 211 We found that clutch losses could be partcularly hgh (70%) dependng upon areas. When comparng our fgures to those reported by AUBINEAU (1981) and BRUN (1991) n the s and revewed by BIRKAN and JACB (1988) n North-Central France, t seems that the falure rate of clutches has ncreased. However, such hgh clutch losses have been reported n other countres (revews n PTTS, 1980; AUBINEAU, 1981; BIRKAN and JACB, 1988). ur data clearly show that the man proxmate cause of the falure of frst clutches s predaton, wth farmng practces beng secondarly mportant, whereas replacement clutches suffered more from farmng practces v other causes than frst clutches. However, the relatve balance of falure due to predaton v farmng practces vared only moderately n space, suggestng that the dfferences n nest success rather result from a global rsk of falure than a factor n partcular. The fate of clutches hghly depended upon nestng cover used by partrdges - and related rsks of farmng destructon and predaton. Thus, to be effcent management should address these ponts. ur most orgnal result s the correlaton between the spatal varaton of the frst clutch-predaton rate and the spatal varaton of musteld abundance. However, ths correlaton does not demonstrate a causal relatonshp. lt could reflect a convergence of choces for the same habtat by partrdges (for nstance choce of nestng cover, BR et al., 2000a, dependng upon cover avalablty), and by mustelds. Nevertheless, n practce, the relaton suggests that reducng the abundance (or trappng near the most rsky nestng cover) of targeted speces lke the stone marten mght reduce clutch losses, at least on areas where ths very concern has been dentfed. Farmng practces such as rrgaton and harvestng operatons occur n most cover types durng the ncubatng perod of the grey partrdge, so that even the most sutable cover may turn nto an ecologcal trap, especally for replacement clutches. Accordng to PTTS (1980) revew, mowng could account for 10-90% of clutch losses dependng upon studes. Ths s a general pont of concern for ground-nestng brds lvng n farmlands (TUCKER, 1997). ur study provdes four man results on the relaton between clutch success rate and nestng cover. 1) We confrm that the preferred nestng cover, cereals (BR et al., 2000a), s also the best cover for frst clutches (despte deserton cases due to rrgaton as prevously reported, e.g. BlRKAN ef al., 1990; SERRE et al., 1995), but the destructon rate of replacement clutches due to harvestng s sgnfcant. 2) Clutches lad n the second preferred nestng cover, lnear landscape features (BR et al., 2000a), sulfer from predaton and to a lesser extent from weed management. Although ths dlemma has been rased numerous tmes, t could be solved n practce by concentratng trappng efforts on these partcular landscape features. The queston s to know whether such local trappng could be effcent. 3) From a global pont of vew, the applcaton of the set-asde scheme dd not appear (n ) to be benefcal to clutch success because cuttng occurred durng the ncubaton perod. Although hunter assocatons have advocated the benefts for Wldlfe of the Envronment and wldlfe" opton (prohbton of mowng, weed control by cuttng replaced by chemcal control- e.g. RAMEAU and CITRN, 1996) whle provdng fnancal ncentves, ths opton has only been adopted for a small proporton of set-asde land (1.7% n 1995; ARNAUDUC, 1996). 4) For the same reasons, fodder cropssuch as lucerne or clover were not a favourable type' of nestng cover despte beng,attractve (e.g. PEETERS and DECAMPS. 1998). Ths result has been well documented elsewhere (e.g. BARBIER,

8 E. Bro et al Ecologcal correlates of nestng success of grey partrdge V. MANAGEMENT IMPLICATINS: STRATEGY FR GREY PARTRIDGE CNSERVATIN UNDER THE CTE SCHEME Detaled research studes on the grey partrdge n western Europe have dentfed the man threats to ths speces, Whch led to prescrbe both Wse management of target-predator populatons and changes n farmng practces for habtat management as remedal actons to allevate clutch losses. Predaton rate may. lkely be reduced drectly by controllng predator abundance. The seasonal local control of targeted predator speces (deally dentfed n a prelmnary dagnoss survey) could consttute Wse management of predator presence. However, the queston of the mpact of predator control on populaton dynamcs of the prey s stll under debate despte sold scentfc research on ths topc. The experment conducted n England by the Game Conservancy Trust (GCT) produced substantal results both on autumn and sprng denstes of partrdges (TAPPER et al., 1996). ne could argue that such ntensty of predator control could only be acheved underexpermental condtons and dd not correspond to the norm for large areas. lndeed, other smlar experments conducted elsewhere dd not lead to convncng results«(serre et af., 1995; revew n CTE and SUTHERLAND, 1997). However, recent evdence ndcates that motvated people reduce red fox densty over large areas (GCT, 1998) despte the need for contnuous and hgh ntensty control (HEYDN and REYNLDS, 2000). Less ntensvely, trappng could be focused on partcular features known, on the one hand, to be sutable nestng cover for grey partrdge (feld margns and lnear landscape features) and, on the other hand, used by predators (carnvores such as the stone marten travel along lnear tems to fnd ther preys; e.g. LÉGER, 1996). lf focused trappng could provde satsfactory results, t would be effcent because efforts would be concentrated on hgh-rsk areas for nestng partrdges. Han/estng and rrgaton are the two man farmng operatons affectng grey partrdge clutches. An effcent remedal acton Would be to delay/suppress these farmng operatons n feld margns Where most clutches are located (BR et al., 2000a). Ths measure s not unrealstc n the context of the new orentaton of the European Unon's Common Agrcultural Polcy (CAP). lndeed, owng to current surpluses and ncreasng envronmental problems (e.g., polluton of sol and Water, endangered speces), the CAP has recently focused on the envronmental aspects of agrculture and provdes fnancal ncentves for farmers that voluntarly address these concerns. ln France, ths European drectve s appled Wthn the natonal scheme Contrat Terrtoral d'explotaton' (CTE). Dfferent measures n favour of Wldlfe such as plantng hedges or strp covers, Introducng supplementary crop types, smplfyng ploughng, dvdng large felds, mprovng the envronmental value of set-asde, etc. are proposed to farmers through a contract wth the mnstry of Agrculture. All optons are grantaded (from 300-4,000 FF/ha/year) to compensate for the loss of ncome, the addtve work or cost (MINISTERE DE L'AGRlCULTURE ET DE LA PECHE, 1999). To maxmze the benefts from ths scheme, management gudelnes must be based on scentfc research, be effectve for Wldlfe, and acceptable by farmers n terms of ease of mplementaton and economcs - to compete nternatonally farmng must ndeed reman proftable.. Wthn ths framework, Whch s a contextual polcy, our results suggest three..,.,..,s...«s= n rentree clutch destructon/deserton *due to farmng.,, at, 1999; BR eta., 2000a): Ths means no pestcdes, rrgaton or cuttng. Ths could be acheved n the context of the CTE scheme by establshng grassy or cereal-based brd-cover strps located ether at feld boundares close to lanes or n the mddle of large felds (to create sutable nestng stes) (optons 14.2, 14.3 of the CTE scheme: MINISTÈRE DE L'AGRlCULTURE ET DE LA PECHE, 1999). These boundares could ncorporate conservaton headlands (opton 16.2) to provde, n addton to undsturbed nestng cover, nsect-rch chck-feedng habtat (selectve herbcdes, no nsectcdes after md-march; GCT, 1995). Flotatonal set-asde strps and conservaton headlands have prevously been shown to mprove reproductve success of grey partrdges (e.g., BLAKE and DWELL, 1992). When/where measure (1) cannot be appled, stubbles should be cut at a mnmum heght 'of 20 cm above ground to preserve ncubatng females and eggs. 2) mplement Envronment and Wldlfe set-asde (MINISTERE DE L'AGRl- CULTURE ET DE LA PÊCHE, 1996; optons 14.1, 14.3, 16.4, 16.6 of the CTE scheme) nstead of tradtonal (e.g. BESNARD, 1996) or ndustral optons. Weed plants are controlled by selectve herbcdes (e.g. RAMEAU and CITRN, 1996). 3) ln regons where fodder crops are abundant, a network of cereal-based mxture strps could be establshed around fodder crop as alternatve nestng habtat (BARBIER,-1979); these strps should be managed as n (1). The two alternatves for mantenance are: to mow feld margns early (before md-may) or late (after 20 July) wth no farmng operatons between these two dates (adaptaton of the 14.4 opton), and to adust cuttng heght at cm above ground (BARBIER, 1979). These measures would help to reduce clutch falure. The ssue s Whether the Wdespread.applcaton of these measures s realstc from a socal and economc pont of vew. Assumng that ncentves are/wll be attractve enough for ntensve arable farms, such a polcy offers an excellent opportunty to assst the grey partrdge and other Wldlfe n cereal ecosystems. ACKNWLEDGEMENTS We acknowledge E. DANCHIN, M. MASST, J.W. CNNELLY, G.R. PTTS and N.J. AEBISCHER, all of whom provded constructve comments on the draft of the manuscrpt that greatly mproved the paper. The study Was funded by the "ffce natonal de la chasse et de la faune sauvage" and the Unon natonale des fédératons départementales de chasseurs". Data Were collected by feld techncans and students of eght "Fédératons départementales de chasseurs" (Aube, Loret, Marne, Nord, Pas-de-Calas, Sarthe, Sene Martme and Somme). We are ndebted to the landowners for gvng us permsson to Work on ther land. We also thank hunters for ther cooperaton n the study. REFERENCES AEBISCHER N.J. & KAVANAGH B. (1997). - Grey partrdge. ln: The EBCC atlas of European Breedng Brds. Ther dstrbuton and abundance, W.J.M. HAGEMEIJER & M.J. BLAIFI, eds. Publshed for the European Brd Census Councl. by T. & A.D. Poyser: 903. * AFINAUDUG J.-F! (1996). - Les contrats jachères faune sauvage. Bull. Mens. ff. Natl. Chasse, 214: AUBINEAU J. (1981). - Quelques résultats sur la reproducton de la perdrx grseldans des terrtores de Beauce. du Sossonnas et de la Marne de Bull..l\Ilens. ff. Natl. Chasse, 21: 7-9. M I Hu-rm mdnnne des coupes de luzerne à déshydrater sur la faune locale. Bull. Mens. J

9 ' ' 21,4 Ecologcal correlates of nestng success of grey partrdge E. Bro et al. 215 BESNARD A. (1996). - Jachères ordnares. Quels sont les couverts les meux adaptés? Bull. Mens. ff. Natl. Chasse, 214: BIRKAN M. & JACB M. (1988). - La perdrx grse. Hater, Pars, 284 p. BIRKAN M., SERRE D., PELARD E. & SKIBNIENSKI S. (1990). - Effects of rrgaton on adult mortalty and reproducton of gray partrdge n a wheat farmng system. lr: Perdx V: Gray partrdge and rngnecked pheasant workshop, K.E. CHURCH, R.E. WARNER & S.J. BRADY, eds. Kans. Dep. Wldl. and Parks, Empora: BLAKE K. & DWELL S. (1992). e Game brds n set-asde. The Game Consenrancy Annual Revew, 24: BLANK T.H., SUTHWD T.R.E. & CRSS D.R. (1967). - The ecology of the partrdge. I. utlne of populaton processes wth partcular reference to chck mortalty and nest densty. J. Anm. Ecol., 36: BR E. (1998). - Corrélats envronnementaux du statut démographque de la perdrx grse (Perdx perdíx) en France. Thèse de doctorat d'ecologle de l'unversté Pars XI, 269 p. + annexes. (n French). BR E., CLBERT J. & REITZ F. (1999). - Effects of radotransmtters on survval and reproductve success of gray partrdge. J. Wldl. Manage., 63: BR E., REITZ F. & CLBERT J. (2000a). - Nest-ste selecton of grey partrdge (Perdx perdx) on agrcultural lands n North-Central France. Game Wldl. Scence, 17: BR E., SARRAZIN F., CLBERT J. & REITZ F. (20000). - Demography and the declne of the grey partrdge Perdx perdx n France.-J. Appl. Eco., 37: BR E., REITZ F., CLBERT J., MIGT F! & MASST M. (2001). - Dagnosng the envronmental causes of the declne n Grey Partrdge survval n France. Ibs, BRUN J.-C. (1991). - La ndfcaton des perdrx : une perode ben délcate. Bull. Mens. ff. Natl. Chasse, 161: CTE l.m. & SUTHERLAND W.J. (1997). - The effectveness of removng predators to protect brd populatons. Consen/aton Bology, 11: GCT (1995). - Gudelnes for the management of feld margns. Recommendatons for the management of feld margns, ncorporatng the feld boundary, boundary strp and the outer few meters of the ' cereal crop. Fact 2 Sheet. The GCT's Farmland Ecology Unt, Fordngbrdge, 16 p. GCT (1998). - Foxes n myth and realty. The Game Conservancy Annual Revew, 30: GREEN R.E. (1984). - The feedng ecology and survval of partrdge chcks pllectors rufa and Perdx perdx) on arable farmland n East Angla. J. Appl. Ecol., 21: » A HAGEMEIJER W.J.M. & BLAIR M.J. (1997). - The EBCC atlas of European breedng brds. Ther dstrbuton * and abundance. Publshed for the European Brd Census Councl by Poyser, London, 903 p. HEYDN & REYNLDS J. (2000). - Demography of rural foxes (l/ufpes vulpes) n relaton to cul ntensty n three contrastng regons of Brtan. J. Zool. London, 251: KAISER W. (1998). - Grey partrdge (Perdxperofx) survval n relaton to habtat qualty. fn: Proceedngs of the XXlIl' congress of the Internatonal Unon of Game Bologsts (IUBG), N.W. STHERTN, F! GRANVAL, R HAVET & N.J. AEBISCHER, eds. Gber Faune Sauvage, Game Wldl., 15 (Hors Sére Tome 1): LEGER F. (1996). - bservatons sur l'écologe de la foune Martes fona dans le bocage du Perche. Premère parte: Utlsaton de l'espace. Bull. Mens. ff. Natl. Chasse, 209: McCULLAGH F! & NELDER J.A. (1983). - Generalzed lnear models. Second edton. Chapman and Hall, New York, 511 p. MINISTERE DE L'AGRlCULTURE ET DE LA PECHE (1996). - Modaltés partculères d'entreten des jachères faune sauvage. Crculare mnstérelle d'applcaton DEPSEISDEAI N C à Messeurs les préfets de régons et de départements. MINISTERE DE L'AGRICULTURE ET DE LA PECHE (1999). - Mse en œuvre des Contrats Terrtoraux d'explotaton. Crculare mnstérelle d'appicat0n DEPSE/SDEN N C à Messeurs les préfets de régons et de départements. ( NICHLS J.D., PERCIVAL H.F., CN R.A., GNRY M.J., HENSLER G.L. & HINES J.E. (1984). - bserver Vstaton and success of mournng dove nests: a feld experment. Auk, 101: PANEK M. (1997). - The effect of agrcultural landscape structure on food resources and survval of grey partrdge Perdx perdx chcks n Poland. J. Appl. Ecol., 34: PANEK M. & KAMIENIARZ R. (1998). - Agrcutural landscape structure and densty of grey partrdge (Perdx perdd populatons n Poland. In: Proceedngs of the Perdx VII Symposum on Partrdges, uals and Pheasants, M. BIRKAN, L.M. SMITH, N.J. AEBISCHER, EJ. PURRY & RA. RBERT- SN, eds. Gber Faune Sauvage, Game Wldl., 15: PEETERS A. & DEGAMPS C. (1998). - Chox et geston de couverts herbacés dans les jachères et tournères faunstques. ln: Proceedngs of the XXlIIrd congress of the Internatonal Unon of Game Bologste. N.W. STHERTN, Ph. GRANVAL, F! HAVET & N.J. AEBISCHER, eds. Gber Faune Sauvage, Game Wldl., 15 (Hors Sére Tome 1): PDLER H. & RGERS D. (1975). - A new method for the dentfcaton of key factors from lfe-table data..j. Anm. Ecol., 44: V - * 1 l É 1 1 PTTS G.R. (1997). - Chapter 6. Cereal farmng, pestcdes and grey partrdge. ln: Farmng and brds n Europe. The common agrcultural polcy and ts mplcatons for brd consomaton, D.J. PAIN & M.W. PIENKWSKI, eds. Academc Press, London: PTTS G.R. & AEBISCHER N.J. (1991). - Modellng the populaton dynamcs of the grey partrdge: conservaton and management. fn: Brd populaton studes: relevance to conservaton and management, C.M. PERRINS, J.D. LEBRETN & G.J.M. HIRNS, eds. xford Unversty Press, xford: PTTS G.R. Bt AEBISCHER N.J. (1995). - Populaton dynamcs of the grey partrdge Perdx perdx : montorng, modellng and management. Ibs, 137: S29-S37. RAMEAU C. & CITRN G. (1996). - Entreten des jachères. Bull. Mens. ff. Natl. Chasse, 214: REITZ F. (1990). - La stuaton de la Perdrx grse dans le Nord et le Bassn parsen en Bull. Mens. ff. Natl. Chasse, 143: REITZ F. (1997). - Les perdrx dans le centre~nord de la France. Bull. Mens. ff. Natl. Chasse, 129: 2-8. REITZ F. & BERGER F. (1995). - Les perdrx en 1994 dans le Nord, le Bassn parsen et le Centre : attenton, danger l Bull. Mens. ff. Natl. Chasse, 197: REITZ F., BR E., MAYT F! & MIGT F! (1999). - Influence de I'habtat et de la prédaton sur la démographe des perdrx grses. Bull. Mens. ff. Natl. Chasse, 240: SAS INSTITUTE (1994). - SAS / STAT user's gude. Verson 6.0 Cary, North Carolna, USA. SERRE D., REITZ F., MIGT P., STAHL F!, DAVID Y & STTEJEAU 1! (1995). - Mleu, prédaton et pégeage en Beauce du Loret : quel mpact sur la démographe de la perdrx? Bull. Mens. ff. Natl. Chasse, 201: SKAL R.R. & RHLF F.J. (1981). - Bometry, the prncples and practce of statstcs n bologcal research. Second edton. W.H. Freeman, New York, 859 p. STHERTN N.W. 8. RBERTSN F!A. (1990). - ndrect mpacts of pestcdes on the producton of wld game brds En Brtan. In: Perdx V: Gray partrdge and rng-necked pheasant workshop, K.E. GHURCH, R.E. WARNER & S.J. BRADY, eds. Kans. Dept. Wldl. and Parks, Empora: STAHL R (1990). - Suv de l'abondance d'une populaton de renards (I/ulpesvulpes) par comptages nocturnes : évaluaton de la méthode. Gber Faune Sauvage, Game Wldl., 7: STAHL F! & MIGT R (1990). - Varablté et sensblté d'un ndce d'abondance obtenu par comptages nocturnes chez le renard (Vufpes vulpes). Gber Faune Sauvage, Game Wldl., 7: TAFPER S.C., PTTS G.R. & BRCKLESS M.H. (1996). - The effects of an expermental reducton n predaton pressure on the breedng success and populaton densty of grey partrdges Perdx perdíx. J. Appl. Ecol., 33: TUCKER G. (1997). - Chapter 4. Príortes for brd consenraton n Europe: the mportance of the farmed landscape. In: Farmng and brds n Europe. The common agrcultural polcy and ts mplcatons for brd conservaton, D.J. PAIN & M.W. PIENKWSKI, eds. Academc Press: APPEN DIX (NEXT PAGES) Descrpton of data set. Sample sze, clutch sze and clutch fate of frst and replacement clutches of grey partrdges, Perdx perdx, and predator abundance are gven for each study area. North-Central France, ANNEXE (PAGES SUIVANTES) Données concernant les premères pontes, les pontes de remplacement, les trosèmes pontes, les pontes de rang ndétermné de perdrx grses, Perdx perdx et fabondance des prédateurs, suvant dx stes d'étude (A-J) du Centre-Nord de la France, Nombre total de pontes, nombre de pontes réusses, talle de ponte, nombre de pontes ratées, au devenr nconnu, désertées à cause du radopstage. Pertes de pontes par mortalté des poules couveuses, destructon des œufs ou abandon du nd. Causes de perte de ponte : prédaton sur poule couveuse, prédaton sur œufs, pratques culturales, autres causes, causes non dentfées. Abondances des mustéldés en pourcentage detransects postfs, et de renards en ndvdus par 10 km.

10 7 21 & B m Mm Ø%Q mmå%ñ9 BN@gbØCQÈQWW mvgg μ25tm$umug W \gg 8š ÊQN g 82 ÎE2W kg gfag FäóÊ:gVw"šU \gm mwa N mg33 gmä \àfm fge8w:mê2μmaåwuœœ 8%% 2NUm NM 10 C m 5 švmmm E 1 ØNG "N Qm22 uw= NN mrq înî *NN W N 2M N mn \ W m m N ggã UmCm L ggg550 $UäQ 0μg9n $me0 m g0c CCüUQn 9M N NG W r WNmu VW NØmu " \IGN N Ctm$u μ Cμš0u ãšo V $gp med F Ego œ mh 0=`oV vo mh 0Ø GM 0 A UVUœ;0Um oum Uœfœœwu wü050 * Ô EE É š 5 åš F N NQ? m 5 UQÊUF r M gg n g rw Da ' S U CCQSS sm nesmg C0rrGmœ S ECm g mm ÊÊÊÊ ÎEHÊEädGä:îflä \ QØ NØ! V HI whahj"h\v l H` "Uhå m l`îiu` 'ri PU' îu ` flv"»%`" `\l Il`:lII l;l : W vh(vv HHI VfïLhH H > VI VÊww HEgum Uœsåíñ ÊÊÊ ÊT8 N Hš3 HNN2 HQ:md HQQN Hom 0% :E k1\`1\ œmbg wgo5q ãmœmszm06 ñ :QFN2œ 1tW œ m wwm Wmμ0 hoa $2U E28N% 5W mm N mg *N NWW QN ägg m0nmu0emu N ØgãóE50 wgšäãm'm" gg 0ÊUmn $mm whμmgecuãš n Êf rb 06 mg; šuww wnîœm NmW:N mé Ø Y Um NW* *Mm œw 2 tø Uã% r%on μñmu FB'0 M$gmv 0Cãgr Fšgn mš BJ *sg NF CN N 7švgum gn D0m$U $ S 1" 0 ÎÎN c MW WØ WC>`öv5 Ê$Q:o * ÎN W M\H *Wguãu 8œ*og atgqê Ø GNVAN5Ê V agê Nœ GQ Ag ÈH@g HNg HNgÈ HQQÈ HE;Ê HQQ@ HQ:È Hãîd HN:Ê Hme mm HEV gmfg NNaE mn@è 39 W@N@N mœp2ju ãmouãho :IØNFL! Q0 m< È*NmNm \mr QU@Nmö 2NNÈ3 N`WmÊJo ho $$ 5 9 MEM >Uμœ

11 218 Ecologcal correlates of nestng success of grey partrdge Game and Wldlfe Scence, Vol. 17 (4), December 2000, p = Éuss TE DEs N Ds DE PEaDn1x c.r ses (PEnDxPanox) DANs LEs Aeno-EcosYsTÈ v Es Du CENTRE-Nono DE LA FRANCE, =relat on AvEc LE couvent DE N D =1cAT on Er L'AsoNDANcE DE PRÉDATEURS E. BR, F. FIEITZ, J. CLBERT et F! MAYT MTS-CLÉS : Perdrx grse, Perdx perdx, devenr des pontes, prédaton, pratque agrcole, abondance de prédateurs, couvert de ndfcaton, aménagement, agro-écosystème, France. RESUME Nous avons suv 1 09 poules perdrx grses, Perdx perdx, sauvages par radopstage de mars à septembre sur dx stes d'étude contrastes dans le but d'dentfer et de quantfer les causes d'échec des nds. Nous avons noté le devenr de 407 premères pontes et 141 pontes de remplacement. Parallèlement, nous avons estmé l'abondance des renards, Vulpes vulpes, et des mustéldés, Musteldae, et noté l'habtat des stes de ndfcaton. Le taux de succès des pontes a varé sgnfcatvement entre les stes d'étude (de 31%, n = 16, à 73%, n = 46) pour les premères pontes (p = 0,013), mas pas pour les pontes de remplacement (de 0%, n = 6, à 53%, n = 16). Les céréales ont consttué le melleur couvert pour les premères pontes (taux de succès de 66%, n = 232) tands que les pontes de remplacement y ont souffert des mossons en jullet (taux de succès de 29%, n = 45). Cependant, la varaton du taux de succès des premères pontes entre stes d'étude n'état pas corrélée à la varaton de Fmportance des céréales dans Fassolement. La prncpale cause d'échec des premères pontes (n = 150 cas d'échecs avec cause dentfée) a été la prédaton (70%), suve par les pratques agrcoles (22%). Les pontes de remplacement (n = 79 cas d'échecs avec cause dentfée) ont échoué du fat de la prédaton (51%) et des pratques agrcoles (43%). Les carnvores terrestres (renard, mustéldés et chat domestque) ont été responsables de 66% des cas dentfés cle prédaton sur les poules couveuses (n =76) et de 36% de ceux drectement sur œufs (n = 69). Le taux de prédaton des premères pontes état postvement corrélé à l'abondance des mustéldés (p = 0,008), mas pas à celle des renards. Le taux de prédaton a été partculèrement élevé (46%) dans les éléments lnéares du paysage (n = 46). Les résultats obtenus justfent l'usage d'un contrôle localsé de certanes espèces de mustéldés pour protéger des nds de perdrx, en complément de mesures de geston des bords de champs, dans le cadre des nouvelles orentatons de la PAC déclnées en France par le programme "Contrat Terrtoral d'explotaton avec pour objectf de préserver la perdrx grse dans les régons de céréalículture ntensve. Receved 28 August 2000, accepted 12 January EFFICACY AND SELECTIVITY F HARE (LEPUS EURPAEUS) " BX TRAPS " v. BRAY (*) and Y. LENARD (**) ffce natonal de la chasse et de la faune sauvage, Drecton des études et de la recherche (*) 11, avenue de Fontmaure, F Chamaleres, e-mal: ciermont@onc.gouv.f' (**) Le Bouchet, F Dry KEY WRD5: European hate, Lepus europaeus, capture, trappng, box, trap, log-lnear model. - ABSTFIACT The choce of a capture technque depends on the bology of the speces and the problem to be studed. lt also depends on ts effcency, the methods cost and selectvty, and the ensuant rsks of mortalty. Hares, Lepus europaeus, were captured wth box traps n two study areas (area 1, Bellay-en-Vexn, Val d'0l's, and area 2. Charel-Cntrat and Montord, Aller). The 13-km* study area 1 s typcal ofa regon ofntensve cultvaton and the 17-km* study area 2 ofan area of polyculture-breedng. Average sze of the felds n area 1 was larger than that of the felds n area 2 (12 ha vs 2 ha), and ts landscape was less dversfed. Because of the dfferent huntng pressures n each study area, t was possble to sngle out n each of them two zones wth dfferent hare denstes (8 and 26 hares per 100 ha n study area 1; 14 and 49 n study area 2). Unbated box traps wth drop gates were set up n open habtats, on bare ground or grounds wth low vegetaton. Prelmnary tests carred out n 1985 and 1986 showed the greater effcacy (number of captures/100 nghts traps) of large box traps (L >< w h: >< 40 cm) wth respect to small ones ( cm) (3.3 vs 112, P < ). Large box traps (the only ones used between 1987 and 1996) were more effectve on study area 1 (5. 3: 247 hares captured n 4,636 nghts traps n May-ctober ) than on study area 2 (2.3: 186 hares n 7,938 nghts < traps n May-ctober ). Trap mortalty rate and recapture rate both amounted to 3% (n = 433). Trappng yeld was 0.43 captured hare/frapper hour on area 1 and 0.19 on area 2,.e. three to seven tmes hgher than the yeld obtaned for captures by beats nto nets. An analyss of the effect of " trappng area ", " lumnosty at nght ", " sol humdty ", " hare densty n sprng " and the perod of trappng " on capture success, was carred out by fttng log-lnear models to the data and the applcaton of Akake's nformaton crteron (A.l. C). ln all areas, the degree of lumnosty at nght had a greater mpact on trappng success than the degree of sol humdty. n both areas, trappng became more effcent wth greater nght lumnosty (a 2.2 to 5.0 average ncrease). ln contrast, the effect of hare densty on capture success vared by area. Because of ths result we were compelled to analwe n addton, and one by one, the nfluence of the other varables n each trappng area. n each of these two areas, trappng effcency ncreased over the seasons and wth nght lumnosty (whch ncreased on average from 2.1 n May-June n dark nghts, to 7.4 n September-ctober n clear l