Bsic nd Applied Ecology 8 (2007) 354 363 www.elsevier.de/be Ptch choice of vin herbivores long migrtion trjectory From Temperte to Arctic A.J. vn der Grf,, J. Sthl b, G.F. Veen c, R.M. Hving, R.H. Drent Animl Ecology Group, Centre for Ecologicl nd Evolutionry Studies, University of Groningen, P.O. Box 14, 9750AA Hren, The Netherlnds b Lndscpe Ecology Group, University of Oldenburg, 26111 Oldenburg, Germny c Community nd Conservtion Ecology Group, Centre for Ecologicl nd Evolutionry Studies, University of Groningen, P.O.Box 14, 9750AA Hren, The Netherlnds Received 27 Mrch 2006; ccepted 14 July 2006 KEYWORDS Brncle goose; Brnt leucopsis; Flywy; Forge qulity; Grzing experiment; Plnt biomss; Plnt herbivore interctions; Slt mrsh Summry Migrtory wterfowl species seem to trck temporl nd sptil pulses of optiml forge vilbility on their wy from temperte wintering to rctic breeding sites. In order to unrvel the reltive contribution of forge qulity nd forge biomss to forging choices in vin herbivores, we experimentlly mnipulted biomss nd qulity of min forge plnts through fertilistion nd grzing exclusion t three sites long the flywy of brncle geese, Brnt leucopsis. Fertilistion incresed the nitrogen content of the forge nd grzing exclusion incresed biomss levels. Mnipulted plots were offered to wild geese in rndom block experimentl design nd goose visittion ws mesured through dropping counts. At ll sites there ws trend towrds higher preference of plots with incresed qulity nd verge biomss bove plots with n verge qulity nd incresed biomss. Generlly, geese preferred plots with highest stnding crop of nitrogen. The numericl response of the geese to forge chnges ws supported by behviourl observtions t the Bltic site. We conclude tht for migrting brncle geese the bottlenecks in the stnding crop of nitrogen pper to lie in the limited biomss vilbility t the Bltic stopover site nd the limited nutrient content of food in the Arctic breeding site, restricting the potentil nutrient intke on these sites. & 2006 Gesellschft für Ökologie. Published by Elsevier GmbH. All rights reserved. Zusmmenfssung Viele pflnzenfressende Wsservogelrten scheinen entlng ihres Zugweges von gemäßigten Überwinterungs- in rktische Brutgebiete zeitlichen und räumlichen Impulsen optimler Nhrungsverfügbrkeit zu folgen.um den reltiven Beitrg von Corresponding uthor. Tel.: +31 503632264; fx: +31 503635205. E-mil ddress:.j.vn.der.grf@rug.nl (A.J. vn der Grf). 1439-1791/$ - see front mtter & 2006 Gesellschft für Ökologie. Published by Elsevier GmbH. All rights reserved. doi:10.1016/j.be.2006.07.001
Ptch choice of migrting herbivores 355 Nhrungsqulität und -biomsse zur Nhrungswhl der Herbivoren zu bestimmen, hben wir entlng des Zugweges der Nonnengns, Brnt leucopsis, Biomsse und Qulität der Huptnhrungspflnzen experimentell mnipuliert. Düngung erhöhte den Stickstoffgehlt der Nhrungspflnzen, der Ausschluß von Beweidung erhöhte ds Biomsseniveu. Die mnipulierten Flächen wurden durchziehenden Wildgänsen in einem rndom block Experimentldesign ngeboten, Gänsenutzung wurde mit Hilfe von Kotzählungen dokumentiert. In llen Gebieten wurde ein Trend zu erhöhter Präferenz von Nhrungsflächen mit erhöhter Qulität und mittlerer Biomsse im Vergleich zu Flächen mittlerer Qulität und erhöhter Biomsse festgestellt. Die Gänse bevorzugten llgemein Flächen mit höchster Stickstoffverfügbrkeit pro Oberflächeneinheit. Die numerische Rektion der Gänse uf Nhrungsveränderungen wurde durch Verhltensbeobchtungen in einem der Untersuchungsgebiete unterstützt. Wir folgern, dss die Limitierung von Stickstoffverfügbrkeit für ziehende Nonnengänse in den Zwischenstoppgebieten im Ostseerum in der begrenzten Biomsseverfügbrkeit begründet liegt, während in den subrktischen Brutgebieten der Nährstoffgehlt der Pflnzen beschränkend wirkt. Beide Prmeter begrenzen die Nährstoffufnhme in den jeweiligen Gebieten & 2006 Gesellschft für Ökologie. Published by Elsevier GmbH. All rights reserved. Introduction Avin herbivores of the Northern hemisphere breed in the Arctic nd winter in more southern, temperte res. During spring migrtion the birds hve to blnce their energy expenditure nd food intke in order to build up sufficient energy reserves to be ble to migrte to their breeding res nd breed successfully (Bety, Guthier, & Giroux, 2003; Drent, Both, Green, Mdsen, & Piersm, 2003; Wrd et l., 2005). The birds usully migrte in severl distinct steps, nd refuel nd rest t ech stopover site (Eichhorn, Afnsyev, Drent, & Vn der Jeugd, 2006; Green, Alerstm, Clusen, Drent, & Ebbinge, 2002; Nolet, Andreev, Clusen, Poot, & Wessel, 2001). Plnt forge vilbility nd qulity on these sites ply crucil role for these smll herbivores, s the mount of body reserves ccumulted prior to migrtion directly influences breeding success (Ankney & Mcinnes, 1978; Ebbinge & Spns, 1995; Mdsen, 2001; Prop & Blck, 1998). At the breeding sites, food vilbility nd qulity influence finl dult body size becuse of the impct on gosling growth rtes (Cooch, Lnk, Rockwell, & Cooke, 1991; Lrsson, Vn der Jeugd, Vn der Veen, & Forslund, 1998; Loonen, Oosterbeek, & Drent, 1997; Sedinger, Flint, & Lindberg, 1995). Fledgling weight lso influences post-fledgling survivl (Loonen et l., 1997; Vn der Jeugd & Lrsson, 1998) nd the probbility of breeding for individuls tht survive (Sedinger, Herzog, & Wrd, 2004). It is evident tht food vilbility nd qulity shpe fitness nd life history prmeters of these vin herbivores. However, the birds re fcing dilemm on their migrtion, becuse n increse in plnt biomss frequently corresponds with decline in food qulity (Lepge, Guthier, & Reed, 1998; Vn der Grf, Sthl, Bkker, & Drent, 2006). Severl studies suggest tht vin herbivores my not mximise intke s such, but insted mximise nutrient or nitrogen intke (Durnt, Fritz, & Duncn, 2004; Hssll, Riddington, & Helden, 2001; Kristinsen, Fox, & Nchmn, 2000; Lepge et l., 1998; Prop & Blck, 1998). Assuming tht protein intke is function of food intke nd tht food qulity declines with ge of the flush of growth of spring forge, birds hve two lterntives in order to obtin mximum protein intke: either forge on ptches with high biomss nd low qulity nd mximise totl intke or forge on ptches of vegettion with lower biomss but higher qulity. Given the digestive constrints on intke rte (Prop & Vulink, 1992; Sedinger & Rveling, 1988) nd time constrints of birds during migrtion, we hypothesise tht the ltter strtegy will be followed. In this study this hypothesis ws tested using the brncle goose, Brnt leucopsis. During their spring migrtion from stging sites in the Europen Wdden Se to Arctic breeding sites, brncle geese follow consecutive wves of fresh spring growth of their forge plnts s they migrte northwrds (Vn der Grf et l., 2006). At ech stopover site, rrivl is timed to mximise the profit from incresed spring production of the vegettion nd high initil nutritionl qulity of forge grsses. At the Arctic breeding sites, the pek in spring growth nd qulity ppers to coincide with gosling htch nd it is ssumed to fcilitte good growth conditions for goslings. We tested forge ptch choice of migrting brncle geese t three points long the
356 Est-Atlntic flywy of this species. Tissue qulity nd biomss of the min forge species were experimentlly mnipulted nd choice of forging ptches of migrting geese ws mesured through dropping counts nd supplementry behviourl observtions t one site. Methods Study sites The study ws conducted t three sites representing stging, stopover nd breeding res long the Est-Atlntic flywy of the brncle goose during spring 2003 nd 2004. The western-most study site is the islnd of Schiermonnikoog in the Dutch Wdden Se (53130 0 N, 6110 0 E). The islnd is used s winter nd spring stging site by up to 13,000 brncle geese (Bos & Sthl, 2003). The slt mrsh of the islnd consists of n ungrzed re nd n re tht is grzed by livestock from erly My until lte November. Both res re intensively used by wild geese prior to migrtion. We chose to conduct the experiment on the livestock-grzed slt mrsh, becuse vegettion structure here is more similr to tht of the other sites. On the ungrzed mrsh ccumultion of litter complictes the forging choices of the geese (Riddington, Hssll, & Lne, 1997; Summers & Critchley, 1990; Vn der Wl, Vn de Koppel, & Sgel, 1998). The geese leve this site towrds the end of April. The second study site is situted on the Swedish islnd of Gotlnd in the Bltic Se. Here, thousnds of geese use the nrrow bnds of costl slt mrshes nd djcent griculturl pstures s stopover site during period of bout 4 weeks in April nd My (Vn der Grf, Sthl, Veenekls, & Bkker, 2006). These mrshes re grzed by livestock from erly June until lte October. Our study site is slt mrsh on the peninsul Grötlingbo-udd, in the south-est of Gotlnd (57107 0 N, 18127 0 E). The third study site represents breeding site in Northern Russi, t the Kolokolkov By ner the bndoned villge of Tobsed (68135 0 N, 52120 0 E). Geese rrive here by the end of My nd strt breeding upon rrivl. Lrge moulting flocks gther in this re from mid-july onwrds, together with fmily birds. All geese leve the re t the end of September. This colony holds round 1500 breeding pirs with pek htch in erly July (Vn der Jeugd et l., 2003; Vn der Grf, Lvrinenko, Elskov, Vn Eerden, & Sthl, 2004). No livestock grzing is currently tking plce t this site, but the re hs been grzed by horses until round 1995. A.J. vn der Grf et l. On the first two sites the min forge plnt is red fescue, Festuc rubr, ccounting for, respectively, 90% (Vn der Wl et l., 1998) nd 45% (own dt bsed on epiderml counts of plnt remins in feces) of the goose diet. On the Russin site the diet consists of Crex subspthce nd Puccinelli phrygnodes in equl proportions (own dt). Experimentl set-up nd mesurements (Experiment 1) Plnt biomss ws mnipulted by plcing temporry grzing exclosures on slt-mrsh swrds, thereby creting difference in plnt biomss with continuously grzed, non-fenced control plots. At the sme time, forge qulity ws mnipulted by fertilising plots with commercil grnulr fertiliser (NPK, 12 10 18) resulting in n experimentl ddition of 10 g of nitrogen per m 2. We used the sme experimentl design t the three sites. At ll sites the vegettion ws in n erly (spring) growth stge. Additionlly we clculted the stnding crop of nitrogen, combintion of nitrogen content (per unit biomss) multiplied by the mount of biomss (stnding crop) nd expressed in g N m 2. Within the experimentl re t ech site, we creted six replicte, rndomly locted blocks, spced bout 50 m prt from ech other. Ech block contined four plots (4 m 4 m) representing the four tretments: nturlly grzed, ungrzed, fertilised nd ungrzed nd fertilised. A tretment ws rndomly ssigned to ech plot within block. Exclosures t the Wdden Se stging site, the Bltic Se stopover site nd the Russin breeding site, respectively, were set-up on 20 Mrch 2004, 4 April 2003 nd 1 July 2003. After 3 4 weeks exclosures were removed, forge biomss nd qulity were ssessed on ll plots nd ll droppings were removed from previously grzed plots (13 April 2004, 5 My 2003 nd 22 July 2003, respectively). Plots were now ccessible for wild brncle geese. This point in time is clled strt of the experiment throughout this pper nd coincided for ll res with pek utilistion by the geese. Goose dropping ccumultion ws mesured by dropping counts on ll plots, which llowed n ssessment of forging choices of the geese. Droppings were counted s soon s ll replicte blocks were visited by geese, in order to prevent differences being obscured by geese forging on less fvourite plots fter the fvourite plots hd been depleted. For the Wdden Se stging site this ws 7 dys, for the Bltic stopover site 5 dys nd for the Russin breeding site 4 dys. Dropping counts re muchused tool for determining grzing pressure of
Ptch choice of migrting herbivores 357 geese, differences in digestibility of grsses might influence the retention time in the gut nd thus the mount of droppings deposited per time unit (Hssll & Lne, 2005). Since our plots re very smll we expect tht retention time is the sme in ll plots, reflecting the verge digestibility of the grsses in the entire study site. Differences between sites, however, cn be cused by differences in verge digestibility; hence, we only mke comprisons within sites. At the strt of the experiment, the sttus of forge biomss ws determined from combined mesures of tiller density nd tiller weight of the min forge plnts. As destructive biomss smpling is connected with rther lrge mesuring errors in this type of short grss swrds (own experience), the following pproch ws dopted. In ech plot the number of tillers ws counted in 20 rndomly plced qudrtes, ech of 5.5 cm 5.5 cm. Tiller weight ws determined by clipping 50 individul tillers t ground level from ech plot, drying them for 48 h t 60 1C nd subsequently weighing them. In order to obtin representtive smple of tillers, point ws rndomly selected within ech qudrte nd ll tillers round this point were collected until smple of 50 tillers ws chieved. Forge biomss ws clculted by multiplying tiller density by men tiller weight. In order to determine forge qulity, smple of green lef tips ws collected t the sme moment. Smples were dried t 60 1C for 48 h, ground nd nlysed for nitrogen content, using CNHS-utomted element nlysis (Interscience EA 1110). Specific tests including behviourl observtions (Experiment 2) In My 2004 the experiment ws repeted with five replictes t the Bltic stopover site only, with the objective of ssessing forge ptch choice by the geese bsed on direct behviourl observtions, in ddition to dropping counts. Experiment 2 ws conducted close to the re used for Experiment 1 in 2003. Mesurements on biomss were conducted by cutting smll turf (10 cm 10 cm), removing ll bove-ground biomss nd sorting it into grsses nd other species. The mteril ws then dried t 60 1C for 48 h nd weighed. Biomss dt, therefore, comprise the weight of ll grsses present. This destructive method, s opposed to the detiled non-destructive method in Experiment 1, ws used becuse mixed swrd of severl species occurred in the re of Experiment 2. Forge qulity ws mesured following procedures used in 2003. Dropping counts were performed in the sme wy s in 2003. The choice experiment ws conducted in erly My 2004. Corners of ll experimentl plots were mrked by inconspicuous plstic tubes in order to be ble to distinguish plots from distnce. Observtions were mde from hide t distnce of bout 250 m, plced on slightly elevted position. During the first 5 dys fter the strt of the experiment, the number of geese present on ech plot of the different tretments ws noted every five minutes. In ddition, the number of ggressive encounters between geese on the plot nd other flock members ws recorded. For ech plot, goose presence ws clculted s the totl time geese were present on plot multiplied by the number of geese on the plot during the entire observtion period. Goose presence is expressed s goose minutes. Interctions re given s the number of interctions per goose observed for ech minute tht geese were present on the plot. As mesure of forge intke, peck rtes were collected for individul geese on ll tretments. Peck rtes were mesured s the time needed for 50 pecks; these dt were lter converted to pecks per minute. Sttisticl nlyses All dt were tested for normlity using Kolmogorov Smirnov test. When dt were not normlly distributed non-prmetricl Mnn Whitney U-test ws used. To test whether grzing exclusion nd fertilistion hd n effect on plnt biomss nd qulity, multivrite ANOVA with biomss nd qulity s dependent vribles nd exclosure nd fertilistion s fixed fctors ws used for ech site. To test whether grzing exclusion nd fertilistion influence dropping ccumultion, we used univrite ANOVA, with site, grzing exclusion nd fertilistion s fixed fctors. Additionlly, the sme test ws performed for ech site. Univrite ANOVAs with grzing exclusion nd fertilistion s fixed fctors were used for the vegettion nd observtionl dt of 2004. Additionlly, Tukey LSD ws performed to find differences in response of the geese to the individul tretments. All nlyses were performed using the sttisticl pckge SPSS for Windows, version 12.0.1. Results Experimentl mnipultions nd numericl response (Experiment 1) On ll sites exclusion of grzing incresed boveground biomss of forge plnts nd fertilistion
358 incresed nitrogen content of the lef tissue (Fig. 1, Tble 1). At the Russin breeding site there ws significnt interction between grzing exclusion nd fertilistion with respect to forge biomss, with the ungrzed nd fertilised combintion plot showing high biomss response s Figure 1. The effects of exclosure nd fertilistion tretments t ech study site on (A) lef tissue biomss (in g m 2 ), (B) qulity of forge species (%N) nd (C) stnding crop of nitrogen (in g N m 2 ), Experiment 1. Shown re mens+se (n ¼ 6). Test results re shown in Tble 1. A.J. vn der Grf et l. compred to solely ungrzed or fertilised plots. This interction is most likely cused by strong nutrient limittion of plnt growth in this re (Vn der Grf et l., 2004), inducing not only response of fertiliser tretment on plnt qulity but lso on plnt biomss. Becuse the tretments hd the required effect on the vegettion from here on they will be referred to s incresed qulity (fertilised), incresed biomss (exclosed) nd incresed qulity nd biomss (fertilised nd exclosed). When effects of grzing exclusion nd fertilistion on subsequent dropping ccumultion were tested for ll sites in comprison, strong threewy interction ws found between the two tretments nd site. This interction demonstrtes tht the tretments hd different effect on dropping ccumultion t ech site. Further nlyses were therefore performed for ech site seprtely (Tble 2). At ll sites long the flywy, the geese showed strong nd significnt response to the experimentl tretments. The significnt interction between incresed biomss nd incresed qulity for ll sites combined, nd for the Bltic nd Russin sites seprtely, shows tht this tretment is visited more thn expected bsed solely on the sum of effects of the two seprte tretments, which is mde visible in Fig. 2. It shows the response of the geese to the experimentl tretments. Vlues re expressed s the percentge of the totl cumultive dropping ccumultion per unit re in order to compre sites. All sttisticl tests, however, were performed bsed on the originl dt. Highest dropping ccumultion occurred on plots tht hd been exclosed nd fertilised (Fig. 2). At the Wdden Se stging site preference for plots with either incresed qulity only or incresed biomss only ws similr to the selection by the geese of nturlly grzed plots. On the Bltic site both incresed biomss only nd incresed qulity re preferentilly selected, while on the Russin site incresed biomss only is not selected bove the nturlly grzed sitution, in Tble 1. Test results for the effects of grzing exclusion nd fertilistion on tissue biomss nd qulity of forge species t the three study sites in Experiment 1 Effect of On Wdden Se stging site Bltic stopover site Russin breeding site F 1,20 P F 1,20 P F 1,20 P Exclosure Biomss 19.12 o0.001 13.39 o0.001 28.18 o0.001 Qulity 0.97 ¼ 0.337 3.11 ¼ 0.093 0.00 ¼ 0.985 Fertilistion Biomss 1.54 ¼ 0.229 2.89 ¼ 0.104 0.14 ¼ 0.710 Qulity 23.57 o0.001 20.44 o0.001 120.78 o0.001 With significnt interction of exclosure fertilistion on biomss: F1,20 ¼ 6.83, P ¼ 0.017. All other interctions were not significnt.
Ptch choice of migrting herbivores 359 Tble 2. Test results for the response of goose dropping ccumultion to experimentlly incresed biomss nd qulity in Experiment 1 Effect of All sites Wdden Se Stging site Bltic stopover site Russin breeding site F P F 1,20 P F 1,20 P F 1,20 P Site F 2;60 ¼ 3:47 ¼ 0.038 Incresed biomss F 1;60 ¼ 68:44 o0.001 16.87 ¼ 0.001 24.02 o0.001 30.93 o0.001 Incresed qulity F 1;60 ¼ 155:46 o0.001 20.50 o0.001 62.92 o0.001 93.15 o0.001 B Q F 1;60 ¼ 28:97 o0.001 2.09 ¼ 0.164 7.45 ¼ 0.013 26.05 o0.001 S B F 2;60 ¼ 1:88 ¼ 0.161 S Q F 2;60 ¼ 10:26 o0.001 S B Q F 2;60 ¼ 5:28 ¼ 0.008 A percentge of totl grzing intensity 70 60 50 40 30 20 10 0 B mid-april erly-my end-july b Wdden Se stging site b b c Bltic stopover site b c Russin breeding site nturlly grzed incresed qulity incresed biomss incresed biomss nd qulity Figure 2. (A) The flywy of the brncle goose projected from left to right, with the three study sites. (B) The effects of incresed biomss nd qulity on the reltive dropping ccumultion for ech study site, Experiment 1. Shown re men percentges+se (n ¼ 6). Different letters denote significntly different vlues, Po0:05 Tukey LSD test. contrst to plots of incresed qulity which re strongly preferred. Combining dt of ll sites, there were strong correltions between forge plnt biomss, forge qulity nd dropping ccumultion. For the Wdden Se stging site only correltion of dropping ccumultion with forge qulity ws observed, while for the Bltic stopover site there ws only correltion of dropping ccumultion with biomss (Tble 3). Finlly, t the Russin breeding site, dropping ccumultion ws correlted with both biomss nd qulity. In ddition, t ll three sites, there ws very strong correltion between dropping ccumultion nd the combined mesure
360 A.J. vn der Grf et l. Tble 3. Person correltion between dropping ccumultion nd forge plnt chrcteristics such s nitrogen content, biomss nd stnding crop of nitrogen in Experiment 1(n ¼ 24 for ech site, n ¼ 72 for ll sites combined) All sites Wdden Se stging site Bltic stopover site Russin breeding site Nitrogen content (%) R 0.54 0.64 0.36 0.68 Biomss (g m 2 ) R 0.33 0.31 0.72 0.46 N biomss (g N m 2 ) R 0.59 0.64 0.81 0.79 Pp0.05. Pp0.01. Pp0.001. Tble 4. Test results for the effects of grzing exclosure nd fertilistion on vegettion dt nd behviourl observtions t the Bltic stopover site in 2004, Experiment 2 Effects of: Biomss Qulity Dropping ccumultion Goose presence Aggressive interctions Peck rte F 1,16 F 1,16 F 1,16 F 1,16 U # F 1,205 Exclosure 51.00 15.96 10.13 61.89 100.0 90.54 Fertilistion 4.69 35.70 12.88 33.90 39.0 2.56 E F 4.85 0.36 7.10 13.67 0.14 # Dt re not normlly distributed; non-prmetric Mnn Whitney U-test ws used. Pp0:05. Pp0:01. Pp0:001. of stnding crop of nitrogen, which explins 35 65% of the observed vrition in dropping ccumultion (Tble 3). Behviourl response (Experiment 2) In 2004 t Gotlnd, grzing exclusion significntly incresed grss biomss (Nturlly grzed, ungrzed, fertilised, ungrzed nd fertilised, respectively: 28.174.5 A, 75.176.9 B, 27.876.5 A nd 116.6715.9 C g dwt m 2 ; different letters denote significnt differences, post-hoc Tukey test) nd experimentl fertilistion incresed grss qulity (nturlly grzed, ungrzed, fertilised, ungrzed nd fertilised, respectively: 2.870.1 AB, 2.170.1 A, 4.070.3 C nd 3.270.2 B cm) (Tble 4). Dropping counts 1 week fter strt of the choice experiment gve similr results in 2003 nd 2004 t this site (Figs. 2 nd 3); plots with combined increse of biomss nd qulity were preferred bove ll other tretments. The dt from 2004 revel nonsignificnt trend of greter dropping ccumultion on plots with either incresed qulity or biomss, s compred to the nturlly grzed plots. The behviourl observtions indicte tht geese preferred the tretment with combined high qulity nd high biomss (Fig. 3). Here, goose presence ws higher thn on ll other tretments nd more ggressive interctions were observed (Tble 4). Both goose presence nd the number of ggressive interctions lso ws higher on the plots with high biomss compred to nturlly grzed plots. There were no differences detected in these two prmeters for plots with incresed qulity. Peck rte ws lower on plots with incresed biomss. Discussion Forge qulity or biomss? The gret preference of geese for combintion of plnt qulity nd biomss s seen in Figs. 2 nd 3, is no surprise s these plots nturlly provide forge with the highest stnding crop of nitrogen. In nturl, non-experimentl sitution, however, there is lwys trde-off between forge qulity nd quntity: with incresing biomss, plnt qulity decreses (Lepge et l., 1998; Vn der Grf et l., 2006) nd digestive constrints limit n incresed intke of plnt mteril to compenste for low forge qulity (Prop & Vulink et l., 1992). In this pper we, therefore, imed to investigte the ptch choice of wild geese, uncoupling biomss nd qulity by fertilistion nd exclosure tretments. Idelly, mesure of ctul intke rte would be needed to unrvel whether
Ptch choice of migrting herbivores 361 Figure 3. Behviourl observtions on the use of experimentl plots on the Bltic stopover site in 2004, Experiment 2. Shown re men percentges+se (n ¼ 5). Different letters denote significnt different vlues, Po0:05 Tukey LSD test. For sttistics see Tble 4. nutrient or dry mtter intke is mximised. Becuse intke rte is difficult mesure to obtin, we circumvent this problem using tretments tht either increse nutrient content (g N/g biomss) nd keep biomss constnt or increse biomss nd keep nutrient content constnt. At ll sites long the goose flywy we witness similr trend (though only significnt t the Russin site); plots with incresed qulity nd verge biomss re preferred by the geese bove verge qulity nd high biomss (Fig. 2, Tble 2). It is striking tht on ech site the order of preference for the tretments is consistent (Fig. 2). Moreover, t ech site ptch choice of the geese mirrors more closely the order of stnding crop of nitrogen thn tht of biomss (Fig. 1), which is lso shown in our correltion nlyses where stnding crop of nitrogen consistently bsorbs higher proportion of the vrince in dropping ccumultion rtes thn biomss. These findings re consistent with our hypothesis tht selecting high-qulity plnts reduces the time necessry to collect sufficient body reserves nd thus potentilly gives the birds the dvntge of rriving erlier on the breeding sites. For the sme flywy, we hve shown tht brncle geese pper to follow peks of stnding crop of nitrogen long their spring migrtion route, the so-clled Green wve (Vn der Grf et l., 2006). In this wy the geese mximise the ccumultion of body reserves during migrtion, while on the breeding sites the chicks cn profit from the pek in stnding crop of nitrogen. Differences in site choice long the flywy Additionlly, the interctions between stging site nd plnt nitrogen content nd forge biomss nd the interction between these mesures (Tble 2) indictes tht ptch choice differs between stging sites. We suggest tht this is cused by differing limiting resources long the migrtion route. At the Russin breeding site overll levels of biomss re high due to low levels of goose grzing, but forge qulity is low (this
362 study, Vn der Grf et l., 2004). Here, forge qulity is the choice criterion nd contributes most to the ptch choice of the geese (Tble 3, Fig. 2). At the Bltic stopover sites where high grzing intensities during spring migrtion result in comprtively low levels of biomss (Vn der Grf et l., 2006) nd where overll tissue qulity is high (this study), stnding crop of nitrogen is highest on previously ungrzed plots with high stnding biomss. Here, brncle geese ggressively monopolise food ptches with incresed biomss s our behviourl dt showed (Fig. 3, Tble 4). Therefore, lso plots with incresed biomss only re slightly preferred bout control plots in this re (Fig. 2). In both these res, the interction between incresed biomss nd incresed qulity (Tble 2) shows tht the rection of the geese towrds combintion of these two prmeters is significntly stronger thn cn be expected from simple ddition of the two single effects. At the spring stging grounds in the Wdden Se neither biomss nor qulity per se re limiting on the livestock-grzed slt mrsh nd probbly rised nutrient intke cn be chieved by selecting plots with incresed levels of either of these two prmeters. In this re the preference for the combintion of the two effects is not more thn would be expected from n ddition of the two single effects (Tble 2, Fig. 2). The bottlenecks in stnding crop of nitrogen for these herbivores thus pper to lie in the limited biomss vilbility t the Bltic stopover site nd the limited nutrient content of food in the Arctic breeding site. Acknowledgements We re grteful to Vereniging Ntuurmonumenten Schiermonnikoog, Länsstyrelsen Gotlnd nd the Russin uthorities who kindly provided permissions to conduct our work. We thnk Mennobrt vn Eerden nd Kjell Lrsson for logistic support. Jn Bkker, Joost Tinbergen nd Henk vn der Jeugd joined stimulting discussions. Criticl comments of Bob Jefferies, Mrk Hssll nd nonymous referees on erlier versions gretly improved the qulity of our mnuscript. Bert Venem nd Nellie Eck nlysed plnt smples. The project ws supported by the Deutsche Forschungsgemeinschft (JS), the Europen Science Foundtion in BIRD trvel grnts to AJvdG nd GFV, the Dutch Institute for Inlnd Wter Mngement nd Wste Wter Tretment (RIZA), the Netherlnds Arctic Progrmme NAP nd the University of Groningen. References A.J. vn der Grf et l. Ankney, C. D., & Mcinnes, C. D. (1978). Nutrient reserves nd reproductive performnce of femle lesser snow geese. Auk, 95, 459 471. Bety, J., Guthier, G., & Giroux, J. F. (2003). Body condition, migrtion, nd timing of reproduction in snow geese: A test of the condition-dependent model of optiml clutch size. Americn Nturlist, 162, 110 121. Bos, D., & Sthl, J. (2003). Creting new forging opportunities for Drk-bellied Brent Brnt bernicl nd brncle geese Brnt leucopsis in spring Insights from lrge-scle experiment. Arde, 91, 153 166. Cooch, E. G., Lnk, D. B., Rockwell, R. F., & Cooke, F. (1991). Long-term decline in body size in snow goose popultion: Evidence of environmentl degrdtion? Journl of Animl Ecology, 60, 483 496. Drent, R. H., Both, C., Green, M., Mdsen, J., & Piersm, T. (2003). Py-offs nd penlties of competing migrtory schedules. Oikos, 103, 274 292. Durnt, D., Fritz, H., & Duncn, P. (2004). Feeding ptch selection by herbivorous Antide: The influence of body size, nd of plnt quntity nd qulity. Journl of Avin Biology, 35, 144 152. Ebbinge, B. S., & Spns, B. (1995). The importnce of body reserves ccumulted in spring stging res in the temperte zone for breeding in Drk-bellied Brent Geese Brnt b. bernicl in the high Arctic. Journl of Avin Biology, 26, 105 113. Eichhorn, G., Afnsyev, V., Drent, R. H., & Vn der Jeugd, H. P. (2006). Spring stopover routines in Russin brncle geese Brnt leucopsis trcked by resightings nd geoloction, Arde, 94(1), in press. Green, M., Alerstm, T., Clusen, P., Drent, R. H., & Ebbinge, B. S. (2002). Site use by drk-bellied brent geese Brnt bernicl bernicl on the Russin tundr s recorded by stellite telemetry: implictions for Est Atlntic Flywy conservtion. Wildlife Biology, 8, 229 239. Hssll, M., & Lne, S. J. (2005). Prtil feeding preferences nd the profitbility of winter-feeding sites for brent geese. Bsic nd Applied Ecology, 6, 559 570. Hssll, M., Riddington, R., & Helden, A. (2001). Forging behviour of brent geese, Brnt b. bernicl, on grsslnds: Effects of swrd length nd nitrogen content. Oecologi, 127, 97 104. Kristinsen, J. N., Fox, A. D., & Nchmn, G. (2000). Does size mtter? Mximising nutrient nd biomss intke by shoot size selection mongst herbivorous geese. Arde, 88, 119 125. Lrsson, K., Vn der Jeugd, H. P., Vn der Veen, I. T., & Forslund, P. (1998). Body size declines despite positive directionl selection on heritble size trits in brncle goose popultion. Evolution, 52, 1169 1184. Lepge, D., Guthier, G., & Reed, A. (1998). Sesonl vrition in growth of greter snow goose goslings: The role of food supply. Oecologi, 114, 226 235.
Ptch choice of migrting herbivores 363 Loonen, M. J. J. E., Oosterbeek, K., & Drent, R. H. (1997). Vrition in growth of young nd dult size in brncle geese Brnt leucopsis: Evidence for density dependence. Arde, 85, 177 192. Mdsen, J. (2001). Spring migrtion strtegies in Pink- Footed Geese Anser brchyrhynchus nd consequences for spring fttening nd fecundity. Arde, 89, 43 55. Nolet, B. A., Andreev, V. A., Clusen, P., Poot, M. J. M., & Wessel, E. G. J. (2001). Significnce of the White Se s stopover for Bewick s Swns Cygnus columbinus bewickii in spring. Ibis, 143, 63 71. Prop, J., & Blck, J. M. (1998). Food intke, body reserves nd reproductive success of brncle geese Brnt leucopsis stging in different hbitts. Norsk Polrinstitutt Skrifter, 200, 175 193. Prop, J., & Vulink, T. (1992). Digestion by brncle geese in the nnul cycle: The interply between retention time nd food qulity. Functionl Ecology, 6, 180 189. Riddington, R., Hssll, M., & Lne, S. J. (1997). The selection of grss swrds by brent geese Brnt b. bernicl: Interctions between food qulity nd quntity. Biologicl Conservtion, 81, 153 160. Sedinger, J. S., Flint, P. L., & Lindberg, M. S. (1995). Environmentl influence on life-history trits: Growth, survivl nd fecundity in blck brnt (Brnt bernicl). Ecology, 76, 2404 2414. Sedinger, J. S., Herzog, M. P., & Wrd, D. H. (2004). Erly environment nd recruitment of blck brnt (Brnt bernicl nigricns) into the breeding popultion. Auk, 121, 68 73. Sedinger, J. S., & Rveling, D. G. (1988). Forging behvior of cckling Cnd goose goslings: Implictions for the roles of food vilbility nd processing rte. Oecologi, 75, 119 124. Summers, R. W., & Critchley, C. N. R. (1990). Use of grsslnd nd field selection by brent geese Brnt b. bernicl. Journl of Applied Ecology, 27, 834 846. Vn der Grf, A. J., Lvrinenko, O. V., Elskov, V., Vn Eerden, M. R., & Sthl, J. (2004). Hbitt use of brncle geese t subrctic slt mrsh in the Kolokolkov By, Russi. Polr Biology, 27, 651 660. Vn der Grf, A. J., Sthl, J., Bkker, J. P., & Drent, R. H. (2006). Surfing on green wve How plnt growth drives spring migrtion in the brncle goose, Arde, 94(1) in press. Vn der Grf, A.J., Sthl, J., Veenekls, R.M., & Bkker, J.P. (2006). Vegettion chrcteristics of brckish mrsh on Gotlnd nd forging choices of migrting nd brood rering geese. Annles Botnici Fennici, in press. Vn der Jeugd, H. P., Gurtovy, E., Eichhorn, G., Litvin, K. Ye., Mineev, O. Y., & Vn Eerden, M. R. (2003). Breeding brncle geese in Kolokolkov By, Russi: Number of breeding pirs, reproductive success, nd morphology. Polr Biology, 26, 700 706. Vn der Jeugd, H. P., & Lrsson, K. (1998). Pre-breeding survivl of brncle geese Brnt leucopsis in reltion to fledgling chrcteristics. Journl of Animl Ecology, 67, 953 966. Vn der Wl, R., Vn de Koppel, J., & Sgel, M. (1998). On the reltion between herbivore forging efficiency nd plnt stnding crop: An experiment with brncle geese. Oikos, 82, 123 130. Wrd, D. H., Reed, A., Sedinger, J. S., Blck, J. M., Derksen, D. V., & Ctelli, P. M. (2005). North Americn Brnt: Effects of chnges in hbitt nd climte on popultion dynmics. Globl Chnge Biology, 11, 869 880.