The role of demographic and life-history variables in de termining the extinction risk of populations has been studied extensively (Terborgh & Winter 1980, Pimm et al 1988; Laurance 1991), and the results have been ap plied in a revision of the threatened specits categories used by the World Conservation Union (IUCN; Mace & linde 1991; Macc & Stuart 1994). A final version of the revised categories has recently been formally adopted by The World Conservation Union (1994).
Studies comparing Red Lists of threatened animals and
plants with nonthreatened taxa have usually been re
stricted to a small region (Lahti et al. 1991; Berg et al
1994).Studies of cite biology arid disrribtrczort ofrlireat ened organism groups at a global level, which are impor tant for nature conservation, are lacking (see, however, Laurila &järvinen 1989). One reason for this is that data on the status of organisms are usually lacking in some parts of the world. Information on the status and biology of the Anatidae (ducks, geese, and swans) ts relatively good on a worldwide scale (del Hoyo et al 1992; Rose & Scott 1994). In this study the new IIJCN red list catego ries (World Conservation Union 1994) are applied to the Anatidae at the subspecific level to produce a list of the globally threatened taxa. This list is justified in detail elsewhere (Green 1996) and follows from earlier at tempts to identify the globally threatened Anatidae (Ellis Joseph et al 1992; Green 1992; callaghan & Green 1993), which used an earlier version of the proposed new IUCN criteria (Mace & Lande 1991).
Humans have long had a unique relationship with the Anatidae, reflected in their widespread harvesting for food, domestication, breeding in captivity, and their in fluence on human culture (Kear 1990). In modern times conservationists have paid particular attention to Anatidac to minimize the impact of recreational hunting (Moser et al. 1993) and to attempt to resolve increasing problems of agricultural damage (van Roomen & Madsen 1992). The Anatidae have acquired great importance as flagships for wetland conservation (Kear 1990), and concern for their
status was central to the adoption of the first global treaty for nature conservation (Matthews 1993), the Ramsar Con vention (Convention on Wetlands of Intemational Impon tance, especially as Waterfowl Habitat [Ramsar 19711).
1 consider the various threats facing the globally threat ened Anatidae in order to establish which factors are most responsible for putting Anatidae at risk of extinction. I analyze the distribution of globally threatened taxa, identifying those countries and regions where they are concentrated. Their distribution is related to that of the growing network of wetlands protected under the Ram sar Convention to establish what protection this network provides to globally threatened Anatidae. I also compare the distribution and threats facing the globally threatened
.Anatidac and thoseforallglobatlythreatened birds,using analyses recently conducted by Collar et al. (1994) for all
nonthreatened Anatidae in an effort to identify parame
ters related to extinction risk, as identified by the IUCN categories, it is well known that extinction risk is corre lated with population size, range size, extent of fragmen tation, and extent of population fluctuation, and the new 1UCN criteria are based on such parameters (Mace
& Linde 1991; World Conservation Union 1994). Thus, these parameters will obviously differ for threatened and nonthreatened taxa. Previous work has shown that there are no clear differences between globally threat ened and nonthreatened Anatidae taxa in terms of Iite history traits such as body size and cltttch size (Laurila & jarvinen 1989). This study compares parameters of dis tribution, habitat use, and migration patterns. Island birds have been found to have a high extinction risk (Moors
1985; Terborgh & Winter 1980), and I compare the sta tus of insular and noninsular Anatidae. 1 also compare the latitudinal breeding distribution of threatened and nonthreatened Anatidac to identify the consequences of the concentration of economically developed countries and of arctic and subarctic land masses (exploited for breeding by migratory Anatidae) in the northern hemi sphere Anatictac arc recorded in a wide range of habitat
typcs, and diffcrent ratcs of destruction sean likcly to make taxa using certain habitats morc prone to cxtinc tion than others. Thus, I compare the habitat usc of thrcatcncd and nonthrcatcned taxa and consider thc sig nificance of habitat loss as a threat. ft might be predicted that migratory taxa are more likcly to bc thrcatcned bc cause they arc dependent on different areas during thc brecding scason, wintcr, and migration. A thrcat opcrat ing in just onc of these areas could have a severe impact (SalathE 1991). Therefore, Iconipare the status of migra tory and nonmigratory Anatidae. 1 am aware of only one previous study in which migrant or rcsident status has been related to extinction risk in birds (Pimm ct aL
1988). The results of these analyses have important mp
licationsfor waterbird and wetland conservation pro grams and will be influential in shaping an IUCN action plan for the Anatidae, which is currently being prepared
by the Wetlands international and The Wildfowl & Wet
Methods The new ILJCN criteria (World Conservanon tJnLon 1994)
were applied to aD Anatidae at the subspecies level (Green
1996). The members of the Anatidac family considcred follow the taxonomy of Morony et al (1975). The list of subspecies considered is that of Madge and Burn (1988) and del Hoyo et al. (1992). The three globally threatened categories are, in order of increasing risk of extinction, vulnerable (VU), endangered (EN), and critically endan gcred ofthe thrce threatened categories are categorized as lower risk (LR).
Data on distribution, migratory status, and habitat use patterns for all Anatidae tan were taken from Madge and Burn (1988), Marchant and Higgins (1990), del Hoyo et al. (1992), Green (1992), and Callaghan and Green (1993). Rarnsar site directories (Ramsar 1990; jones 1993) together with access to the Ranasar site database held by the Wet lands International at Slimbridge, United Kingdom, iden tified which threatened taxa are recorded on the 685 wetlands in the world that are protected by the Ramsar Convention.
I divided taxa into those that are migratory and nonmi
gratory, using a broad definition of migratory which in cluded aLl those taxa that regularly undergo movements across nationaL frontiers, incLuding partiaL migrants. Taxa
making predictablc, long-distance movements between brecding and wintcring grounds that do not cross na tional frontiers (c.g thc Alcutian Canada Goosc (Branta canactensis teucopareia])were aLso included. Taxa that have both migratory and nonmigratory populations were categorized as migratory. Taxa wcre divided into those rcstrrcted to islands and those not restrictcd to islands. I uscd two size classcs of islands: smaD islands with an arca of up to 20,000 km2 (including the Falkland islands [to
United Kingdomj and the Kerguelen islands [to FranceI)
and large islands with an area of between 20,000 km2 and 110001000 km2 [Cereopsisnovaeholtandiaegriseajand thc Alcu tian Canada Goosc), wcrc regarded as smaD-island taxa.
In analysts of habitat use I uscd the following broad
categories of habitat type: forcst (including woodland and mangroves); marine cnvironment
The thrcats assigned to each taxon arc those consid ered the major causes o its globally threatened status (Green 1992; Callaghan & Grecn 1993; Grccn 1996) and do not give a complete list of all thosc negative factors that are impinging on the taxon. The various threats are divided into the categories used by Collar et al (1994) for all threatened bird specics, with thc cxccption that a threat of hybridization duc to the introduction of a closely related taxon is treated separatcly from othcr cf fects of exotic introductions (of predators, compctitors, introduced plants, etc).“Habitat loss” includes all kinds of habitat alteration apart from pollution. “Hunting” in cludes egg-collecting, accidental trapping, and other kinds of persecution. “Small range” includcs small popu lation and is only listed as a threat when it is thc princi pal rcason thc taxon qualifies for globally threatened sta tus “Pollution” incltides poisoning from pesticides or other chcmicals Statistical analyses were conducted using x2 tcsts on contingency tables, joining some categories in habitat analyses in order to prevent cxccssively low ex pected frequencies (Siegel & Castellan 1988).