Long-Distance Dispersal in the Barnacle Goose Branta Leucopsis Based on Individual Case Histories

Home , Barnacle goose, Brant (goose)

Travels and traditions: long-distance dispersal in the Barnacle Goose Branta leucopsis based on individual case histories

Henk P. van der Jeugd1,2,* & Konstantin Y. Litvin3

van der Jeugd H.P. & Litvin K.Y. 2006. Travels and traditions: long-distance dispersal in the Barnacle Goose Branta leucopsis based on individual case histories. Ardea 94(3): 421–432.

Long-distance dispersal can have far-reaching consequences for the amount of gene flow between populations. However, individuals usually disperse to places well outside the study area, which makes long-distance dispersal hard to study. In this study, the fate of c. 4500 individually marked Barnacle Geese Branta leucopsis hatched on the Swedish island of Gotland between 1984 and 2002 was followed. Only 18 dispersed to other Barnacle Goose populations along the flyway, 16 males and two females. Most long-distance dispersers were very large compared to other cohort members, and were observed to become independent of parents and siblings early in life. It is estimated that about 6.6% of all juvenile Barnacle Geese disperse over long-distances, but at most 11% of these long-distance dispersers were actually detected. Most dispersed to the large Russian population, where several birds were observed breeding and/or retrapped during fieldwork between 2002 and 2004. One male dispersed from Gotland to the Svalbard population. Breeding was only proven in three out of the 18 dispersers, and only one is known to have successfully fledged young. These results confirm that many years of intensive studies at different localities are needed to define long-distance dispersal rates.

Key words: body size, Branta leucopsis, dispersal, gene-flow, individual, personality

1Animal Ecology Group, Centre for Ecological and Evolutionary Studies, University of Groningen, P.O. Box 14, 9750 AA Haren, The Netherlands; 2SOVON Dutch Centre for Field Ornithology, Rijksstraatweg 178, 6573 DG Beek-Ubbergen, The Netherlands; 3Bird Ringing Centre, Leninsky prospekt 86–310, Moscow 119313, Russia; *corresponding author ([email protected])

INTRODUCTION modern evolutionary ecology. The phenomenon of individuals that are sighted outside their normal Dispersal is a process that influences population range has inspired research for at least a century dynamics, mediates gene flow between popula- and has supported a growing body of literature tions, determines the rate of colonisation of new (see Clobert et al. 2001 for a review). However, habitats and is a major component in the process much has still to be learned as the study of disper- of speciation. It therefore lies at the heart of sal is beset with many practical problems. Natal 422 ARDEA 94(3), 2006

dispersal, the movement of individuals from their 2001) and Great Tits Parus major (Dingemanse et natal site to the site of first breeding, is especially al. 2003) is associated with dispersal. difficult to study because dispersing individuals Here we investigate the phenotypic correlates are not hindered by the boundaries of the study of long-distance dispersal in Barnacle Geese area and potentially can settle anywhere where Branta leucopsis hatched and ringed in two colo- suitable habitat exists (van Noordwijk 1995). This nies in the Baltic and resighted or recovered far inevitably underestimates long-distance dispersal, outside their natal area. Colonial breeding birds although the introduction of genetic techniques like Barnacle Geese are highly suitable for the has partly overcome these problems. However study of dispersal, since colonies are clearly sepa- rare, long-distance dispersal potentially has far- rated, often conspicuous, and comparatively easily reaching consequences as it can inhibit genetic monitored, so that dispersers are identified unam- divergence between seemingly isolated popula- biguously (Pradel 1996). tions. Practical problems aside, long-distance dispersal probably remains rare in most species and the distribution of distances between natal and first breeding sites is typically skewed, with many METHODS individuals settling in or very close to the natal site, and few individuals dispersing long distances. Of the three East Atlantic flyway populations of For those staying ‘at home’, there are the potential the Barnacle Goose, the Russian-based population benefits of interacting with kin, while long-dis- is by far the largest, with a nesting range tradition- tance dispersers might benefit from new and bet- ally centred on Novaya Zemlya and Vaigach, and ter opportunities or (re-)colonise new areas. There wintering on the European mainland (mainly in are also costs associated with both philopatry and Germany and The Netherlands). Within three dispersal. Individuals remaining ‘at home’ face the decades, this population has undergone a dramatic risk of inbreeding depression, whereas individuals change in breeding range and numbers through dispersing far might not be well adapted to their colonisation of new breeding areas to the south- new environments. west of their original breeding grounds (Ganter et The relative costs and benefits of dispersal dif- al. 1999). In 1971, the first breeding pair was fer between individuals of differing quality, and found in the Baltic (Larsson et al. 1988), and this this determines which individuals stay ‘at home’ population has since grown rapidly (Larsson & van and which disperse (e.g. Greenwood 1980, der Jeugd 1998; Fig. 1). More recently, a breeding Verhulst et al. 1997, van der Jeugd 2001, Clobert population was established in the southwest of et al. 2001). Long-distance dispersal might require The Netherlands, also growing rapidly (Ouweneel special attributes or even morphological adapta- 2002, Pouw et al. 2005, van der Graaf et al. 2006; tions to be successful. Indeed, in Naked Mole-rats Fig. 1). New populations have adopted consider- Heterocephalus glaber the probability of death dur- ably shorter migration distances or become seden- ing dispersal is extremely high, but recently a spe- tary. Southern populations breed six to seven cial dispersal morph has been discovered amongst weeks earlier than populations in the north. Three captive-bred Mole-rats (O’Riain et al. 1996). This populations, (Barents Sea, Baltic and North Sea) morph is behaviourally as well as morphologically are now recognised, although information on the adapted to long-distance dispersal between amount of exchange between these populations is colonies. In insects, special life-history stages currently lacking. There is evidence that, although adapted to dispersal, usually recognised by the winter ranges overlap, habitat choice on a smaller development of wings, are common (Zera 2003). scale differs, i.e. even at the same staging site a There is evidence that exploratory behaviour in subtle form of segregation prevails (van der Jeugd Common Lizards Lacerta vivipara (Clobert et al. et al. 2001). van der Jeugd & Litvin: DISPERSAL IN BARNACLE GEESE 423

1000000

100000 RUS

10000

1000

100 SE population size SE 10 NL

1 NL 1960 1970 1980 1990 2000

Figure 1. Population trend of the East Atlantic Flyway population of Barnacle Geese (filled triangles) and number of breeding pairs in two recently established populations in the Baltic (SE) and The Netherlands (NL). The location of these populations, as well as the third study population mentioned in the text in northern Russia (RUS), is shown in the map.

Smaller numbers of Barnacle Geese also breed niles. Each year until 2000 between 77 and 245 in Greenland and Svalbard. Birds from these popu- juveniles have been ringed with colour rings. In lations winter in Northern Ireland and on the addition to these, up to 154 fledglings were ringed island of Islay, and in South-western Scotland, annually with metal rings only. From 2001 respectively. Their wintering ranges do not overlap onwards, ringing was performed using metal rings with those of the Barents Sea, Baltic and North only, and between 163 and 420 fledglings have Sea populations. Small numbers of individuals been ringed annually until 2003. In total, 2662 from the Svalbard population winter on the fledglings were ringed with colour rings and European mainland however, based on resightings another 1984 with metal rings only. Taken of ringed birds (Black et al. 2007). Exchange of together, more than 4500 birds were available for individuals between the wintering areas of the detection at breeding localities away from their Greenland and Svalbard populations has also been natal colony. observed (Black et al. 2007). All birds were measured and weighed upon In this study, we only consider birds hatched ringing when they were five to eight weeks of age, and ringed in the Baltic population. Ringing com- just prior to fledging. Sex was determined by cloa- menced here in 1984 and continues to the present. cal examination (Owen 1980). Body mass, tarsus Flocks of adult and juvenile Barnacle Geese were length and head length of captured birds were captured by round-ups at up to five different measured to the nearest 25 g, 0.1 mm and 1 mm, moulting localities on Gotland and Öland in the respectively (see Larsson et al. 1998 for details on Swedish Baltic in mid-July each year. Captured measuring techniques). Tarsus length and head birds were individually marked with metal rings, length have reached their final size when birds are usually in combination with engraved coloured one year old. Tarsus length, head length and body plastic leg rings, which allowed individual identifi- mass of juveniles are affected by genetic factors cation up to a distance of a few hundred meters. (Larsson & Forslund 1992, Larsson 1993, Cooke et Among captured birds, adults, defined as birds one al. 1995) but, to a considerable extent, also by the year old or older, were distinguished from juve- quantity and quality of the food eaten during 424 ARDEA 94(3), 2006

growth (Cooch et al. 1991a,b, Larsson & Forslund this bird had dispersed to Svalbard. Two more 1991, Sedinger & Flint 1991, Lindholm et al. birds observed in Scotland had to be excluded 1994). Body size measures therefore vary both because they could not be identified at individual within and among years and mean values of level. In only a few cases is good evidence of cohorts can be used as an index of the availability proven breeding available. of high quality food and other environmental con- Body size and body mass at fledging have pre- ditions to which the juveniles have been exposed viously been found to affect post-fledging survival after hatching (Loonen et al. 1997, Larsson et al. in both males and females (van der Jeugd & 1998). Here, tarsus length is used as an index of Larsson 1998). The probability of natal dispersal the overall size of individual juveniles. Because of increased with increasing mass and tarsus length sexual size dimorphism, tarsus length of juvenile in males, but not in females (van der Jeugd 2001). males was standardised to female equivalents In the latter analysis, birds were classified as hav- before further analyses by multiplying male tarsus ing dispersed when they were known to be alive at length with a constant (0.938) reflecting the the age of five years – the maximum age of first female/male ratio observed in the total sample of breeding – but had not recruited into their natal juveniles of known sex where tarsus length had colony. However, the resighting rate on the winter- been measured (n = 4344). After this procedure ing grounds has decreased in later years because no difference in mean or variance between the the number of observations received did not keep sexes was detected. To correct for differences in up with the increasing amount of ringed birds. the overall size and mass of juveniles between Therefore it was not meaningful to select only years and colonies, relative standardized tarsus birds that had survived up to a certain age for the length was calculated by subtracting colony and analyses in this paper because too many birds year specific means from individual values. would have been classified as being dead. Observations from the public, birds shot by hunters, and resightings and recaptures by the authors and other workers in Barnacle Goose RESULTS colonies were used to trace birds that had dispersed to places far from their natal area. Out of 4646 birds that were ringed, 18 satisfied the Sampling was especially intensive in one of the criteria outlined in the methods (Table 1). Only recently established Barnacle Goose populations three out of these 18 birds were observed breeding along the Barents Sea coast at the mouth of the (nest with eggs or a brood), in 14 cases breeding Kolokolkova Bay (68°21'N, 52°12'E; Fig. 1). During was likely because birds were observed in suitable July 2002, c. 4000 adult birds were captured dur- breeding habitat during the breeding season. In ing moult (van der Jeugd et al. 2003). In 2003 and one case probable breeding was inferred because 2004, some 2000 breeding pairs were monitored the bird wintered in Scotland during two years closely each year and 2950 and 3775 adult birds among birds belonging to the Svalbard population. were captured during moult, respectively. All Of the 18 birds, 16 were males and two were ringed birds that were observed outside the nor- females. Although ringing in the Baltic started in mal Baltic breeding range during the breeding sea- 1984, most birds were hatched recently, 14 after son (May–August) when at least two years of age, 1995, and only 4 before that year, probably due to i.e. when capable of breeding, were considered to increased search effort at potential breeding areas have bred or at least to have attempted to breed in in recent years. For example, observations and the area where they were observed. In addition, catching of Barnacle Geese at Kolokolkova bay, we added one bird that was observed among win- Russia started in 2002 and has continued annually tering Svalbard Barnacle Geese in Scotland during since. Six out of the 18 long-distance dispersers two consecutive winters, since it was likely that were detected here (Table 1). van der Jeugd & Litvin: DISPERSAL IN BARNACLE GEESE 425 (km) dish Baltic (57°17'N, to fledging compared to the to fledging compared

Characteristics of 18 long-distance dispersers among 4500 Barnacle Geese ringed as juvenile on the island Gotland in Swe CW2 M 1997 23 (184) 0 2003–2004 Russia, Bay, Kolokolkova Breeding 68°34'N, 52°22'E 2075 able 1. other members of its cohort. 9164473WDGE MBKGHBHGE M 2000WFGP MWKR2 2000 1 (131) MR5B- 1996 FY 2 (150) ? 1996 M9164432 5 (218) 2000B5G9 M 0 7 (218) 1988 MW=G= F 11 (150) 1993BEGY 0 14 (125) 2000 May 2004 M Russia, Bay, 9167052 Shot Kolokolkova M M 19 (197) M, FW=G1 Hellegatsplaten, Netherlands, 2004 breeding Probably 30 (131) 1995 M FBPWH 1999 F F, Netherlands, 1991 Summer Oostvaardersplassen, Svalbard Scotland, UK, 1999–2001, breeding Winters 9164493 ? M 31 (200) July 2002 dead Mezen, Russia, Found 2001 51°42'N, 04°22'E 1996G6RK 32 (157) July 2002 Russia, Bay, Caught Kolokolkova M M9166685 0 15°37'E 78°13'N, 1999 M, F 36 (267) 34 (110) 68°35'N, 52°18'E Hellegatsplaten, Netherlands, 2004 breeding Probably 52°26'N, 05°19'E 1116 1991 2000 M M 38 (157) ? ? Observed summer 2002 Southern Finland, 2335 48 (126) 51°42'N, 04°22'E 2074 1994 48 (131) 1014 0 2001 May 2002 Shot Kanin, Russia, 2001 Russia, Shot Kolguev, 68°36'N, 52°27'E 0 ? 60 (181) 65°51'N, 44°14'E 77 (267) 1116 Observed summer 2003 southern Finland, May 1999 Shot Shoina, Russia, F 2080 ? 2004 1634 Russia, Bay, Kolokolkova Breeding 60°24'N, 22°12'E 2003 August Russia, Bay, Caught Kolokolkova May 2004 Russia, Bay, Shot Kolokolkova 401 ’t Zwin, Belgium, 1996–2002 Breeding Observed summer 2003 southern Finland, 60°24'N, 22°12'E 68°00'N, 44°30'E 68°36'N, 52°27'E 69°00'N, 49°00'E 68°34'N, 52°22'E 401 67°55'N, 44°19'E 1756 68°35'N, 52°18'E 2080 1971 2075 60°24'N, 22°12'E 1741 2074 51°22'N, 03°21'E 401 1194 Code Sex Hatched rank Tarsus Siblings Evidence for long-distance dispersal Coordinates Distance T 18°45'E) between 1988 and 2001. Tarsus rank refers to the rank of the bird according to its tarsus length at capture just prior to its tarsus length at capture according to the rank of bird rank refers 18°45'E) between 1988 and 2001. Tarsus 426 ARDEA 94(3), 2006

The total number of Gotland immigrants in 30 long-distance dispersers other populations is a function of the proportion of other birds juveniles actually dispersing over long distances, 25 and their annual survival. Van der Jeugd (2001) identified 231 dispersers among a sample of 1302 20 individuals that were hatched and ringed on Gotland between 1984 and 1993. Of these, 59 15 were found breeding at other localities in the Baltic. The rest (74%) was never observed during 10

summer but they were inferred to have dispersed relative frequency (%) as they never returned to their natal colony yet 5 were known to be alive from winter observations at an age of at least five years, when Barnacle 0 Ð16 Ð12 Ð8 Ð4 0 4 8 12 Geese normally would breed. Assuming that few relative fledging tarsus length (mm) or no birds that dispersed to other areas within the Baltic were missed, this would mean that 172/ Figure 2. Frequency distribution of the relative tarsus 1302 = 13.2% had dispersed to areas further length at fledging of long-distance dispersers, n = 18) away. Applying that figure to the 4500 birds that and other birds ( n = 3693). were the subject of this study, and taking an annual adult survival of 95% into account (Lars- son et al. 1998), c. 330 Gotland dispersers should captured immigrants is 58–182, which is of similar have been present in 2003 in populations outside order of magnitude. the Baltic. Because the Russian population is by far Although the sample size is too small to allow the largest, most of these birds should be attempt- for rigorous analyses, one thing stands out: most ing to breed somewhere in northern Russia. Under of the 18 long-distance dispersers were large or less stringent assumptions, for example when half very large compared to the other members of their of the dispersers in the study by van der Jeugd cohorts (Fig. 2). Seven out of 18 birds were among (2001) that settled in the Baltic were actually the largest ten percent of their cohort, and the four missed, the proportion dispersing decreases to largest birds (all males) ranked number 1, 2, 5 and 6.6%, and the total number of dispersers to 170. 7 in their cohorts, which ranged in size from 131 Only 18 birds were actually found, a detection to 218 birds (Table 1). Before calculating these probability of 5.5 or at most 11%. ranks, male body size was standardized to female During three moult catches at Kolokolkova Bay equivalents to control for sexual size dimorphism in 2002, 2003, and 2004 a total of 10 725 adult (see methods). In an analysis where differences in Barnacle Geese were captured. Among these were mean size between cohorts were controlled for by three Gotland-hatched immigrants. Given that the calculating relative values, tarsus length of long- Russian Barnacle Goose population numbered distance dispersers was clearly larger than that of about 370 000 individuals in 2003 this means 103 other birds (t18,3693 = 4.16; P < 0.001). ringed Gotland immigrants must have been present in the entire Russian population at that time. The 18 individual cases are briefly described This is less then the 170–330 calculated above, but below. given that Arctic-breeding Barnacle Geese have lower survival (Ebbinge et al. 1991) that number 9164432 Hatched in 2000 at the island Skenholmen, male, unknown reduces to c. 110–230. According to Pearson & hatch date, with a tarsus length of 87.3 mm it ranked num- Hartley (1962) the 95% confidence interval of our ber 30 out of 131 fledglings ringed there that year. Because estimate from Kolokolkova Bay based on three he lacked colour rings there were no observations until he van der Jeugd & Litvin: DISPERSAL IN BARNACLE GEESE 427

was seen (ring number read) at various locations near 21 July without signs of breeding (but few birds bred that Turku in southern Finland in August 2002. year). She was never caught or seen again, and probably bred elsewhere in the area. 9164473 Like the previous bird, also hatched in 2000 at WKR2 Skenholmen, male, unknown hatch date, with a tarsus of Hatched at Laus holmar 22 May 1988, male. With a tarsus 93.0 mm one of the largest fledglings we ever caught, and of 88.6 mm the largest of four siblings (three brothers and the largest of the 131 members of his cohort. Because he a sister) that made it to fledging. All four were seen during had no colour rings we have no observations before he was their first autumn without their parents, three of them, shot at Kolokolkova Bay on 26 May 2004. WKR2, a brother and the sister together, the other brother alone. None of the brothers ever returned to Gotland, WDGE although observations from the wintering grounds showed Hatched 2000 at Laus holmar, male, unknown hatch date, they were alive. The sister recruited into her natal colony at the only surviving son of two ringed parents. A tarsus of the age of three years and still breeds there to the present. 90.0, the second largest fledgling of his cohort (n = 150). WKR2 attempted breeding at Oostvaardersplassen where Seen on 10 March 2001 at the Frisian coast in The he was seen on 10 June 1991. Netherlands with his parents, but never seen thereafter until May 2004 when observed in suitable breeding habitat R5B- at Hellegatsplaten, The Netherlands. Hatched on 24 may 1993 at Laus holmar. Male, ringed parents and two ringed sisters. Its tarsus length of 85.1 mm is BKGH well above average, rank 19 out of 197. Parents were very Hatched 31 May 1996, at Laus holmar, male, one surviving faithful to their wintering site in German Schleswig- sister, two ringed parents. Tarsus length of 86.7, fifth Holstein, where they were observed more than one hun- among 218 fledglings in his cohort. Seen 6 February 1997 dred times. They never travelled to The Netherlands. in northwestern Germany without his parents, which were During the winter of 1993 they went to Germany again, seen almost simultaneously 200 km to the southwest. accompanied by their two daughters. Son R5B- however Parents always wintered in Germany, but BKGH was seen a did not accompany them and was alone right from the start few times during winter in The Netherlands, and then of that autumn and went straight to the southwest of The turned up in northern Scotland in December 1999, to be Netherlands. He went back there during subsequent win- seen again at Eastpark, Caerlaverock, Scotland in January ters. On 10 April 2004 he was observed among the local 2001 among wintering birds from the Svalbard population breeders at Hellegatsplaten, The Netherlands, apparently with an unringed female. Presumably he had migrated to preparing for breeding. His two sisters both returned to Svalbard during summer 1999 or 2000. their natal colony on Gotland and breed at less than 100 m from their parents and each other. BHGE Hatched also in 1996 at Laus holmar, male, unknown hatch YCW2 date, ringed one day before the previous bird, the only sur- Hatched 24 May 1997 at Laus holmar, male. Tarsus length viving son of two ringed parents. A tarsus length of 86.5 85.0 mm (23/184). The only surviving fledgling of marked mm was seventh largest among the 218 fledglings in his parents. After fledging it was never seen again, while the cohort. No observations of parents or son in the first winter, parents were seen regularly without young. It was caught but the second winter there is one observation of BHGE in six years later, on 6 August 2003 at Kolokolkova Bay, and Friesland, The Netherlands. On 5 July 2002, in the middle seen shortly after with two fledged young and a marked of the breeding season, he was found dead near Mezen at female. One of the three birds out of the 18 that we know the eastern coast of the White Sea in northern Russia. are breeding in their new surroundings, and as yet the only one that also produced fledged young. WFGP Astonishingly, the whole family was seen again on 24 One of two females among the 18 long-distance dispersers. April 2004 on Gotland among Gotland breeding birds, and Hatched at Laus holmar as early as 17 May 2000. Fledged less than 1 km away from the place where the male was together with one brother. With a tarsus length of 80.1 mm raised 7 years before. Russian birds are not seen in these number 11 among 150 fledged young. Parents both ringed. parts of Gotland generally, and most birds from the colony Although not observed together, parents and daughter at Kolokolkova Bay do not visit Gotland during their spring spent the first three winters in the same area in the south- migration (Eichhorn et al. 2006). In fact, YCW2 and family west of The Netherlands. Brother still with the parents in were the only ones identified in the field on Gotland in the January 2001, but sister already on her own. During the spring of 2004. The next day, they were gone, and two first expedition to the Kolokolkova Bay she was caught on months later, on 30 June 2004, YCW2 and partner were 428 ARDEA 94(3), 2006

back at their breeding site at Kolokolkova Bay, where they already alone during its first autumn. It was then observed produced a nest with six eggs that hatched on 12 July. only a few times until it was trapped during moult at Kolokolkova Bay in August 2003. B5G9 Hatched on 27 May 1995 at Laus holmar, male. Tarsus 9164493 length 85.7 mm (31/200). The only surviving son of two Male, hatched at Skenholmen in 2000 with a tarsus of 85.9 ringed parents. In winter, it was never seen with the par- mm (48/131). There are no observations of this bird wear- ents, which usually stayed in the southwest of The ing only a metal ring until it was shot at Kolokolkova Bay in Netherlands and were observed many times. After two May 2004. years it was observed for the first time wintering along the Elbe river in Germany. On 23 August 1999 it was shot on G6RK Kolguev Island in northern Russia, where large colonies of This male was hatched on 28 May 1994 at Laus holmar. Barnacle Geese exist. Upon fledging its tarsus measured 82.7 mm, which made it an average bird (48/181). In contrast to other long-dis- W=G= tance dispersers, it wintered in the same general area as its Hatched in 1999 at Laus holmar, male, unknown hatch parents and sister in the southwest of The Netherlands. date, tarsus length of 86.3 mm (rank 32 out of 157), ringed Already in 1996, at the age of two, it settled as a breeding together with one brother, one sister and both their par- bird in ‘t Zwin in nearby Belgium, together with an ents. There are no observations of this bird until it was shot unmarked female. It was the first breeding Barnacle Goose at the Kanin Peninsula in Northern Russia in May 2002 at in Belgium of proven wild origin. It was reported by many the age of three years. Parents, daughter, and the other son observers and is generally ‘well-known’ in Belgium. It was where regularly observed at various localities in The seen in the area again during subsequent breeding seasons Netherlands during winter. until 1999. During at least two seasons it was seen with goslings but there is no proof that any of those ever BEGY fledged. Its sister settled close to her parents at Laus hol- The second of the two female dispersers. Hatched in 1996 mar, while its brother was never seen again and presum- at the island of Skenholmen with unknown hatch date, ably died young. where it was retrapped one year later. With a tarsus length of 81.7 mm it ranked number 34 out of 110 fledglings 9166685 ringed there that year. There are four observations at vari- Hatched at Laus holmar in 2001 with unknown hatch date ous sites in The Netherlands in winter before it was shot on with a tarsus of 83.0 mm (77/2670). No observations and 26 May 1999 at Shoina, northern Russia at the age of three. never retrapped on Gotland until it was observed in southern Finland during the summer of 2003. 9167052 Hatched in 2001 at Laus holmar, male, unknown hatch date, tarsus length of 86.2 mm (rank 36 out of 267). Because it had no colour rings, there are no observations of The common denominators of these case histories this bird until it was seen in Southern Finland in August are that long-distance dispersers are predomi- 2002, where it was killed by a golf ball. nantly males, that they are relatively large (Fig. 2), W=G1 and that they seem to have become independent Hatched on 28 May 1999 at Laus holmar, male. Tarsus from their parents relatively early in life. Normally, length 85.7 mm (38/157). Only son of two ringed parents. offspring of geese stay with their parents during Between its birth and 2004 the parents were observed 64 their first winter, and female offspring can even times during winter in The Netherlands, whereas their son join their parents during subsequent winters (Ely was seen three times during the same period, and never with its parents, not even when it was still young. In 2004 1993, van der Jeugd & Blaakmeer 2001, van der it was found breeding at Kolokolkova Bay, northern Russia Jeugd unpubl.). In some cases, the stories of these with an unringed female. dispersing male Barnacle Geese are strikingly different from those of their sisters hatched in the BPWH same year. Quite the opposite of males, female This male was hatched on 25 may 1991 at Laus holmar and had no surviving siblings. A tarsus length of 83.8 mm Barnacle Geese tend to return to their natal colony, earned it rank 48 out of 126 birds. Its parents, also ringed, and actively settle close to female kin (van der where observed many times during winter, but BPWH was Jeugd 2001, van der Jeugd et al. 2002). van der Jeugd & Litvin: DISPERSAL IN BARNACLE GEESE 429

DISCUSSION exploratory personalities are more prone to disperse (Dingemanse et al. 2003). The fact that Natal dispersal is nearly always female-biased in long-distance dispersers were disproportionately birds, but the opposite is observed in waterfowl. large might suggest that the origin of their behav- Greenwood (1980) suggested that male dispersal iour might lie in the conditions they experienced is the rule in this group because the mating system during early childhood. in waterfowl is based on mate-defence rather than Making calculations about the true number of resource-defence, in contrast to most other bird long-distance dispersers based on the 18 birds that species. In addition, Lessells (1985) and Rohwer & are the subject of this paper is hazardous and Anderson (1988) proposed that the benefits of bound to be biased given all the factors that con- knowledge of the breeding area may be greater in tribute to the probability of such birds actually female than in male waterfowl because of their being detected. Different calculations yielded larger investment in reproduction. In an earlier somewhat different estimates of the total number study on the Baltic population, van der Jeugd of birds that disperses over long-distances. A pro- (2001) found that 63% of dispersers that were portion of at most 6.6% seems realistic and yields found breeding in other colonies in the Baltic were a detection probability of around 11%. Johnson males. Out of 172 that were not observed in the (1995) reported similar results from a thirteen- Baltic and, at least partly, must have bred at other year study on Lesser Snow Geese Chen caerulescens localities further away, 70% were males. In the caerulescens in Northern Alaska. Only 17 out of study reported here, 16 out of 18 (89%) true long- 4000 birds ringed as juveniles were resighted in distance-dispersers were males. Taken together, other colonies situated between 500 and 1200 these results suggest that especially long-distance kilometres away from their natal colony, pointing dispersal in Barnacle Geese seems to involve to a detection probability of at most 15%. These mainly males. Moreover, many of these males were figures illustrate that finding long-distance dis- disproportionately large compared to other mem- persers is difficult, and that many years of inten- bers of their cohorts, and were observed indepen- sive studies at various localities are needed to dently of parents and siblings early in life. determine its full extent. The use of large numbers Interestingly, in addition to three birds ringed as of cheap tracking devices that have a lifespan of juveniles on Gotland, two males that were ringed more than one year and can transmit data would as adults on Gotland were captured during moult be a big step forward, but as yet such devices are in Kolokolkova Bay, Russia. Apart from their initial not available. Geolocators (Eichhorn et al. 2006) capture, these males were never seen again on and GPS loggers, although very useful for other Gotland. Both were over one standard deviation types of studies, have the disadvantage that birds above the average size for adult males. carrying them have to be retrieved to download That dispersers are larger than philopatric indi- data and thus are of little help in the study of dis- viduals has been shown before in this species (van persal. der Jeugd 2001, Black et al. 2007) as well as Breeding was very likely in the 18 long-dis- others (Verhulst et al. 1997). Van der Jeugd tance dispersers detected in this study, but was (2001) explained this by proposing that larger definitely proven in only three birds (all males), of males are better at winning agonistic interactions which one was successful in at least one season. at unfamiliar sites. The present study may suggest One other bird breeding in Belgium produced that there is something special about long-distance goslings in several years, but there are no observa- dispersers that make them actively search for new tions of fledged young. This shows that long-dis- horizons. Their early independence may suggest tance dispersers are capable of breeding in envi- that they are behaviourally programmed to do so, ronments that differ widely from their natal one, in a similar way that Great Tits with certain and that gene flow between different populations 430 ARDEA 94(3), 2006

along the east-Atlantic Flyway does occur. So far, and others. Financial assistance from The Netherlands only males have actually been observed breeding, Arctic Program NAP of The Netherlands Organisation of Scientific Research (NWO), the BIRD program of the and the majority of long-distance dispersers were European Science Foundation and the University of males. Hence, gene flow probably only occurs Groningen is gratefully acknowledged. The success of the along the male-line, and this could be detected study at Hellegatsplaten, The Netherlands depended using molecular techniques that can distinguish among others on support by the Faunafonds and between male and female-biased gene flow Staatsbosbeheer. Almost 1000 volunteer ring readers con- (Helbig et al. 2001, Prugnolle & de Meeus 2002). tributed their observations of marked Barnacle Geese. During the first eight years, these observations were One could indeed wonder whether female gathered and processed by Bart Ebbinge and Jan Burgers. Barnacle Geese are at all capable of breeding at The Bird Ringing Centres of Sweden, The Netherlands latitudes that are very different than their natal and Russia provided the metal rings and the administra- one, considering that the traits that are associated tive backbone that is needed for every study that tracks with timing of breeding, and hence differ between individual birds, and the work of many scientists and latitudes, are female traits, which could have a volunteer bird ringers around the globe depends on the prolonged existence of such institutes. genetic origin or result from cultural transmission from mother to daughter. The study of long-distance dispersal and the mechanisms behind it requires long-term field REFERENCES studies of large numbers of individuals, as this study showed. Only large numbers of detailed observations can reveal the different strategies Black J.M., Prop J. & Larsson K. 2007. Wild goose dilem- mas: Population consequences of individual decisions that individuals follow to achieve their goals and in barnacle geese. Branta Press, Groningen, The how personality, previous experience, and learning Netherlands. act in concert to shape the dispersal patterns that Clobert J., Danchin E., Dhont A.A. & Nichols J.D. 2001. we observe. Undoubtedly, more fascinating results Dispersal. Oxford University Press, Oxford. and extraordinary tales of individuals’ travels will Dingemanse N.J., Both C., van Noordwijk A.J., Rutten A.L. & Drent P.J. 2003. Natal dispersal and personali- emerge during the coming years. ties in great tits (Parus major). Proc. Roy. Soc. Lond. B 270: 741–747. Ebbinge B.S., van Biezen J.B. & van der Voet H. 1991. Estimation of annual adult survival rates of Barnacle ACKNOWLEDGEMENTS Geese Branta leucopsis using multiple resightings of marked individuals. Ardea 79: 73–112. This study would have been impossible without the help Eichhorn G., Afanasyev V, Drent R.H. & van der Jeugd of many people. As with all long-term population studies, H.P. 2006. Spring stopover routines in Barnacle the vision and continued effort of a few people is crucial Geese Branta leucopsis tracked by resightings and to their success. Kjell Larsson and Pär Forslund had that geolocation. Ardea 94: 667–678. vision and initiated the long-term study on Gotland in Ely C.R. 1993. Family stability in greater white-fronted 1984 and maintained it to date. Their efforts form the geese. Auk 110: 425–435. basis for the analyses presented here. Mennobart van Ganter B., Larsson K., Syroechkovsky E.V., Litvin K.Ye., Eerden of RIZA (Lelystad) paved the way in 2002 by Leito A. & Madsen J. 1999. Barnacle Goose Branta including our study in his PRISM programme in the leucopsis: Russian and Baltic populations. In: Madsen Russian Arctic and leading our first expedition. The sup- J., Fox T., Cracknell J. (eds) Goose Populations of the port of the institute of Biology in Syktyvkar, the admi- Western Palearctic Wetlands International Publ. No nistration of the Nenets Autonomous District in Naryan 48: 270–283. Wetlands International, Wageningen, Mar, the State Nature reserve Nenetskiy, Naryan Mar, and The Netherlands. National Environmental Research the Russian Hunters Association for the studies in Russia Institute, Rønde, Denmark. is gratefully acknowledged. Apart from the authors, the Greenwood P.J. 1980. Mating systems, philopatry and Russian field team consisted of Yura Anisimov, Rudi dispersal in birds and mammals. Anim. Behav. Drent, Götz Eichhorn, Lena Gurtovaya, Julia Karagicheva 28:1140–1162. van der Jeugd & Litvin: DISPERSAL IN BARNACLE GEESE 431

Helbig A.J., Salomon M., Bensch S. & Seibold I. 2001. Sutherland W.J. 1998. Evidence for flexibility and con- Male biased gene flow across an avian hybrid zone: straint in migration systems. J. Avian Biol. 29: evidence from mitochondrial and microsatellite DNA. 441–446. J. Evol. Biol. 14: 277–287. van der Graaf A.J., Feige N., van der Jeugd H.P., Leito A., Johnson S.R. 1995. Immigration in a small population of Larsson K., Litvin K.Ye., Drent R.H. & Stahl J. 2006. snow geese. The Auk 112: 731–736. Breeding range expansion of arctic geese facilitated Larsson K. & Forslund P. 1994. Population dynamics of by changes in human land use? In: van der Graaf A.J. the Barnacle Goose, Branta leucopsis, in the Baltic (ed) Geese on a green wave: flexibe migrants in a area: density-dependent effects on reproduction. J. changing world. PhD Thesis University of Groningen, Anim. Ecol. 63: 954–962. Groningen. Larsson K., Forslund P., Gustafsson L. & Ebbinge B.S. van der Jeugd H.P. 2001. Large barnacle goose males can 1988. From the high Arctic to the Baltic: the success- overcome the social costs of natal dispersal. Behav. ful establishment of a Barnacle Goose Branta leucop- Ecol. 12: 275–282. sis population on Gotland, Sweden. Ornis Scand. 19: van der Jeugd H.P. & Blaakmeer K.B. 2001. Teenage love: 182–189. the importance of trial liaisons, sub-adult plumage Larsson K. & van der Jeugd, H.P. 1998. Continuing and early pairing in barnacle geese. Anim. Behav. 62: growth of the Baltic barnacle goose population: num- 1075–1083. ber of individuals and reproductive success in differ- van der Jeugd H.P., Gurtovaya E., Eichhorn G., Litvin ent colonies. Research on Arctic Geese. Norsk K.Ye., Mineev O.Y. & van Eerden M.R. 2003. Breeding Polarinstitutt Skrifter 200: 213–219. barnacle geese in Kolokolkova Bay, Russia: number of Lessells C.M. 1985 Natal and breeding dispersal of breeding pairs, reproductive success, and morphol- Canada geese Branta canadensis. Ibis 127: 31–41. ogy. Polar Biol. 26: 700–706. Lindholm A.G., Gauthier G. & Desrochers A. 1994. Effects van der Jeugd H.P. & Larsson K. 1998. Pre-breeding sur- of hatch date and food supply on gosling growth in vival of barnacle geese Branta leucopsis in relation to Arctic-nesting greater snow geese. Condor 96: fledgling characteristics. J. Anim. Ecol. 67: 953–966. 898–908. van der Jeugd H.P., Olthoff M.P. & Stahl J. 2001. Loonen M.J.J.E., Oosterbeek, K. & Drent, R.H. 1997. Breeding range translates into staging site choice: Variation in growth of young and adult size in Baltic and arctic barnacle geese use different habitats Barnacle Geese Branta leucopsis: evidence for density at a Dutch wadden sea island. Ardea 89: 253–265. dependence. Ardea 85: 177–192. van der Jeugd H.P., van der Veen I.T. & Larsson K. 2002. Ouweneel G.L. 2001. Snelle groei van de broedpopulatie Kin clustering in barnacle geese: familiarity or phe- Brandganzen Branta leucopsis in het deltagebied. notype matching? Behav. Ecol. 13: 786–790. Limosa 74: 137–146. van Noordwijk A.J. 1995. On bias due to observer distrib- O’Riain M.J., Jarvis J.U.M. & Faulkes C.G. 1996. A disper- ution in the analysis of data on natal dispersal in sive morph in the naked mole-rat. Nature 380: birds. J. Appl. Stat. 22: 683–694. 619–621. Verhulst S., Perrins C.M. & Riddington R. 1997. Natal dis- Pearson E.S. & Hartley H.O. 1962. Biometrika tables for persal of great tits in a patchy environment. Ecology statisticians. Cambridge University Press, Cambridge. 78: 864–872. Pouw A., van der Jeugd H.P. & Eichhorn G. 2005. Broed- Zera A.J. 2003. The endocrine regulation of wing poly- biologie van Brandganzen Branta leucopsis op de morphism in insects: state of the art, recent surprises, Hellegatsplaten. Unpubl. report. and future directions. Integr. Comp. Biol. 43: 607–616. Pradel R. 1996. Animal dispersal within subdivided populations: an approach based on monitoring individuals. Acta Oecol. 17: 475–483. Prugnolle F. & de Meeus T. 2002. Inferring sex-biased dis- SAMENVATTING persal from population genetic tools: a review. Heredity 88: 161–165. Dispersie kan grote gevolgen hebben voor de mate van Rohwer F.C. & Anderson M.G. 1988. Female-biased philo- uitwisseling van genetisch materiaal tussen populaties, patry, monogamy, and the timing of pair formation in maar is is vaak moeilijk te bestuderen omdat individuen migratory waterfowl. In: Johnston R.F. (ed) Current zich ver buiten het studiegebied kunnen vestigen. In dit Ornithology 5: 187–201. Plenum Publ., New York. onderzoek werd het lot nagegaan van circa 4500 individu- Sedinger J.S. & Flint P.L. 1991. Growth rate is negatively eel herkenbare Brandganzen Branta leucopsis die tussen correlated with hatch date in black brant. Ecology 1984 en 2002 uit het ei waren gekropen op het Zweedse 72: 496–502. eiland Gotland. Van deze vogels werden er slechts acht- 432 ARDEA 94(3), 2006

tien (zestien mannetjes, twee vrouwtjes) teruggevonden werd een op Gotland geboren mannetje broedend vastge- in andere populaties binnen het verspreidingsgebied van steld. Broeden kon echter maar in drie van de achttien de Brandgans (Noordzee, Barentszzee, Spitsbergen). De gevallen worden bewezen, en slechts één van deze broed- meeste van deze ganzen waren erg groot vergeleken met gevallen was succesvol. Tijdens de voorjaarstrek op weg andere vogels die in hetzelfde jaar waren geboren. Ze naar Rusland werd deze vogel, vergezeld van vrouw en waren bovendien vaak al vroeg in hun leven zelfstandig. drie geringde jongen, gedurende één dag waargenomen We schatten dat ongeveer 6,6% van alle jonge Brand- nabij zijn geboortekolonie op Gotland, waar normaal ganzen zich over zulke grote afstanden verspreidt, maar gesproken geen Brandganzen van de Barentszzeepopu- dat ten hoogste 11% van deze bewegingen ook daadwer- latie verblijven. Ongetwijfeld gaan er meer intrigerende kelijk kon worden vastgesteld. De meeste vogels vertrok- verhalen schuil achter het fenomeen van langeafstands- ken naar de grote Barentszzeepopulatie, waar verschei- dispersie. Het is echter duidelijk dat onderzoek naar dit dene vogels tijdens veldwerk in 2002–2004 broedend proces vele jaren in beslag neemt en dat daarbij verschei- werden gezien of werden teruggevangen. Er werd één dene populaties intensief bestudeerd dienen te worden. geval vastgesteld waarbij een mannetje geboren op Gotland naar Spitsbergen was getrokken. Ook in België Received 15 February 2006; accepted 20 April 2006

ARDEA is the scientific journal of the Netherlands Ornithologists’ Union (NOU), published bi-annually in spring and autumn. Next to the regular issues, special issues are produced frequently. The NOU was founded in 1901 as a non-profit ornithological society, composed of persons interested in field ornithology, ecology and biology of birds. All members of the NOU receive ARDEA and LIMOSA and are invited to attend scientific meetings held two or three times per year.

Subscription ARDEA – Separate subscription to Ardea is possible. The 2007 subscription rates are € 33 (The Netherlands), € 37 (Europe), and € 43 (rest of the world). See website for institutional subscription rates. Payments to Postbank account 125347, as indicated below under Membership NOU. Correspondence concerning subscription, and orders for back volumes to: J.J. de Vries, Oosterend 10 b, 8897 HZ Oosterend, The Netherlands ([email protected]). Internet – www.ARDEAjournal.nl

NETHERLANDS ORNITHOLOGISTS’ UNION (NOU)

Membership NOU – The 2007 membership fee is € 41 (or € 24 for persons <25 years old at the end of the year). Foreign membership amounts to € 53 (Europe), or € 63 (rest of the world). All payments to Postbank account 125347 in the name of Nederlandse Ornithologische Unie, Ir. van Stuivenbergweg 4, 6644 AB Ewijk, The Netherlands (BIC: PSTBNL21 IBAN: NL65 PSTB 0000 125347). Correspondence concerning membership: J.J. de Vries, Oosterend 10 b, 8897 HZ Oosterend, The Netherlands ([email protected]). Internet – www.nou.nu