Occurrence of hybrid geese in Sweden – a conservation problem ?

Inter-specific hybridisation is exceptionally common in wildfowl, and even inter-generic hybridisation is not unknown (Johnsgard 1960, Scherer & Hilsberg 1982, McCarthy 2006). Among true geese, almost every combination of hybrid is possible, both within and between the two genera, Anser and Branta (Ogilvie 1978). Hybrids are often produced in captivity but are comparatively rare in the wild (Rutschke 1997). Fertile hybrids are usual within goose genera but not always between them (Owen 1980; but see Kear 1990). All the grey geese are similar genetically (Ruokonen et al. 2000), and many fertile crosses have been produced. Crosses between the Greylag Goose Anser anser and Swan Goose Anser cygnoides, for instance, are completely fertile (Darwin 1859, 1880), but a slight chromosome difference can cause reduced hatchability at back-crossing (Lärn-Nilsson 1996). The several fertile inter-generic crosses between Anser and Branta indicate the very close relationship existing between these two morphologically fairly Occurrence of hybrid geese in Sweden – a conservation problem?

distinct groups (Johnsgard 1960). Still, few studies have tried to measure the importance of hybridisation in terms of actual number of individuals (but see Delany 1993, Randler 2000, Rowell et al. 2004. For Sweden, no such study has previously been done, even if trends for hybrids have been discussed shortly in two papers (Lerner 2005a, Lerner & Kampe-Persson 2006. That closely related species interbreed and produce fertile hybrids is relatively rare in nature, where interbreeding is usually avoided by the action of geographical, ecological, behavioural or morphological barriers (May 1970). When separated by geographical barriers, species may be less likely to evolve ecological, behavioural or morphological mechanisms to avoid interbreeding. In such case, hybridisation may be facilitated when a species is introduced in a region that contains a close relative. Hybridisation caused by fauna manipulation is of that reason a conservation issue, because it may alter the genetic structure and reduce the fitness of native populations. A number of cases corroborate the importance of this phenomenon in birds.
Through fauna manipulation, hybridisation can become a severe conservation problem, sometimes threatening native populations by extinction. The introduction of the Ruddy Duck Oxyura jamaicenses to England in the 1940s is an illustrative example of this (Green & Hughes 1996, Hughes 1996, Persson & Urdiales 1997, Muñoz-Pomer 2006, Hughes et al. 2006. After establishing a naturalised population on the British Isles, the Ruddy Duck started to spread to other countries, where it came in contact with its native relative, the Whiteheaded Duck Oxyura leucocephala. The two species started to hybridise, and the hybrids turned out to be fertile. The situation was aggravated by the fact that both Ruddy Duck males and hybrid males were dominant over White-headed Duck males, at the same time as the White-headed Duck was recovering from a very low population level (Torres 2001).
Also the Fennoscandian Lesser White-fronted Goose Anser erythropus conservation project has been affected by hybridisation (Aarvak & Timonen 2004). When genetic studies revealed that approximately 25% of the studied captive Lesser Whitefronts, used in the Swedish re-establishing project (Andersson 2004) and the Finnish restocking project (Markkola et al. 1999), carried the mitochondrial DNA of Siberian White-fronted Goose Anser albifrons albifrons (Ruokonen 2000, Tegelström et al. 2001, further releases were stopped in both countries. Delaying conservation measures within this field can prove very expensive. Reducing the British Ruddy Duck population to the level it had when the problem of hybridisation with White-headed Duck became known will take several years and be very costly (Smith et al. 2005). To that must be added the cost of eradicating all Ruddy Ducks and hybrids in those countries in Europe, North Africa and the Middle East where the species has spread to since 1980. These are the extra costs, coming on top of the cost to eradicate the British population in 1980 of about 300 individuals. Accordingly, when actions must be taken to prevent hybridisation, it is of utmost importance that these can be decided on a sound basis and implemented without unnecessary delay.
This study aims at providing basic data about hybrid geese and mixed pairs in Sweden, combinations of species, numbers, trends and origins, which can serve as a framework for future studies, especially when conservation measures, such as control programmes or eradication, are actualised. Special attention was paid to species included on the Swedish Red List. Included on this list are the Lesser White-fronted Goose as critically endangered and the Taiga Bean Goose Anser fabalis as near threatened, while the nominate race of the latter species is included as vulnerable (Gärdenfors 2005). Also, attention was paid to the Red-breasted Goose Branta ruficollis, on the IUCN Red List regarded as vulnerable (IUCN 2006).

Material and methods
Hybrid geese often present problems of identification. In some hybrids, morphological characters from both parents are obvious, as for example in crosses between Greylag and Greater Canada Goose Branta canadensis. Hybrids between closer relations than these might be harder to distinguish, as for example crosses between Lesser and Siberian White-fronted Goose (see, for example, Lerner 2005a). Identification is further complicated by individual differences within crosses (Randler 2001), differences depending of which species the parent male was, and the existence of second-generation hybrids (Lack 1974). For crosses between Ruddy and White-headed Duck, the characters of the hybrid depend on which species the parent male was (Urdiales & Pereira 1993). The same seems to apply to hybrid geese (cf. Olsson 1954). In some crosses, large differences exist even among individuals from the same brood (see e.g. Lebret 1983, Rodríguez & Palacios 1993, Palacios & Rodríguez 1999).
These problems of identification, combined with inexperience in observing odd geese, have resulted in a number of misidentifications in Sweden (Amcoff et al. 1989, Carlsson et al. 2006. The first crosses between Greylag and Greater Canada Goose seen in a winter flock, for instance, were recorded as Barnacle Geese Branta leucopsis (Bolund 1964). Several crosses between Barnacle and Bar-headed Goose Anser indicus were reported as hybrids between Bar-headed Goose and either Greylag or Greater Canada Goose, while crosses between Barnacle and Snow Goose Anser caerulescens often were reported as Blue Goose; the Blue Goose is a colour morph of Snow Goose. Crosses between Greater Canada and Siberian Whitefronted Goose have been reported as hybrids between Barnacle and Lesser White-fronted Goose. In Blekinge, crosses between Lesser White-fronted Goose and Cackling Goose Branta hutchinsii minima were initially reported as Lesser White-fronted Geese (Nilsson 1983), while in Skåne, a male Cape Shelduck Tadorna cana was reported as a hybrid between Ruddy Shelduck Tadorna ferruginea and Greylag Goose. Albinistic Greater Canada Geese have been reported as hybrids between Greater Canada and either Snow or Bar-headed Goose, while some juvenile Bar-headed were reported as hybrids. Most of the sighted second-generation hybrids have, of obvious reasons, most likely been reported as first-generation ones.
This report is based on data published in national, regional and local magazines and reports, observations made by the authors (HKP and HL), and unpublished sightings made by others, reported either to the authors or on Svalan (www.artportalen.se), up to August 2007. Also, several regional rarities committees have provided information and two preliminary papers (Lerner 2005a, Lerner & Kampe-Persson 2006) enhanced reports of hybrids. All existing data were not available for this report however, partly because many bird-watchers pay little attention to these birds and partly because all regional and local publications were not accessible.
In general, sightings in this report were based on the observers' suggestion of parent species. In Sweden, hybrids are rarely considered by local, regional or national rarities committees. In some cases, for instance when the same bird was given more than one combination of species by different observers, or as in the cases given above, misidentifications were corrected. Omitted were all hybrids where only one of the parent species was identified. Combinations of species for which all available sightings were denoted either probable or possible were excluded. In all but a few instances when sightings are included in this report, the word probable is deleted, a word that some observers have used for nearly every single hybrid reported by them. Also, mixed pairs, where a male of one species was paired to a female of another species, were included. These are more seldom reported.
The nomenclature follows recommendations of Commissie Systematiek Nederlandse Avifauna (Sangster et al. 1999(Sangster et al. , 2003, which means that Bean Goose is represented by Taiga  Combinations of species in the headings in the Appendix are given in alphabetical order, without indicating which of the parents being male or female. For second and third generation hybrids sci-entific names within brackets refer to one of the parents. Abbreviations of geographical provinces, given in brackets after each site name, are in accordance with those used for national reports (SOF 2002(SOF , 2006. For observations published in bird reports or on Svalan (www.artportalen.se), as well as earlier unpublished ones, name(s) of observer(s) is/ are given in brackets if there was/were only one or two, while only the first name is given in case of more than two. This applies also to sightings where the identification was changed or modified by the authors.
For most combinations of species, all available data were included, in the text or in an Appendix, to facilitate future updates.

Results
No less than 17 species were involved in the hybrid geese sighted in Sweden through August 2007 (Tabell 1, Appendix). Involved in most first-generation combinations of species were the Barnacle Goose (ten), followed by the Greylag Goose (nine), the Greater Canada Goose (eight), and the Siberian White-fronted Goose and the Snow Goose (seven each). Represented by only one type each were the Ross's Goose Anser rossi, the Pale-bellied Brent Goose and the Red-breasted Goose. One species, the Greylag Goose, was also represented by a cross with a swan species, and mixed pairs with a duck species and Crane Grus grus as well.
Some combinations of species have been known for a long time, the first ones (Siberian White-fronted Goose x Snow Goose, Siberian White-fronted Goose x Taiga Bean Goose, Siberian White-fronted Goose x Bar-headed Goose and Greylag Goose x Greater Canada Goose) appeared already in 1918. Most of the other combinations of species were only reported during the last 30 years; second-generation ones mainly since the mid-1980s (Tabell 2, 3).
Ever since the Greater Canada Goose was introduced in Sweden three quarters of a century ago, crosses between this species and the Greylag Goose have been by far the most common combination of species. During the non-breeding season, some of the hybrids were seen in Greylag Goose flocks but the vast majority among Canada geese. In flocks of Greylag and Greater Canada Geese staging and wintering in South Sweden, there were no clear trends in the frequencies of hybrids during the last 36 years (Tabell 4). In Greylag flocks, the average frequency was higher during the years 1991-  N=350) in 1964N=350) in /1965N=350) in and 4.00‰ (N=500) in 1965N=350) in /1966N=350) in (Axelsson 1967. The second most commonly reported combination of species is Greater Canada and Barnacle Goose, and the third Lesser White-fronted and Barnacle Goose.
An estimate of the total number of hybrids in Sweden was made for 2005, the latest year with good data (Table 5). For crosses between Greylag and Greater Canada Goose, an estimate of the total number of individuals in Sweden was obtained by multiplying observed frequencies (Table 4) with population sizes of the two species. Estimated sizes of the post-breeding populations in 2005 were 210,000 Greylag Geese and 100,000 Greater Canada Geese (Nilsson 2006, Svenska Jägareförbundet 2006. In that way, a total of 226 hybrids was obtained, 13 among the Greylags and 213 among the Greater Canadas. A similar calculation was made for crosses between Pink-footed and Taiga Bean Goose, as 20% of all staging bean geese in Sweden in the late 2005 autumn were carefully checked by HKP. However, as it was uncertain whether there were any hybrids in the unchecked flocks, the result is presented as an interval, ranging from the number of hybrids actually seen to the calculated figure. For other combinations of species, estimates were based on available sightings.

The problems of identification
A chain of factors, running from occurrence to inclusion, such as observer availability, knowledge and devotion of the field observers, detectability, reporting frequency, publishing frequency, and the ability of the reviewers to find the sources, creates the necessary conditions for how well a review article can describe the true picture.
Staging and wintering geese in Sweden started to be checked more in detail from the late 1950s (e.g. Lennerstedt 1962, Markgren 1963. Since then, the number of persons checking the geese have doubled many times over. Better binoculars and telescopes, as well as better roads and access to watch towers, have facilitated field observations. At the same time, however, most goose populations regularly occurring in Sweden have increased markedly in size (Nilsson 2006). These larger numbers, in combination with more and more multi-species flocks, make it in several places harder to spot hybrids today than earlier. During the period covered by this report, hybrids went unnoticed in two ways,   1910-1919 1 1920-1929 1 1930-1939 2 1940-1949 0 1950-1959 1 1 1960-1969 2 1970-1979 5 1980-1989 14 1990-1999 10 2000-2007 9 1 = before 1959. partly by flocks and families remaining unchecked, especially during the breeding season, and partly by bird-watchers paying little attention to hybrids. As a result of the latter, at least up to the late 1990s, the bulk of all sightings in most geographical areas was reported by only 1-3 persons. But even if many hybrids passed unnoticed for long periods, there were good chances that they, due to the longevity and mobility of geese in general (Owen 1980), were seen at least occasionally. Identification of hybrid geese is not covered by field guides, and is rarely the theme of identification articles. By that, knowledge gained by the most devoted field observers is not passed on to others. Of that reason, the ability among bird-watchers in general to correctly identify encountered hybrids has improved slowly over time, mainly reflecting a  (Rowell et al. 2004) and in 1991 (Delany 1993) and Germany in 1998 (Randler 2000 (Rowell mfl 2004) och 1991 (Delany 1993) och Tyskland 1998 (Randler 2000 general improvement in identification skill. Above all, individual variation within each combination of species have and will give rise to misidentifications. Individual variation is supposedly partly caused by which species the male is. Encountering hybrids where the male parent was of the opposite species compared to the normal can cause the ob-server to believe that these birds were of another combination. Improvements in identification skill over time might explain some of the new combinations of species reported in Tables 2 and 3, but far from all. As the aim of the study was to describe the occurrence of hybrid geese in Sweden, it was impossible to restrict the material to only those hybrids that for certain were correctly identified. The more so as hybrids have to be studied genetically to be sure of their ancestry. It is not enough to see the hybrid together with its parents, as extra-pair fertilisation occurs also in mixed pairs (Berg 1937). Obvious misidentifications were corrected before inclusion in this report, but all of them were most likely not found. As the majority of all observations, also of the least reported combinations of species, were made by devoted field observers, the overall frequency of misidentifications was probably low. The largest remaining uncertainty regards crosses between Emperor Anser canagica and Barnacle Goose. Available sightings, some of them reported as unidentified Emperor Goose hybrids, were not included until one of them had been reported as a cross between these two species. Another kind of misidentifications was when second-and thirdgeneration hybrids were reported as first-generation ones. This was probably of minor importance as such hybrids apparently occurred in very low numbers. After three generations of back-crossing it is probably no longer possible to identify the individual as a hybrid in the field.
The chance of being overlooked in the field is, of obvious reasons, higher for crosses between closely related species, with high morphological resemblance, than for inter-generic crosses, and higher for second-and third-generation hybrids than for first-generation ones. The most likely candidates of having passed unnoticed during the period covered by this study were the following combinations of species: Siberian White-fronted Goose x Lesser White-fronted Goose, Greylag Goose x Pink-footed Goose, Greylag Goose x Swan Goose, Greylag Goose x Taiga Bean Goose, Greylag Goose x Bar-headed Goose, Dark-bellied Brent Goose x Light-bellied Brent Goose, Snow Goose x Barnacle Goose, Greater Canada Goose x Barnacle Goose, Bar-headed Goose x Barnacle Goose and Lesser Canada Goose x Barnacle Goose. There are no data indicating to which extent hybrids were overlooked.
The interest of reporting and publishing observations of hybrid geese in Sweden has varied markedly during the past decades, among areas as well as years, which becomes evident when examining regional and national bird reports. The introduction of Svalan (Report system for Birds, www.artportalen.se) changed this by stimulating a higher reporting frequency and enhancing data availability. However, in the only study of reporting frequency so far it was found to be extremely low. Of hybrids found during four mid-monthly counts of geese in South Skåne during the period September 2006-January 2007, only 5% (N=61) were reported on Svalan (Kampe-Persson 2007). This low figure was obtained even though data were extracted for a 16day period each counting month.
The picture given in this study might be somewhat skewed, partly due to the omission of all hybrids where only one of the parents was identified and partly to an unknown number of misidentifications. If it had been possible to include also these sightings correctly identified, the total number of combinations of species might have been higher than now, especially second-and third-generation ones, and the relative frequencies of the different combinations of species had very likely been slightly different. For instance, as Swedish birders rarely differentiate between Lesser and Greater Canada Goose as hybrid parent, generally reporting all of them as the latter, the total number of sightings of crosses between Lesser Canada and Barnacle Goose had very likely been higher than now. As few geese of captive or semi-captive origin are reported correctly to the rarities committees, some sightings of small Canadas might have been hybrids. A better understanding in this issue might be reached with more written descriptions and photos of these birds.
In spite of the shortcomings discussed above, there are good reasons to accept the picture given in this study as a good representation of the true one.

Trends
During the latter half a century, the number of hybrid geese in Sweden shows a positive trend similar to that of the breeding populations of Greylag, Greater Canada and Barnacle Goose , but also to the summer occurrence of northern breeders and exotic geese (Kampe- Persson & Lerner, in prep.). To which degree this trend is due to increasing breeding populations, to differences in number of birds of captive and semi-captive origin, to the summer occurrence of species normally breeding farther north, or to a higher reporting frequency is and will remain unknown.
The trend of an increasing number of first-sightings of new hybrid types from the 1980s is applicable to also other goose areas in Europe. In goose flocks in North and West Europe, it is nowadays normal to find a great variety of hybrids, especially crosses involving introduced species (Delany 1993, Lensink 1996, Randler 2000, van Horssen & Len-sink 2000, Rowell et al. 2004. When originating from birds used in re-stocking and reestablishment projects, also parts of native populations can be regarded as alien (Randler 2000). In Sweden, that applies to Taiga Bean, Greylag, Lesser White-fronted and Barnacle Goose (Svensson et al. 1999).
There were large similarities among Sweden, Germany and Great Britain, in total number of hybrids, that the majority (61-82%, though lower in Great Britain in 2000) was made up of crosses between Greylag and Greater Canada Goose, and that the minority was made up of at least 12 different combinations of species (Delany 1993, Randler 2000, Rowell et al. 2004. Betweencountry differences for combinations of species with low numbers (less than 20 individuals) may be due to stochastic reasons, but notably crosses between Greylag Goose and Swan Goose were rarer in Sweden in 2005 than in Great Britain 2000 and Germany 1998. As geese are long-lived (for the Greylag Goose, see Kampe-Persson 2002), a single mixed pair can produce a large number of hybrids during their reproductive life, hybrids that can be alive long after their parents have died. In Skåne, one hybrid was more than 17 years old when last seen (this study). Hence, one reason that so many combinations of species were represented by less than 15 individuals might be that each combination was produced by a single pair.

Origins
Of hybrid geese seen free-flying in Sweden, some were born in captivity, semi-captivity or close to sites with captive breeding, others in the wild abroad, but the vast majority in the wild in Sweden. Also among those born in the wild most had at least one parent or ancestor of captive origin. The captive origin is obvious when it comes to the introduced Greater Canada Goose and aliens, but rarely for birds emanating from re-stocking or re-establishment programmes, or for escaped individuals of native species. Geese that have been bred in captivity for a number of generations are more prone to hybridise than their wild con-specifics (Randler 2000). In the Greylag Goose, for instance, the frequency of hybrids was markedly higher in naturalised than in natural populations (Bruns 1982, Sibley 1994, Kreutzkamp 1996, Randler 2000. Hence, effects of captive origin must be taken into consideration for a long time after that geese have been released for re-stocking or re-establishment. As most hybrid geese have an obvious captive origin, it might be more fruitful to ask: Which of the reported combinations of species were of natural origin? If, besides crosses involving introduced species, also those involving released and escaped parent birds are excluded, only seven combinations of species remain: Siberian White-fronted Goose x Lesser White-fronted Goose, Siberian White-fronted Goose x Barnacle Goose, Dark-bellied Brent Goose x Light-bellied Brent Goose, Dark-bellied Brent Goose x Barnacle Goose, Pink-footed Goose x Taiga Bean Goose, Pink-footed Goose x Tundra Bean Goose and Barnacle Goose x Red-breasted Goose. Most likely, Taiga Bean Goose x Tundra Bean Goose does not belong to this group. Very high frequencies of hybrids between Taiga and Tundra Bean Goose were formerly reported from Central Europe (Johansen 1962, Bauer & Glutz von Blotzheim 1968, Litzbarski 1974, Cramp 1977, Ogilvie 1978, Owen 1980, Rutschke 1983a, 1983b, Klafs & Stübs 1987, Liebherr & Rutschke 1993, Rutschke & Liebherr 1996, with no less than 54%, 73% and 97% of all bean geese in Mecklenburg, Brandenburg and Hungary, respectively, reported as hybrids. These false conclusions were due to a typological approach, which did not account for the individual variation that occurs in the subspecies fabalis and rossicus, especially in the shape and coloration of the bill (Persson 1995b(Persson , 1997b. A careful check of mixed flocks, where one of the species occurred in low numbers, were recently carried out in Germany. The outcome of this study was the sighting of one mixed pair (Thomas Heinicke, pers. comm. to HKP).
During more than 40 years of careful checks of the European bean goose haunts neither Georges Huyskens nor Leo van den Bergh found any crosses between Taiga and Tundra Bean Goose (Huyskens 1986(Huyskens , 1999, Leo van den Bergh, in litt.), or any crosses or mixed pairs between Pink-footed and Taiga Bean Goose (Leo van den Bergh, in litt.). The lack of sightings south of the Baltic highlights the origin of the bean goose hybrids seen in Sweden, especially as it cannot be ruled out that one, or even both, of the parent birds in each pair was of captive origin. Five Pink-footed Geese of presumed captive origin were, for instance, sighted in Skåne in the autumns of 1995 and 1996 (Elleström et al. 1996, Green et al. 1997. The fact that two of the four sightings of hybrids between Taiga and Tundra Bean Goose in Skåne were done in early autumn, before the parent species started to arrive in this part of Sweden, gives support to a captive or semi-captive origin. The first Pink-footed Goose reported breeding in northern Fennoscandia bred paired to a Greater Canada Goose at Hammervatnet in Nord-Trøndelag in 1994(Husby 1994. This was followed by a female that bred paired to a male Siberian Whitefronted Goose at Lofoten in 2002 and 2003 (Birina 2005). Before the first breeding pair was found at Tromsø in 2003 (Irgens 2004), one Pink-footed Goose was caught together with moulting Taiga Bean Geese at Tana in 1971 (Tveit 1984), breeding was reported but never confirmed from Karasjok (Gjershaug et al. 1994) and a family group of two adults and three young was found at Øysand in Sør-Trøndelag 6 September 1992 (Gustad 1993), the early date indicating that breeding might have taken place on the mainland. An increasing Pinkfoot population in northern Fennoscandia, at the same time as the Taiga Bean Goose population decreases there, could result in hybridisation between these two species.
Hybridisation between Lesser and Siberian White-fronted Goose is believed to be occasional in the natural populations (Nagy 1950, Shackleton 1956, Voous & Wattel 1967, Rogers 1979, Panov 1989, Müller 2001, Lerner 2005a, this study, but see Van Impe 1982), although difficult to detect reliably due to the great morphological similarity of the two species. According to Peter Scott (British Birds 49(1956): 229): "It is my theory that hybrids between Lesser and European [Siberian] White-fronts occur from time to time among the geese visiting Britain. This idea is based (1) on the occurrence, in the winter flocks, of Lesser White-fronts which are paired to European [Siberian] White-fronts; (2) on observations of "Lesser White-fronts" which are unusually large; and (3) on the bill-measurements of a specimen from the Severn Estuary now in the British Museum (Natural History)." A male Lesser White-fronted Goose mated to a Siberian White-fronted Goose with two young, the young not distinguishable from those of Siberian Whitefronts, were seen at Slimbridge, Gloucestershire 6-7 March 1956 (Scott & Boyd 1956). Another mixed pair was seen at the same site 28 February-5 March 1969 (Smith 1970). An indication of the possible magnitude of hybridisation between these closely related species can be obtained by comparing with a similar pair of goose species in the Nearctic. Among 5471 Ross' and 8155 Snow Geese caught for ringing in USA during 1961-1968, there were 32 hybrids, corresponding to a frequency of hybrids of 2.34‰ (Trauger et al. 1971). Overall among birds, however, about one in 50,000 individuals (0.02‰) is a hybrid (Mayr 1970).
Injuries can force migrating geese to remain in or close to their wintering quarters during the following breeding season. Sometimes such birds breed far south of their normal breeding range, paired to either a con-specific or a bird of another species. Examples of the former are the successful breeding of a Greylag Goose pair in northern Spain in 1993 (Moreno 1993) and of Tundra Bean Goose pairs in the Netherlands in 1993 and 1994 (Lensink 1996b) and of the latter, the unsuccessful breeding of a mixed pair of Taiga Bean and Greylag Goose in South Sweden in 1992 (this study). So, hybridisation outside the breeding range of one of the species making up a mixed pair is not always unnatural.
To understand the causes of hybridisation between waterfowl species, several theories have been proposed (for a summary, see Randler 2000Randler , 2006. Field data from Swedish goose haunts support at least two of these: the "Best-Option-Hypothesis" and "Inter-specific mate choice following false imprinting". In most cases, when breeding and breeding attempts of mixed goose pairs were recorded in Sweden, one of the parent species was represented by only one individual at that very site (this study). That applies not only to the rare species, but also to such numerous and wide-spread ones as the Greylag Goose and the Greater Canada Goose (Fabricius & Norgren 1987). These observations support the "Best-Option-Hypothesis". Instead of giving up breeding altogether, the singly bird mate with an individual of another species (Hubbs 1955).
Some of the Lesser White-fronted Goose males from the Swedish re-establishment project, which were released with Barnacle Geese as foster parents, later paired to Barnacle females (Lerner 2005a, this study). Such inter-specific mate choice after false imprinting has been demonstrated in the Greylag Goose (Fabricius 1991). Of 19 males and 16 females that had been cross-fostered by Canada geese in semi-captivity, five males paired to Canada females, while the other males and all females paired to Greylags. When widowed, these five males re-paired to a new partner which was always a Canada female. Two of the males paired as many as nine and five times. In the same study, all of 78 Greylag males that had been reared by conspecifics paired with females of their own species and never even courted Canada Geese. A detailed discussion on the factors involved in sex differences in sexual imprinting was given by ten Cate (1985).
Cross-fostering during natural conditions can oc-cur either by a pair taking over a nest where a female already has started to lay (Fabricius 1983) or by brood amalgamation (Eadie et al. 1988, Beauchamp 1998, Randler 2005 (Dittberner & Dittberner 1976, Hauff 1982, Plath 1985. A strong pattern of natal female philopatry and male dispersal is the usual pattern in Anatidae (Greenwood 1980). In the Greylag Goose, the female rarely disperse more than 10 km, while the male can start to breed more than 1000 km from the site of birth (Nilsson & Persson 2001). A similar pattern was found in a naturalised Greater Canada Goose population (Lessells 1985). One consequence of this sex-difference is that males can get problems in finding a female of their own species during a period of breeding range expansion. The Swedish Greylag Goose population is in such a phase since half a century (Svensson et al. 1999), during which time it has increased hundredfold in numbers. The fact that about 95% of all crosses between Greylag and Greater Canada Goose were found in flocks of the latter species should probably be viewed as most successful mixed pairs were made up of a male Greylag and female Canada. Field evidences are lacking however, as very few parent birds were ever sexed.
As the Greylag and the Greater Canada Goose get breeding ranges that are more or less overlapping, one can expect fewer crosses according to the "Best-Option-Hypothesis". At the same time, the chances of pre-and post-hatch brood amalgamation increase, for instance, by Greylag females laying eggs in Canada nests (Fabricius & Norgren 1987). So far, however, mixed broods of these two species have only been reported from three breeding sites in Sweden (see above).
Owing to problems of identification, secondgeneration hybrids are hardly ever reported except when seen together with both parents. In Sweden, there were reports of nine broods of seven different combinations of species, with a total of 28 young, and one possible young of an eighth combination as well (this study). These eight combinations represented, however, three different crossing possibilities, as six were back-crosses, one a cross between hybrids and one involved a third species. Most combinations of species proving to be fertile in this study were known to be so already half a century ago (Johnsgard 1960), the combination of Greylag Goose and Greater Canada Goose already by Bengt Berg (see Jansson 1984). In the latter combination of species, besides back-crosses with both parent species, also successful breeding of two hybrids was reported (Persson 1997). Moreover, one such hybrid was seen paired first to one of the parent species and later to the other one (Snatter 18(1993): 4). The only observation of a third-generation hybrid was, not surprisingly, recorded between Greylag and Swan Goose (cf. Lärn -Nilsson 1996).
Hybrids differ from individuals of the parental species not only in morphology but usually also in fertility and viability (Mayr 1970, Lack 1974. In general, the hybrids recorded in Sweden seemed to produce very few offspring. Several breeding at-tempts did not result in any hatchlings, and in those producing fledglings, two thirds of all broods numbered only 1-2 young. Larger broods were only produced by pairs where one of the parents was a hybrid between a Greater Canada Goose and a Barnacle or Greylag Goose. When three species were involved, a total of three breeding attempts of three different species combinations resulted in only one hatchling, a bird that never fledged. A fourth threespecies combination, a Siberian White-fronted Goose x Greylag Goose hybrid paired to a Lesser White-fronted Goose, was reported by Kolbe (1972).

Conservation implications
An increased incidence of hybridisation, similar to that at range expansion, often occurs during population declines (Randler 2006). This has implications for Red List species, in Sweden especially due to an increased occurrence of individuals of captive origin.
The only native species in Sweden that seemingly has been affected by hybridisation is the Greylag Goose, in which introgression of genetic material from the Swan Goose has taken place in Skåne. This process passed almost unnoticed however. Instead, it was the revelation that about 25% of the captive stock of Lesser White-fronted Geese, used in the Swedish re-establishing project, carried mitochondrial DNA of Siberian White-fronts that caught attention. Some geneticists are of the opinion that these haplotypes might be a relict from a distant shared ancestor (Kholodova & Severtsov 2002), but the general assumption is that this introgression took place by hybridisation in captivity (Ruokonen et al. 2004). However, it can very well be that it originates from a wild-caught individual, as such hybridisation occurs in the wild (see references above). One reason for genetecists not having found Siberian Whitefront genes in wild Lesser Whitefronts (Ruokonen et al. 2004(Ruokonen et al. , 2007 might be that such hybrids mainly associates with Siberian White-fronted Geese. The scientificly founded reasons to eradicate birds that have been released are of that reason quite weak. In either case, to obtain a more natural, though unknown, frequency of Siberian Whitefront DNA in the re-established Lesser Whitefront population, future releases in Sweden should be of pure-bred birds, and in large scale, by means of which the existing frequency can be diluted (cf. Andersson 2006, Andersson & Holmqvist 2007. As hybrids produced by released Lesser White- The fact that the Lesser White-fronted Goose is critically endangered according to the Swedish Red List (Gärdenfors 2005), urges for action to prevent all kinds of devastating hybridisation. Also the EU Life Nature project "Conservation of the Lesser White-fronted Goose on European migration route" identifies hybridisation as a threat, however, in quite a different way: "An additional threat for the Fennoscandian LWfG is the possible hybridisation with the reintroduced and/or escaped captive LWfG (cf. IUCN guidelines for reintroduction and restocking). As shown by the recent genetic studies of the Finnish and Swedish stocks used or planned to be used in the reintroduction programmes, hybridisation with the White-fronted Goose (Anser albifrons) and Greylag Goose (Anser anser) has occurred several times during the captive history. The reintroduced birds of these captive stocks form also threat in the sense that the small unpedigreed and ill managed captive populations might have accumulated deleterious mutations with untested effects in the wild. When introduced into small wild population, these alleles might become quickly fixed by genetic drift and accelerate the extinction of the wild Fennoscandian population. The present project will not, however, directly target the genetic threat." (WWF Finland). The genetic studies referred to in the cited statement were made by Ruokonen et al. (2007). But, at the same time as the released birds are depicted as a threat, re-stocking is very likely the only measure by which the native population can be prevented from extinction within a near future. It is in the light of this, that projects other than the Life project shall be viewed (for a summary, see Hansson 2005).
Also for the Taiga Bean Goose, all factors negatively affecting the Swedish population (for a summary, see Kampe- Persson et al. 2005) ought to be limited to a minimum, due to the status of this taxon as vulnerable (Gärdenfors 2005). Evidently, hybridisation did not become one of these factors until quite recently (this study). But due to its small population size of only 800-1,250 breeding pairs (Leif Nilsson, pers. comm. to HKP), measures to prevent introgression of genetic material from closely related taxa, especially from birds of captive origin, should be considered for inclusion in a future conservation plan. Fauna falsification of that kind has occurred in the Swedish Greylag Goose population, by introgression of genes from the eastern subspecies rubrirostris (Kampe- Persson 2002), resulting in the occurrence of individuals with pink bills (Kampe- Persson 2003).
Regarding the third Red List species, the Redbreasted Goose, hybrids recorded in Sweden so far seem to have originated from the species' natural breeding grounds. Several summer sightings of free-flying Redbreasts with a suspected captive origin were done during the last years however (see, for example, Lerner 2004a). These birds might be a source for future hybridisation.
Both introduced and naturalised species have been found to hybridise with native geese, which calls for increased vigilance in regard to invasive species. After establishing itself in Great Britain (Sutherland & Allport 1991, Delany 1993, Brown & Grice 2005, Baker et al. 2006) and the Netherlands (Lensink 1996, van der Jeugd et al. 2006, the Egyptian Goose Alopochen aegyptiacus is now invading Sweden (Bengtsson 2007). Except for Common Shelduck, no reports of hybridisation with native species exist so far, only with introduced ones, such as Greater Canada Goose and Ruddy Shelduck Tadorna ferruginea (Lensink 1996). In captivity, however, interbreeding has occurred between Egyptian and Greylag Goose (Gray 1958).
Some culling of hybrid geese is taking place on a regular basis, for instance in Skånes Djurpark (Staffan Åkeby; in litt.), and attempts are made to cull the hybridising Lesser White-fronted Geese of the Swedish re-establishment project (Lerner 2005a). One such bird, paired to a Barnacle Goose, was, for instance, shot at Hjälstaviken (Upl) 20 April 2007 (Karl-Gustav Sjölund et al.). That male, released in Swedish Lapland in 1997, had produced hybrids during the last 2-3 years, and should most likely have continued to do so (Bo Fagerström, in litt.). Another male, which never had shown any indication of hybridisation, was caught in 2005 (Åke Andersson, in litt.). Culling is also used locally to prevent naturalised populations from growing too large. It seems worth considering if culling should be used on a larger scale, as naturalised populations can serve as sources for both hybridisation and the spread of alien genes and behaviours. The more if situations similar to those in Great Britain (Delany 1993, Baker et al. 2006) and the Netherlands (Lensink 1996a, 1998, van Horssen & Lensink 2000, Ouweneel 2001, Sovon Vogelonderzoek Nederland 2002, van Roomen et al. 2003, van der Jeugd et al. 2006) are undesirable. Green, A. & Hughes, B. 1996
Anser anser x Anser fabalis One pair, the male a Taiga Bean Goose, bred unsuccessfully at Yddingen (Sk) in 1992 (HKP), while a pair of captive origin, the male a Greylag Goose, made a breeding attempt at Årsta (Srm) in 2003 (Mats Gothnier). A third pair, the male a Greylag Goose, wintering at the Kampanische Heide in the Netherlands since the early 1980s, was seen at Hagby ( Anser anser x Anser indicus One pair was seen at Hornborgasjön (Vg) 16 May 2007 (Stig Karlsson et al.), while 20 sightings were reported during the period 1984-2006, but more than one individual was seen only twice (Table A2). A Norwegian origin for some of the observed individuals is possible, as breeding was recorded at Namsos in Nord-Trøndelag in July 1991(Gustad 1992. (Anser anser x Anser indicus) x Branta canadensis One pair, where the hybrid was not fully identified, staged at Eggby (Vg) 14 May 2006 (Patrick Åström, Sofia Åström).
Anser anser x Branta canadensis Free-flying hybrids were produced at Kalmarsund (Sm) as early as in the late 1920s (Berg 1930(Berg , 1937. Reports of hybrids among staging and wintering Canada geese started to appear in the 1960s (Borg 1962, Bolund 1964, Nord 1964, Jonasson 1966, Axelsson 1967, Johansson 1968), but observations were made in Norrköping already in the winter 1956-57. By now, this hybrid type has been reported from most parts of the country, but in some regions not until recently, as for instance, Gotland (1997) andNorrbotten (2003). In 2007, a mixed pair was seen in both Lule and Pite lappmark (Stefan Holmberg).
The earliest reported breeding occurred in Boulongerskogen (Hls) in the 1960s (Johan Södercrantz). Since then, breeding has been reported from large parts of the country. At the typical breeding site, there were no pairs of one of the species, while the other was represented by quite a number of pairs (see e.g. Elmfors 1980, Eriksson 1982, Cronert & Svensson 1991. Some pairs were very productive, illustrated by observations of pairs with 11 juveniles at Vesan (Bl) 17 October 1989, 10 juveniles at Trolle-Ljungby (Sk) 8 September 1985, 9 juveniles at N Håslövs By (Sk) 19 November 1999  (Anser anser x Branta canadensis) x Anser anser Pairs were seen in Uppland before 1984 (Amcoff et al. 1984), at Lidingö (Upl)  (Anser anser x Branta canadensis) x Branta canadensis Pairs with one young were seen at Södersjön (Upl) 10 July 1985 (Edholm 1987) (Table A3).
(Anser anser x Branta canadensis) x (Anser anser x Branta canadensis) Eleven 1st cy birds were seen together with hybrid parents at Ellestadsjön (Sk) 29 July-1 August 1993 (Persson 1995a). A similar group, numbering 11 hybrids, was seen in the very same lake 12-14 September 1991, but at that time it was not possible to check if they made up a family (HKP).  (Table A4).

Anser anser x Branta leucopsis
Anser anser x Cygnus olor One pair, the male a Greylag Goose, produced at Staffanstorp (Sk) one fledgling in both 1972 and1975 (Anders Björkman). The Greylag had been raised up by a Mute Swan pair.

Anser caerulescens x Anser indicus
Three different individuals were probably involved in the sightings made during the years 1999-2006 (Table  A5).
Anser caerulescens x Branta canadensis Pairs bred at Ingsbergasjön (Sm) in 1978(Thorin 1981), Viaredssjön (Vg) in 1983 and Kvarnsjön (Vg) in 1989, while a breeding pair was seen at Romanäs (Sm) in 2004 (Anders Ring). In Viaredssjön, Borås a pair occurred in 1982, made the first breeding attempt with three hatched eggs in 1983, while two hybrids were seen and six eggs resulted in at least four fledglings in 1984. During the following years (1985-86 and 1988) hybrids were observed in the same lake (for all observations at Viaredssjön, Magnus Neuendorf). The pair at Kvarnsjön was seen with five young 27 April (Hans-Erik Olausson). Hybrids from Ingsbergasjön were seen in the area around Anebygård, usually from the middle of August to the middle of October, the first four autumns three birds, then two birds up to 1986, and finally, one bird up to 1991. These hybrids were also seen elsewhere in Småland, for instance at Nävelsjö (Westring 1984). These hybrids might have had their winter quarters in western Skåne, where two birds wintered in 1981-1986 (Table A6). Other sightings made up to 1994 might have been of hybrids from either Ingsbergasjön, Viaredssjön or Kvarnsjön (Table A6).
Anser canagica x Branta leucopsis A total of 15 sightings exists, but few were identified with certainty (Table A8).
Anser cygnoides x Branta canadensis One pair was seen at Säve (Boh)  Anser erythropus x Branta hutchinsii At least two different individuals (for descriptions, see Nilsson 1983Nilsson , Österblad 1987 were recorded in West Blekinge and North-east Skåne on a number of occasions during 1979(Nilsson 2000 Table A9). In all probability, these hybrids originated from the breeding of geese at Eriksberg (Bl) (Nilsson 1983). In 1983, one of these hybrids made a breeding attempt paired to a Barnacle Goose (see below).
(Anser erythropus x Branta hutchinsii) x Branta leucopsis One pair, the male a hybrid, made a breeding attempt at Karlshamn (Bl) in 1983, but none of the four eggs hatched (Nilsson 1984).
Two sites frequently used by staging hybrids during the years 2004-2006, at least by those born in Hälsingland, were Gävletravet (Gstr) in August and Hjälstaviken (Upl) from mid-August to mid-October.
Some birds staged at Hjälstaviken also in May and early June. Hybrids were occasionally seen also at other sites in Jämtland, Hälsingland, Gästrikland and Uppland. Sightings further south were mainly done in Gotland and Skåne (Table A10).
Winter quarters of these hybrids were situated in the Netherlands (Bo Fagerström, Koffijberg et al. 2005) and the Doñana area, South-west Spain (Kampe-Persson 2004).
(Anser erythropus x Branta leucopsis) x Branta canadensis One pair, the male a hybrid, bred unsuccessfully at Gussjön (Ång) in 1989 (Sjöberg 1990); none of the six eggs hatched. The pair was seen at the breeding site 11 May-20 July, and at Härnösand 20 September (Leif Johansson).  Essen 1982, Svensson et al. 1999. Several of these families wintered at Höllviken, South-west Skåne (HKP). One male from this project, paired to a Greater Canada Goose, bred at Öster-Malma (Srm) for five years, before he was removed, whereupon the female re-paired with a Greylag Goose and finally with a Canada Goose (Fabricius 1991).

Anser indicus x Branta canadensis
This hybrid type was mainly reported from Skåne, usually singly, but there were also observations of a family at Tåkern (Ög) in 1987 and 1988 (Table A11).

Anser indicus x Branta leucopsis
One pair made up of two hybrids was recorded in South-west Skåne during the years 1985-1994 (see below). After that one of the mates had died, the surviving bird continued to use the same sites, and bred paired to a Greylag Goose in 1997 (see below). This individual was last seen 15 September 2001, when more than 17 years old (HKP). Sightings of other individuals than the two afore-mentioned ones were also they mainly done in South-west Skåne (Table A12).
(Anser indicus x Branta leucopsis) x Anser anser In Skåne in 1997, one young was seen together with its parents at Fjällfotasjön 29 May-24 June, while the pair, without young, was seen at Börringesjön and Foteviken 11 July-21 October (HKP).
(Anser indicus x Branta leucopsis) x (Anser indicus x Branta leucopsis) In Skåne, one pair, first recorded at Yddingen 7 May 1985, bred unsuccessfully at Fjällfotasjön annually during the period 1986-1994 (HKP). After leaving the breeding site, the pair spent the rest of the summer and autumn at Börringe and Foteviken (HKP).
Branta bernicla x Branta leucopsis Ten sightings, all of them from the last three-year period, involved probably at least three different individuals (Table A13).