CHAPTER 7 Procellariform extinctions in the Holocene: threat processes and wider ecosystem-scale implications R. Paul Scofield 7.1 Introduction Procellariiformes are the order of tube-nosed seabirds that includes the albatrosses, petrels, and shearwaters. Members of the order breed mainly on islands; individual species are often extremely widely distributed, with populations on many islands within many island groups, and frequently in different oceans. The order has a global distribution, and contains four extant families: Pelecanoididae (diving-petrels), Procellariidae (shearwaters, petrels, and fulmars), Diomedeidae (albatrosses), and Hydrobatidae (storm-petrels), together with either one or two extinct families. Species are generally either pelagic scavengers primarily taking squid and fish, or are planktivorous. As basal members of the Neoaves, the group has a physiologically restrained breeding system in which birds only produce a single egg, usually annually, with no relaying, do not have complex nest structures, breed either at or below ground level, and generally have a tightly constrained breeding season. These traits leave them open to predation by mammalian predators, and it is hypothesized that for this reason their breeding biology is characterized by a number of behaviours that have evolved to reduce the effect of such predation. On islands where petrels are without predators they occur in huge numbers; indeed, one of the largest concentrations of breeding animals in the world is a petrel colony (Reyes-Arriagada et al. 2007). Although these vast colonies are susceptible to the invasion of exotic mammals, and can disappear rapidly (see Atkinson 1985 for examples), procellariiform populations also typically spawn offshoots of a few hundred individuals on tiny offshore islands or stacks, and when invasions occur it is rare for populations to disappear entirely. Furthermore, a majority of petrel species arrive nocturnally on their breeding grounds, and burrow-nesting and extensive habitat gardening are the norm. Despite being susceptible to mammalian predation, the procellariiforms are an ancient group that has survived comparatively unchanged since the earliest Cenozoic. Although procellariiform communities are dynamic over time (Warheit 2002), responding to environmental changes, mammalian invasions and the emergence of new islands, the taxa themselves have shown remarkable resilience. For example, the short-tailed albatross, Phoebastria albatrus, which is rare but still extant today, has managed to survive since at least the Pliocene (Olson and Rasmussen 2001) and move from the Atlantic, where it is now extinct (Olson and Hearty 2003), to the Pacific. Here I will attempt to determine reasons for the group’s longevity, and discuss their ecosystem impacts and extinction drivers in the Holocene. 7.2 The earliest fossil record of procellariiformes 7.2.1 The Tertiary record (and earlier) Fossil evidence suggests that procellariiforms have survived comparatively unchanged since the 151 07-Turvey-Chap07.indd 151 12/5/2008 12:21:47 PM 152 Perhaps Para here? HOLOCENE EX TINCTIONS earliest Cenozoic, and there is limited (and somewhat contentious) evidence of procellariiformlike birds at the end of the Mesozoic. The late Maastrichtian Lance Creek Formation of Wyoming contains two species of the genus Lonchodytes described by Brodkorb (1963a). Hope (2002) argued that Lonchodytes is likely to be a procellariiform (most closely resembling procellariids) although this has been questioned (Scofield et al. 2006). Of a similar age, and equally enigmatic, is a fragmentary clavicle from the Nemegt Basin in Mongolia, which Kurochkin (1995, 2000) assigned to the Diomedeidae. Olson and Parris (1987) described Tytthostonyx glauconiticus from the Late Cretaceous or early Paleocene of New Jersey, placing it in its own family (Tytthostonichidae) and suggesting that it appeared to be either a basal procellariiform or close to the origin of the Fregatidae and the Pelecaniformes, the later opinion with which Feduccia (1996) concurred. Eopuffinus kazachstanensis, a species assigned to the Procellaridae, is known from the Paleocene of Kazakhstan, although it is currently considered of uncertain status. The earliest unequivocal procellariiform appears to come from the Late Eocene of Louisiana. A distal tibiotarsus has been accepted as morphologically close to the extant genus Pterodroma in the Procellaridae (Feduccia and McPherson 1993). Procellariiforms and petrel-like taxa are known from Eocene deposits from Uzbekistan, Louisiana, and perhaps the London Clays (but see Mayr et al. 2002). Murunkus subitus, based on a carpometacarpus from the mid Eocene of Uzbekistan, has been placed in the Diomedeidae (Panteleyev and Nessov 1987), as has Manu antiquus from the mid to Late Oligocene of Otago, New Zealand (Marples 1946), although both assignments are tentative. Another extinct procellariiform family, the Diomedeoididae Fischer, 1985, has recently been described from the Oligocene of Germany, and subsequent work by Mayr et al. (2002) has shown this group to indeed warrant family status and include a number of species originally described as albatrosses. Procellariidae are reliably reported in the Oligocene. ‘Larus’ raemdonckii from the early Oligocene (Rupelian) of Belgium was placed in the extant genus Puffinus (Procellariidae) by Brodkorb (1963b). Three extinct Diomedea species (Diomedeidae) are known from the North Pacific Neogene, and several more indeterminate Tertiary records exist for the genus (Chandler 1990). The Upper Miocene of Australia has produced a Diomedea species described from an unguis (Wilkinson 1969). The fossil record of Hydrobatidae and Pelecanoididae is much younger. The first stormpetrel is found in the upper Miocene of California (Olson 1985c), while Scofield et al. (2006) and Worthy et al. (2007) have recently described diving-petrels differing little from modern species from the Miocene of southern New Zealand. A diving-petrel has also been described from the early Pliocene of South Africa (Olson 1985c). 7.2.2 Pre-human petrel extinctions In Table 7.1, I summarize a list of published procellariiform taxa that became extinct in the Pliocene and Pleistocene. While undoubtedly incomplete, it is indicative of comparatively low levels of extinction in the group. Our understanding of extirpations of extant species populations is more incomplete for this interval; however, the Pleistocene fossil record for the Palearctic has been comprehensively surveyed by Tyrberg (1998), and the Quaternary in the Mediterranean Region was examined by became Sánchez Marco (2004), and similarly low levels of population extinction have been recorded (Table 7.2). Thus, from what is currently known about the avian fossil record, it would seem that comparatively few species of procellariiforms have become extinct in the last 4 million years, before humans began to impact global ecosystems. However, it is worth noting that, in being primarily pelagic, the bones of procellariiforms will usually only be preserved when birds return to the land to nest. Even then conditions have to be suitable for preservation to occur, as is evident from the fact that the majority of Pleistocene procellariiform fossils are known from limestone or karstic environments. Where these conditions are absent, fossil seabirds are rare. Many modern-day petrel colonies occur on peat soils, which are often extremely acidic because petrels acidify the soil though the nitrification of their guano. Petrel biology therefore typically ensures the destruction of their bones. The outcome of this until 07-Turvey-Chap07.indd 152 12/5/2008 12:21:47 PM HOLOCENE PROCELL ARIFORM EX TINCTIONS 153 Table 7.1 Procellariiform species that became extinct in the Pliocene and Pleistocene. Species Locality Reference Calonectris krantzi Phoebastria anglica Phoebastria rexsularum Pterodroma kurodai Pterodromoides minoricensis Lee Creek (USA) UK; USA Lee Creek (USA) Aldabra (Indian Ocean) Menorca; Lee Creek (USA) Puffinus nestori Puffinus pacificoides Puffinus tedfordi Puffinus nestori Ibiza (Mediterranean) St. Helena (South Atlantic) Western North America Ibiza (Mediterranean) Olson and Rasmussen (2001) Olson and Rasmussen (2001) Olson and Rasmussen (2001) Harrison and Walker (1978) Olson and Rasmussen (2001); Seguí et al. (2001) Alcover (1989) Olson (1975) Howard (1971) Alcover (1989) Table 7.2 Procellariiform populations that have been documented as becoming extinct in the Pliocene and Pleistocene. Species Locality Reference Calonectris diomedea diomedea Phoebastria cf. albatrus Bermuda Bermuda; Lee Creek (USA) Phoebastria aff. immutabilis Phoebastria aff. nigripes Puffinus mauretanicus Lee Creek (USA) Lee Creek (USA) Pityusic Islands (Mediterranean) Olson et al. (2005b) Olson and Rasmussen (2001); Olson and Hearty (2003) Olson and Rasmussen (2001) Olson and Rasmussen (2001) Alcover (1989) cultural putative is that our understanding of the extinction of many Neogene procellariiform species (e.g. Puffinus felthami and Puffinus kanakoffi, described from the Pliocene of California; Chandler 1990) remains incomplete, and these species could well have survived considerably later into the human-impacted Late Pleistocene. For these reasons, it is possible that apparently Pleistocene (and even Pliocene) extinctions were in fact human-induced. 7.3 An unparalleled series of extinctions? The Holocene fossil record of the Procellariiformes is considerably more complete than that of the Pliocene and Pleistocene. However, many seemingly geologically young sites are not reliably dated, and may actually be pre-Holocene in age. Some subfossil dune deposits dated as Holocene may include birds that have been forced ashore by storms or 07-Turvey-Chap07.indd 153 by vagrancy, and so these records may not actually indicate breeding populations. Furthermore, archaeologists frequently interpret the presence of species in middens to indicate nearby breeding sites, but the occurrence of species known to be high-latitude specialists in tropical sites (Steadman 2006b) indicates that that some indigenous populations may have exploited vagrant or migratory populations. Nevertheless, there is strong evidence that species-level extinctions have occurred more often in the Holocene than is documented in the Pliocene and Pleistocene (see Chapter 4 in this volume), and 56% (76 of 136; Onley and Scofield 2007) of Holocene procellariiform species have lost populations (Table 7.3). ‘In terms of its rate and geographical extent, its potential for synergistic disruption and the scope of its evolutionary consequences, the current mass invasion event is without precedent and should be regarded as a unique form of global change’ 12/5/2008 12:21:47 PM 07-Turvey-Chap07.indd 154 Table 7.3 Procellariiform populations that have become extinct in the Holocene (some of these populations have since reintroduced themselves). Species Locality Likely cause of extinction1 Pre- or postEuropean? Reference Diomedea exulans Diomedea sanfordi Phoebastria albatrus Macquarie Island Pitt Island (Chatham Group) Agincourt Island and Pescadore Islands (Taiwan); Izu, Bonin, Daito, Senkaku and western volcanic groups of Japan Johnston, Marcus, and Wake Islands, Izu Islands Johnston, Marcus, Volcano, Wake, and Marshall Islands, and Northern Marianas North Island (New Zealand) Southern New Zealand Macquarie Island North Island (New Zealand), Amsterdam Island 1 1 1 Post Pre Post de la Mare and Kerry (1994) Millener (1999) Hasegawa and DeGange (1982) 1 1 Post Post Rice and Kenyon (1962) Rice and Kenyon (1962) 1 1 1 1 Pre Post Post Pre/post 1 1 Post Pre/post Worthy and Holdaway (2002) Worthy and Holdaway (2002) Clarke and Schulz (2005) Worthy and Jouventin (1999); Worthy and Holdaway (2002) Imber (1994); Worthy and Holdaway (2002) Megyesi and O’Daniel (1997); Steadman (2006b) 1 1 1 1 1 1 1 1 1 1 1 1 1 Post Post Pre Pre Post Pre Post Pre/post Post Post Pre Post Post 1 1 1 Post Post Pre/post Phoebastria immutabilis Phoebastria nigripes Thalassarche bulleri Macronectes halli Halobaena caerulea Pachyptila turtur Pachyptila vittata Bulweria bulwerii Pseudobulweria aterrima Pseudobulweria becki Pseudobulweria rostrata Pterodroma alba Pterodroma arminjoniana Pterodroma axillaris Pterodroma baraui Pterodroma brevipes Pterodroma cahow Pterodroma cervicalis Pterodroma cookii Pterodroma defilippiana Pterodroma cf. feae Pterodroma gouldi Pterodroma hasitata Pterodroma hypoleuca Southern New Zealand, Main Chatham Island Main Hawaiian Islands, Midway Atoll, and many southeast Pacific Islands; islands off China; Tenerife and islands off Lanzarote (Canary Islands) Amsterdam Island ?Solomon Islands Ofu (Samoa), Aitutaki, Huahine Huahine; Ua Huka Amsterdam Island Pitt Mangere Island (Chatham Islands) Amsterdam Island Many central Pacific Islands Bermuda Raoul Island (Kermadec Islands) North Island (New Zealand) Isla Robinson Crusoe (Juan Fernández Group) Scotland, The Netherlands, Denmark, Sweden 12/5/2008 12:21:47 PM North Island (New Zealand) Guadeloupe, Martinique Main Hawaiian Island and islands of north-west chain (including Kure) Worthy and Jouventin (1999) BirdLife International (2007a) Steadman (2006b) Steadman (2006b) Worthy and Jouventin (1999) Tennyson and Millener (1994) Worthy and Jouventin (1999) BirdLife International (2007b) Olson et al. (2005b) Veitch et al. (2004) Worthy and Holdaway (2002) Brooke (1987) Lepiksaar (1958); Ericson and Tyrberg (2004); Leopold (2005); Serjeantson (2005) Worthy and Jouventin (1999) Bent (1922) Kepler (1967); Olson and James (1982) 07-Turvey-Chap07.indd 155 Pterodroma inexpectata Pterodroma lessonii Pterodroma macroptera Pterodroma cf. madeira Pterodroma mollis New Zealand mainland Campbell Island Amsterdam Island El Hierro (Canary Islands) Amsterdam Island, Macquarie Island 1 1 1 1 1 Pre Post Post Post Post Pterodroma neglecta Raoul Island (Kermadec Islands); Isla Robinson Crusoe (Juan Fernandez Group) Raoul Island (Kermadec Islands), Henderson Island, many central Pacific Islands Easter Island Lord Howe Island, Norfolk Island Oahu and other Hawai’ian Islands Norfolk Island Tahuata, Easter Island Amsterdam Island New Zealand mainland Northern South Island (New Zealand) Giraglia Island (Mediterranean) Islands off southern Japan and south-east Russia Amsterdam Island San Benedicto Island (Mexico) Lifuka and Hu’ano (Tonga), Ofu (Samoa), Henderson and many southeast Pacific Islands Main islands of Bonin Group North Island and some offshore islands (New Zealand) Amsterdam Island Chatham Island North Island and northern South Island (New Zealand) New Zealand mainland New Zealand mainland, Main Chatham Island Northern South Island (New Zealand) Bermuda Cabrera, Formentera (Balearics) Ogasawara Island, Marcus Island, Wake Island, Henderson Island, Pitcairn Island, and some Marquesas Islands O’ahu, Maui, and Laana’i (Hawaii) Guadalupe Island (Mexico) 1 Post Worthy and Holdaway (2002) Taylor (2000) Worthy and Jouventin (1999) Rando (2002) Worthy and Jouventin (1999); Clarke and Schulz (2005) Veitch et al. (2004) 1 Post Wragg and Weisler (1994); Steadman (2006b) 1 1, 2? 1 1 1 1 1 1 2 2 1 3 1 Pre Post Pre Pre Post Pre Pre Post Pre/post Post Post Pre Steadman (2006b) Meredith (1991); Holdaway and Anderson (2001) Olson and James (1982) Medway (2002) Steadman (2006b) Worthy and Jouventin (1999) Worthy and Holdaway (2002) Worthy and Holdaway (2002) Thibault and Bretagnolle (1998) Oka (2004) Worthy and Jouventin (1999) Jehl and Parkes (1982) Wragg and Weisler (1994); Steadman (2006b) 1 1 1 1 1 1 1 1 1 1 1 Post Pre Post Pre Pre Pre Pre Pre Pre Post Pre/post 1 1 Post Post Pterodroma nigripennis Pterodroma phaeopygia Pterodroma ?pycrofti Pterodroma sandwichensis Pterodroma solandri Pterodroma ultima Procellaria cinerea Procellaria parkinsoni Procellaria westlandica Calonectris diomedea Calonectris leucomelas Puffinus assimilis Puffinus auricularis Puffinus bailloni Puffinus bannermani Puffinus bulleri Puffinus carneipes Puffinus elegans Puffinus gavia Puffinus gravis Puffinus griseus Puffinus huttoni Puffinus lherminieri Puffinus mauretanicus Puffinus nativitatis 12/5/2008 12:21:47 PM Puffinus newelli Puffinus opisthomelas Hasegawa (1991) Worthy and Holdaway (2002) Worthy and Jouventin (1999) Millener (1999) Worthy and Holdaway (2002) Worthy and Holdaway (2002) Imber (1994); Worthy and Holdaway (2002) Worthy and Holdaway (2002) Olson et al. (2005b) Oro et al. (2004) Wragg and Weisler (1994); Seto (2001); Steadman (2006b) Ainley et al. (1997) Everett and Anderson (1991) 07-Turvey-Chap07.indd 156 Table 7.3 Continued Species Locality Likely cause of extinction1 Pre- or postEuropean? Reference Puffinus pacificus Puffinus puffinus Puffinus yelkouan Pelecanoides georgicus Pelecanoides urinatrix Main Hawaiian Islands, Tonga and many southeast Pacific Islands Bermuda; many European islands Mediterranean Islands Macquarie Island Macquarie Island, Chatham Island, New Zealand mainland, Amsterdam Island Main Chatham Islands Amsterdam Island Henderson, Ua Huka, Tahuata Rapa Island ‘Eua (Tonga), Samoa, Mangaia, Tahuata, Easter Island, Henderson Island Many islands and one mainland site around coast of Europe Main islands of Azores, Canary Islands and Hawai’i Central Kuril Island (Russia); some islands in north-western California Many islands and one mainland site around coast of North America and Ireland Many Californian Islands Numerous islands in Gulf of California Midway Island, Izu Islands 1 1 1 1 1 Post Post Post Post Pre/post 1 1 1 1 1 Post Post Pre Post Pre Olson and James (1982); Steadman (2006b) Brooke (1990) Martin et al. (2000) Clarke and Schulz (2005) Worthy and Jouventin (1999); Worthy and Holdaway (2002); Clarke and Schulz (2005) Imber (1994) Worthy and Jouventin (1999) Steadman (2006b) Murphy and Snyder (1952) Wragg and Weisler (1994); Steadman (2006b) 1 1 1 Post Pre Post Martin et al. (2000) Olson and James (1982); Rando (2002) Boersma and Silva (2001) 1 Post 1 1 1 Pre and post Post Post Podolsky and Kress (1989); Huntington et al. (1996) Everett and Anderson (1991) Donlan et al. (2000) Hasegawa (1984); Baker et al. (1997) Garrodia nereis Pelagodroma marina Fregetta grallaria Fregetta tropica Nesofregetta fuliginosa Hydrobates pelagicus Oceanodroma castro Oceanodroma furcata Oceanodroma leucorhoa Oceanodroma melania Oceanodroma microsoma Oceanodroma tristrami 1 Key to likely causes of extinction: 1, introduction of predators; 2, hunting; 3, volcanic eruption. 12/5/2008 12:21:47 PM of neccesity HOLOCENE PROCELL ARIFORM EX TINCTIONS (Ricciardi 2007). Being burrowing species with low fecundity, procellariiform populations have evolved in the absence of mammalian predators. The arrival of humans on oceanic islands during the Late Pleistocene and Holocene precipitated a wave of extinctions among birds, especially seabirds, caused largely by the introduction of exotic mammals (see also Chapter 2 in this volume). The magnitude of this extinction event varies markedly between islands, and correlates of recent bird extinctions are becoming increasingly understood (e.g. see Blackburn et al. 2004; Duncan and Forsyth 2006; Chapter 12 in this volume). But what are the mechanisms driving procellariiform extinctions? 7.4 Human-induced petrel extinction 7.4.1 Hunting versus introduced predators Evidence for the effect of hunting by prehistoric settlers being responsible for local or global extinction of petrel species is limited. MourerChauviré and Antunes (2000) found the bones of 157 the extinct shearwater Puffinus holeae associated with Upper Pleistocene Neanderthal middens in Portugal, and Rando and Alcover (2007) found cut marks and burning on bones of the extinct lava shearwater Puffinus olsoni on the Canary Islands (Fig. 7.1). However, while these observations indicate that at least some now-extinct procellariiforms were actively persecuted by humans before the historical period, they provide little information on the intensity of anthropogenic persecution or whether it represented a significant factor in the disappearance of these species. It is also rarely possible to tell from the fossil record how quickly petrel populations disappeared, although some evidence can be gained from St. Helena (discovered in 1502 and first described in written accounts in 1588), where the St. Helena petrel, Pseudobulweria rupinarum, became extinct shortly after human arrival. Here the species was extinct or very rare before any documented record could be made of it (Olson 1975). In another more recent case feral cats, Felis catus, probably reached Figure 7.1 Burnt and cut bones of extinct Puffinus olsoni from the Canary Islands, indicating that the species experienced prehistoric exploitation by humans. Scale bar = 2 cm. Courtesy of J.C. Rando and J.A. Alcover. 07-Turvey-Chap07.indd 157 12/5/2008 12:21:47 PM 158 HOLOCENE EX TINCTIONS Little Barrier Island, off New Zealand’s North Island, in about 1870. By the 1980s the population of Parkinson’s petrel, Pterodroma parkinsoni, was functionally extinct (Veitch 1999). 7.4.2 Introduction of predators 7.4.2.1 Rats Invasive rats are some of the largest contributors to seabird extinction and endangerment worldwide (Jones et al. 2008a). The most problematic species of the genus Rattus are native to Asia. Rats spread out of Asia at times of major human diaspora, with the Pacific rat or kiore, Rattus exulans, being spread through the Pacific by Melanesian and Polynesian peoples in the Late Pleistocene and Holocene (Spennemann 1997; Matisoo-Smith et al. 1998). The black rat, Rattus rattus, reached Europe in Roman times while the brown or Norway rat, Rattus norvegicus, arrived in the Middle Ages (Kurtén 1968). Thus, although rats are generally thought of as ubiquitous, their presence is a relatively new phenomenon worldwide. Being commensal with humans, rats generally accompany humans accidentally wherever they settle, including islands representing important sanctuaries for procellariiforms. The introduction of rats to island communities is frequently devastating. For example, the accidental introduction of black rats to Big South Cape Island (Taukihepa) (Atkinson and Bell 1973) led to a plague of rats, the rapid extinction of three species of landbird, and the decline to virtual elimination of two seabird species within 5 years of their introduction (Bell 1978). Among predators, R. rattus and to a lesser extent R. norvegicus are well known as the leading agents of bird extinction on islands, but the effects of R. exulans introduced by Polynesians throughout Remote Oceania is now thought to be significant, especially on seabirds (Steadman 2006b). Whereas the larger rat species kill the adults of smaller species of petrel, all species kill chicks and eggs. Modern techniques such as stable-isotope analyses (e.g. Hobson et al. 1999; Stapp 2002) have not only confirmed that rats do actually feed on seabirds, but have also suggested that the traditional studies of rat stomach contents were underestimating the importance of seabirds in the diet of rats. One study looking at 07-Turvey-Chap07.indd 158 the relative effects of predation by rats on species of differing size showed that small species of petrel are most susceptible to egg and adult predation whereas larger species such as Cory’s shearwater, Calonectris diomedea, are more likely to be affected by chick predation (Igual et al. 2006). These studies indicate that our knowledge of the way rats impact island populations is limited and that more work is needed. There can be no doubt, however, that rats severely impact seabirds, reducing their populations and in many cases triggering their local extinction (Atkinson 1985). Wherever archeological evidence is examined in the Pacific, extinctions begin with the introduction of kiore, and, whether prehistoric or recent, seabird extinctions on islands normally occur very shortly after the introduction of rats. 7.4.2.2 Mice The ubiquitous, commensal house mouse, Mus musculus, is the most widely introduced of all mammals, but its effect on native biota is poorly known. On Gough Island in the South Atlantic, unlike most other sub-Antarctic islands, the house mouse is the only introduced mammal. Mice were known to affect populations of the small storm-petrels and diving-petrels but were thought to pose little population level risk to large seabirds. However, recent video evidence has shown house mice killing chicks of the Tristan albatross, Diomedea dabbenena, and Atlantic petrel, Pterodroma incerta, and has indicated that mouse-induced mortality is a significant cause of poor breeding success in these species. Population models show that the levels of predation recorded there are sufficient to cause population decreases (Wanless et al. 2007). It is suggested that mice may not be a significant issue to larger seabirds on islands when constrained by other introduced predators, but when these mouse populations are released from the ecological effects of predators and competitors by eradication and restoration programmes, they too may become predatory on seabird chicks (see section 7.4.2.9, below). an other 7.4.2.3 Cats Feral cats, F. catus, have a devastating effect on seabird colonies. Researchers have estimated that cat 12/5/2008 12:21:48 PM HOLOCENE PROCELL ARIFORM EX TINCTIONS mortality on Marion Island prior to rat eradication was 450 000 seabirds per annum (van Aarde 1980), whereas on Macquarie Island 47 000 broad-billed prions, Pachyptila vittata, and 110 000 white-headed petrels, Pterodroma lessonii, were estimated to be killed annually (Jones 1977), and on Kerguelen Island 1.2 million seabirds were estimated to be killed annually (Pascal 1980). This level of mortality has dramatic effects on the population viability of seabird breeding colonies. For example, cats have been responsible for the local extinction of 10 petrel species on the Crozet Islands in the last century (Derenne and Mougin 1976), whereas in New Zealand the extinction of most seabirds from main Chatham Island has been attributed to cats (Imber 1994). However, the effects of cats are not always predictable, and their removal may upset delicate (though artificial) ecosystems that have developed on some islands (see section 7.4.2.9, below). 7.4.2.4 Foxes More than 10 million seabirds belonging to 29 species formerly bred on the Aleutian Islands off Alaska (Croll et al. 2005). The region lacked large terrestrial predators prior to the arrival of humans, but in the past 5000 years Arctic foxes, Alopex lagopus, have been introduced to at least 400 of these islands. Being primarily carnivorous, foxes preyed on the local seabirds that had evolved in the absence of predators, and only those species that nested on unreachable cliff faces were able to survive. Burrownesting species and surface-nesters (e.g. gulls) were predated easily, and their populations became locally or regionally extirpated. Fox removal from 40 islands has resulted in an increase by two orders of magnitude in whiskered auklet, Aethia pygmaea, and other seabird populations. Similarly, nine foxfree islands have nearly 100 times more seabirds than nine geographically and ecologically similar but fox-infested islands (Croll et al. 2005). 7.4.2.5 Pigs Pigs, Sus scrofa, are a frequently ignored but extremely important predator of burrowing petrels. They were responsible for the eradication of Buller’s petrel, Puffinus bulleri, on one of the two main islands in The Poor Knights group off northern New Zealand (Medway 2001), they have been 07-Turvey-Chap07.indd 159 159 implicated in the decline of Galapagos petrel, Pterodroma phaeopygia, in the Galapagos Islands (Cruz and Cruz 1987), and they have been responsible for the degradation of habitat on the main Auckland Islands and (with cats) caused the extinction of most breeding seabirds in this island group (Challies 1975). on the large main 7.4.2.6 Ungulates island The indirect effects on ungulates on habitat are discussed below. Livestock on some isolated islands may not only trample burrows, but also exhibit unusual behaviours caused by an absence of nutrients that can directly affect seabird survival. In particular, they may eat chicks and/or fledglings in order to ingest nutrient-rich bone. On Foula in the Shetland Islands, sheep, Ovis aries, have been observed biting off the legs, wings, or heads of unfledged young Arctic tern, Sterna paradisaea, and Arctic skua, Stercorarius parasiticus, chicks, and on Rhum in the Inner Hebrides, red deer, Cervus elaphus, have been seen biting the heads off chicks of Manx shearwater, Puffinus puffinus, and occasionally also chewing the chicks’ legs and wings (Furness 1988). 7.4.2.7 Mongooses The mongoose, Herpestes auropunctatus, was introduced to many tropical islands to control rats and other species in sugar cane plantations. This has severely affected many ground-dwelling species and especially seabirds. For example, in some years more than 60% of all egg and chick mortality in the Hawai’ian petrel, Pterodroma sandwichensis, on Oahu is caused by cats and mongooses. Although rats also prey on P. sandwichensis eggs, the major threat that they pose is providing a prey base for these larger exotic predators. The lower limit of the Hawai’ian petrel’s breeding range on Maui now coincides with the upper limit of permanent mongoose infestation. In contrast, the species has a high breeding success rate on Kaua’i, where mongoose are not established (Simons 1983). 7.4.2.8 Snakes The accidental introduction of the brown tree snake, Boiga irregularis, to Guam around 1950 induced a cascade of extirpations that may be unprecedented among historical extinction events in taxonomic 12/5/2008 12:21:48 PM 160 HOLOCENE EX TINCTIONS scope and severity. Birds (including wedge-tailed shearwater, Puffinus pacificus), bats, and reptiles were affected, and by 1990 most forested areas on Guam retained only three native vertebrates, all of which were small lizards (Fritts and Rodda 1998). It is clear that some petrels and shearwaters can coexist with snake species (e.g. tiger snakes, Notechis spp., with wedge-tailed shearwater and short-tailed shearwater, Puffinus tenuirostris, in Australia), so it would appear that it is the invasive nature of the Boiga population and the Guam fauna’s lack of adaptation to snake predation that has caused extirpation. goats, Capra aegagrus hircus, are very destructive to vegetation; by browsing on seedlings they slow or even halt the regeneration of the forest canopy and reduce native plant diversity. Cattle in particular can damage and destroy burrows and consolidate soils, and have been implicated in the decline of the Galapagos petrel (Cruz and Cruz 1987). Goats are present on at least nine island groups in the Pacific and four in the sub-Antarctic. Wherever they are present in dense populations they cause great destruction to vegetation and landscapes, typically compounded by subsequent soil erosion, which often results in total habitat loss. Forested environments can therefore be converted into degraded grasslands (e.g. on Isabela Island, Galapagos) or become more vulnerable to further invasion by weeds or to cyclone damage, all of which can adversely affect burrowing seabirds as well as wider island ecology. 7.4.2.9 Mesopredator release The example of Little Barrier Island, in New Zealand’s northern Hauraki Gulf, demonstrates the complexity of the delicate (though artificial) ecosystems that have developed on some islands. Cats were introduced in the 1870s to an island that previously only hosted an introduced population 7.4.4 Fisheries interactions of kiore. Contrary to previous expectations that cats would preferentially forage on smaller prey, Fishery waste is unquestionably an important food cats virtually exterminated not the much smaller source for seabirds, with about 6 million birds Cook’s petrel, Pterodroma cooki, but the large including procellariiforms being supported by this Parkinson’s petrel from the island. It has been sugresource in the Baltic Sea alone (Garthe and Scherp gested that the timing of Parkinson’s petrel chick 2003). An interesting and unexpected outcome of emergence from their burrows, during a period such high levels of food availability is that numbers of low food availability for cats, led to their disof generalist gulls, Larus, have reached unnaturally proportionate decline (Veitch 1999). Cook’s petrel high population levels, and exclude more specialsurvived in large numbers on the island until cats ized seabird species (e.g. Manx shearwater) from were eradicated in 1990. However, the initial eradiresources such as limited nesting grounds in these cation of cats on Little Barrier Island led not to an highly populated areas (Garthe and Scherp 2003). increase but a decrease in Cook’s petrel, due to the However, fisheries interactions are not always negareduced breeding success of these small petrels tive; northern fulmar, Fulmarus glacialis, was hiswhich are vulnerable to predation by kiore (Rayner torically restricted to Arctic regions, but a gradual et al. 2007). What led to this apparently counter-insouthward expansion begun during the mid-1700s tuitive situation? Elimination of the top introduced from Iceland through the Faeroes, Shetland and predators from islands can, in fact, lead to the Orkney, and down the British and Irish coasts to decline of smaller prey populations throughcomma the the Channel Islands and France (Fisher 1952). This ecological release of smaller introduced predators, expansion has been postulated to have resulted in a process termed mesopredator release (Crooks from increased availability of offal from whaling, and Soulé 1999). and more recently due to fisheries discards (Fisher and Lockley 1954), although other explanations have also been suggested; some researchers have 7.4.3 Habitat destruction suggested that a behavioural or genetic transformation may have occurred and a colonizing phenoThe indirect effects of ungulates on burrowing pettype emerged among the boreal fulmar population rels are poorly documented. Cattle, Bos taurus, and 07-Turvey-Chap07.indd 160 12/5/2008 12:21:48 PM HOLOCENE PROCELL ARIFORM EX TINCTIONS 161 space which was able to spread into lower latitudes (Wynne-Edwards 1962), or that a gradual change in sea temperature or oceanographic conditions may have taken place (Salomonsen 1965). Evidence now suggests that the spatial overlap between fulmars and commercial fisheries is far from complete, and while it is indisputable that northern fulmars are major consumers of fishery waste in the southern part of their range, the extent to which their distribution is or was constrained by the availability of this resource is debatable (Phillips et al. 1999). The real question arises about what will happen when supplementary fisheries offal is inevitably removed by overfishing. Will fulmar populations stay high? 7.4.5 Direct fisheries mortality The recent decline in procellariiforms in the Southern Ocean is thought to be largely driven by birds getting accidentally caught on fishing long lines or tangled in trawling gear. It is believed that the high levels of mortality offset any population increases accrued by the ready availability of fishery waste. The magnitude of by-catch mortality is exemplified by black-browed albatross, Thalassarche melanophris, the IUCN Red List status of which rose from Least Concern in 1998 to Near Threatened in 2000, Vulnerable in 2002, and Endangered in 2003, an increase greater than virtually any other bird species. Many different methods have been developed to mitigate fisheries by-catch, but these can only be enforced rigorously within Exclusive Economic Zones of individual countries, or within areas designated by international agreements such as the Commission for the Conservation of Antarctic Marine Living Resources. Implementation in international waters, where many of these pelagic species feed primarily, is currently politically difficult. example, show that there have been consistent declines in oil release into the southern North Sea in recent years (Furness and Camphuysen 1997). As predators high in marine food webs, procellariiforms can also be affected by pollutants that accumulate at higher trophic levels. Recent work on mercury in seabirds has permitted an analysis of spatial patterns and of the rates of increase in mercury contamination of ecosystems over the last 150 years, since mercury concentrations in feathers of museum specimens can be used to assess contamination in the birds when they were alive. This work has led to hypotheses about the cause and effect of these levels of pollutants, but specific mechanisms and pathways remain unclear. Surprisingly, pelagic procellariiforms show higher pollutant loads than coastal seabirds, and increases have been greatest in seabirds feeding on mesopelagic prey (Furness and Camphuysen 1997). Floating plastic is a pollutant of considerable concern, especially in the Pacific, where voluntary ingestion in surface-feeding birds may occur due to floating particles of plastic being confused with prey items, or through plastic already being incorporated within the bodies of prey species. Plastic is often passed from parents to chicks in regurgitated food (Blight and Burger 1997). The effects of this pollution on procellariiforms is poorly understood, but assimilation of polychlorinated biphenyls (PCBs) is recognized as inhibitory to fecundity (Powell et al. 1996), and chicks that ingest plastic fledge at lower weights and are liable to death from dehydration (Sievert and Sileo 1993). 7.5 Financial impacts of a decline in procellariiforms As well as ecosystem impacts, economic impacts may result from alterations in seabird abundance. 7.4.6 Pollution 7.5.1 Guano production Many studies have shown that procellariiforms are sensitive to marine pollution. For this reason, procellariiform species have frequently been used as monitors of pollution, especially oil pollution. Beached bird surveys provide important evidence of geographical and temporal patterns, and, for Guano is not primarily produced by procellariiforms, but they are a contributor to this valuable economic resource. The growth of seabird populations from 1925 to 1955 in the Peruvian Current was probably a response to increased productivity of the Peruvian upwelling system, but a subsequent 07-Turvey-Chap07.indd 161 12/5/2008 12:21:48 PM 162 HOLOCENE EX TINCTIONS drastic decline in seabird abundance was likely to be due to competition for food with the large-scale regional fishery, which caught approximately 85% of the region’s anchovies that would otherwise would have been available for seabirds. This crash lead to a financially devastating decline in guano production on the islands of the Peruvian coast (Jahncke et al. 2004). The decline in guano led to economic depression and political turmoil in western South America that still has repercussions today. 7.5.2 Impact of birds on fisheries Whereas the impacts of fisheries on procellariiforms are widely published, the impacts of procellariiforms on fisheries are less well known. Prior to effective mitigation it was estimated that albatrosses and petrels removed over 20% of all baits from tuna long-line hooks prior to their deployment (Tasker et al. 2000). Brooke (2004) estimated that petrels take approximately 16 million tonnes of prey per annum (approximately twice the annual take of the Japanese fishing fleet), whereas annual pelagic marine fisheries take 70 million tonnes. Brooke (2004) further estimated the trophic levels at which the majority of fisheries and petrels operate, and concluded that seabirds and fisheries overlap considerably. Whether seabirds and fisheries are actually in competition is arguable, though, as they could be taking fish of different ages and size ranges. 7.5.3 Harvest Three significant human harvests of procellariiforms still occur—northern fulmar on the Faeroe Islands and Iceland; short-tailed shearwater in southern Australia; sooty shearwater, Puffinus griseus, in southern New Zealand (Mallory 2006)— and at least 10 other populations are quasi-legally or illegally harvested. These harvests require mechanisms to be put in place to ensure that they remain sustainable and economically viable. However, much ongoing seabird harvesting is sporadic and unobserved. In particular, poaching and difficulties of enforcing harvest prohibitions make it clear that existing legislative restrictions and reservations are not 07-Turvey-Chap07.indd 162 sufficient safeguards to halt ongoing risk to seabirds and the economies they sustain. 7.6 The impact of procellariiforms on ecosystems Little is known about the potential consequences of widespread disappearance of fish-eating and scavenging bird species. There is an urgent need to investigate whether ongoing declines in seabird populations may have unanticipated top-down or bottom-up consequences as a result of trophic cascades or significant reductions in nutrient deposition (Sekercioglu et al. 2004). 7.6.1 Trophic cascades This term has been coined to describe a situation where predators in a food chain suppress the abundance of their prey, thereby releasing the next lower trophic level from predation or herbivory (Paine 1980). When human or other agencies alter this balance, then large-scale ecological shifts may occur. In a classic example, when the abundance of the northern fulmar, a large piscivorous seabird, increased in the North Atlantic, the abundance of its prey (capelin, Mallotus villosus, a zooplanktivorous fish) decreased, large zooplankton abundance increased and phytoplankton biomass decreased. The combined effects of environmental conditions and overfishing have led to dramatic fluctuations in pelagic fish stocks. These changes have, in turn, induced large decreases in some seabird populations (although no information is available for northern fulmars) (Cherel et al. 2001). It would appear that birds reliant on single prey species are more likely to represent such keystone species (Bruno and O’Connor 2005). 7.6.2 Allochthonous nutrients There is increasing recognition that a diverse range of terrestrial ecosystems are in fact supported by nutrients that originate from marine systems (Polis et al. 1997; Erskine et al. 1998; Vitousek 2004; Ellis 2005; Ellis et al. 2006; Harrow et al. 2006). Seabird droppings are enriched in important plant nutrients such as calcium, magnesium, nitrogen, 12/5/2008 12:21:48 PM HOLOCENE PROCELL ARIFORM EX TINCTIONS phosphorus, and potassium. By feeding in productive oceanic waters and defecating inland, seabirds act as a primary vector for transferring nutrients to terrestrial systems. The size of these nutrient inputs cannot be underestimated. It is estimated that seabirds around the world transfer more than 104–105 tonnes of phosphorus from sea to land every year (Murphy 1981). Ironically, many marine currents that facilitate spectacular marine productivity (e.g. Benguela, California, Humboldt) also create temperature inversions that result in low-productivity deserts on nearby landmasses, and seabirds help offset this inbalance by providing allochthonous inputs in the form of guano and carcasses. The effects of seabird nutrient inputs on terrestrial systems can be ably illustrated by several examples. The two species of seabird on 17 ha Heron Island in Australia’s Great Barrier Reef contribute 129 tonnes of guano per annum, including 9.4 tonnes of nitrogen and 1.4 tonnes of phosphorus (Staunton Smith and Johnson 1995). These high nutrient levels (over three times the recommended levels of nitrogen fertilization for nearby agricultural areas) have created unique soil environments that cannot be replicated by terrestrial processes. On the Aleutian Islands, Croll et al. (2005) found that guano was the main source of fertilizer for terrestrial vegetation. When introduced arctic foxes nearly eliminated breeding seabirds in the region, the annual input of guano was reduced from 362 to 5.7 g/m2, resulting in substantial declines in soil phosphorus, marine-derived nitrogen, and plant nitrogen content, and triggering an ecosystem switch from grassland to maritime tundra on fox-infested islands. Similarly, increased numbers of snow geese, Chen caerulescens, due to intensification in agricultural practices along migration pathways in North America has altered plant productivity and community structure on their Arctic breeding grounds thousands of kilometres away (Jefferies et al. 2004). Twenty eight to 38% of the nitrogen in the biota of streams near Westland petrel, Procellaria westlandica, breeding colonies in the South Island of New Zealand is marine-derived (Harding et al. 2004), and possible correlations have been shown between Westland petrel numbers and growth rates of long-lived tree species (Holdaway 07-Turvey-Chap07.indd 163 163 et al. 2007). These effects are also of critical importance today in islands in the Southern Ocean. On high-latitude Marion Island, plants influenced by seabirds were 55% more enriched in nitrogen (1.59 compared with 2.46% N) and 88% in phosphorus (0.17 compared with 0.32% P) than plants away from guano areas (Smith 1978). Plants collected from a range of sites on sub-Antarctic Macquarie Island varied by up to 30% in their leaf δ15N values, with the majority of nitrogen utilized by plants growing in the vicinity of animal colonies or burrows being animal-derived (Erskine et al. 1998). Other studies have demonstrated that the presence of breeding seabirds increases plant productivity (Bancroft et al. 2005; Wait et al. 2005), leaf nutrient status (Anderson and Polis 1999; García et al. 2002), and, probably most telling of all in terms of understanding biodiversity, insect abundance (SanchezPinero and Polis 2000) Any disruption to these nutrient imports may drastically affect ecosystems at both small and large scales (Vanni et al. 2004), and this is especially true in coastal areas and unproductive island systems. In particular, vast colonies of burrowing petrels and shearwaters existed on the New Zealand mainland before human contact (Holdaway 1989), but today these have been virtually eradicated (Worthy and Holdaway 2002). Although New Zealand and many other islands are generally interpreted by conservation planners and restoration ecologists as being naturally nutrient-poor systems, the removal of huge numbers of inland seabird colonies by introduced mammalian predators during the Holocene has undoubtedly had a massive impact on island nutrient cycling and productivity, which needs to be recognized in future conservation management (Harding et al. 2004). However, it remains extremely difficult to make meaningful inferences about ecosystemlevel changes that lack direct scientific observation data, even from the recent past (see also Chapter 10 in this volume). For example, several studies have attempted to address the impact of the now-extinct passenger pigeon, Ectopistes migratorius, on North American ecosystems (Webb 1986; Ellsworth and McComb 2003). Between the seventeenth and nineteenth centuries, ‘countless numbers’ and ‘infi nite multitudes’ of these birds were 12/5/2008 12:21:49 PM 164 HOLOCENE EX TINCTIONS described on their annual migration from eastern and central Canada and the north-east USA to the southern USA, with flocks sometimes a mile or more in width and taking several hours to pass overhead (Lewis 1944), but the species was extinct in the wild by 1900. The role that such huge numbers of birds played in shaping their environment cannot be underestimated, but calculating factors such as the effect of nutrients from passenger pigeon droppings on plant growth is confounded by the actual physical impact of such huge numbers of birds on vegetation, and the effects of other large-scale vegetational changes brought about by agricultural changes over the same time period. Similarly, the consequences of the loss of marine nutrient inputs to the main islands of New Zealand are unknown; given the unique character of the archipelago’s now-compromised combination of colonial seabirds breeding in mammal-free temperate forests on a relatively large landmass, predicting the consequences for New Zealand terrestrial ecosystems of the extinction of most burrow-nesting seabirds from such studies is difficult. Indeed, the effects of seabird removal on the soil chemistry and ecology of larger areas are often poorl y understood. For example, on the New Zealand mainland, recent work has shown that topography and underlying soil chemistry influence the effects of nitrification (Harrow et al. 2006). Furthermore, the effects of nitrification are difficult to predict because the responses of different tree taxa to nutrient enhancement vary considerably (Islam and Macdonald 2005), and excessive nutrient inputs may inhibit growth, change species composition, or even kill certain species of plants (Hogg and Morton 1983). remove space 7.6.3 Environmental modification The term ecosystem engineer describes a species that modulates resource flows and species composition within an ecosystem through the physical modification of habitat (Jones et al. 1994). Procellariiforms are classic ecosystem engineers, and burrow building is a common example of ecosystem engineering, as burrows may provide habitats for a range of other burrowing and non-burrowing species. High burrow densities can also reduce plant growth rates and seedling establishment (Mulder and Keall 2001). Biopedturbation has been shown to drive decreased diversity and structural complexity in island ecosystems (Bancroft et al. 2005), and influences entire island ecosystems through the effects of burrowing and underground deposition of vegetation on biotic and abiotic island processes (McKechnie 2006). Comparison of ratfree and rat-invaded offshore islands in New Zealand has shown that predation of seabirds by introduced rats has led to altered soil properties, thereby structuring plant and animal communities (Bancroft et al. 2004, 2005; Fukami et al. 2006). 7.6.4 Seabirds as scavengers and their role in reprocessing nutrients Although not generally obligate scavengers, most seabird species will scavenge opportunistically. In particular, giant petrels, Macronectes spp. , together with skuas, Catharacta spp. , are considered the marine equivalents of vultures on subAntarctic islands, although giant petrels also feed on marine invertebrates and chicks of various marine birds (Hunter and Brooke 1992). Albatross and petrel diets contain large amounts of deepwater species of squid and crustaceans that have not been recorded by oceanographers at the surface even at night, and it is hypothesized that other than scavenging at fishing boats the only way such food items could be obtained is through scavenging at the surface following natural death (Croxall and Prince 1994). We know little about the potential ecological consequences of the changes in the numbers of these scavenging seabirds, but it has been theorized that a lack of alternate pathways for the remineralization of nutrients may decrease ecosystem biodiversity and resilience (Nixon 1981). 7.7 The original terrorists? The structure of procellariiform populations is very like the structure suggested by Louis Auguste Blanqui (1885) for revolutionary human organizations to avoid persecution and survive. By their very nature, the secretive yet widespread "ecosystem engineer" 07-Turvey-Chap07.indd 164 12/5/2008 12:21:49 PM HOLOCENE PROCELL ARIFORM EX TINCTIONS 165 Table 7.4 Procellariiform populations that have become established in the Holocene. Species Locality Reference de la Mare and Kerry (1994) Miskelly et al. (2006) Gummer (2003) Fisher (1952) Four species of Antarctic petrel Macquarie Island Pitt and Main Islands (Chatham Group) Hawaiian Islands Iceland to the Faeroes, Shetland and Orkney Islands, and down the British and Irish coasts to the Channel Islands and France Antarctic coastline Human-encouraged Phoebastria immutabilis Pachyptila turtur Puffinus newelli Pterodroma axillaris Pterodroma leucoptera Pterodroma phaeopygia Pterodroma pycrofti Puffinus gavia Pelacanoides urinatrix Oceanodroma leucorhoa Hawai’ian Islands Mana Island (New Zealand) Kilauea Point, Kauai (Hawai’i) Pitt Island Boondelbah Island (NSW, Australia) Santa Cruz Island (Galapagos) Cuvier Island (New Zealand) Maud and Mana Islands (New Zealand) Mana Island (New Zealand) Old Hump Ledge and Ross Island (Gulf of Maine, USA) Natural Diomedea sp. Diomedea antipodensis Phoebastria immutabilis Fulmarus glacialis species of procellariiforms have formed comparable isolated, independent ‘cells’ which experience limited genetic interchange over time. By existing in these isolated pockets, populations of many species have survived virtually unchanged over long intervals of geological time, and although populations of more than 50% of extant procellariiform species have been documented to have lost populations during the Holocene (Table 7.3), most species still retain isolated populations that are protected from sources of predation due to their Blanquistic behaviour. However, these refuges are now being progressively invaded by the recent demand for coastal property and the increasing use by humans of even very small islands. 7.8 So, what is being done to prevent this ancient group from extinction? Given the antiquity of the procellariiform lineage, the extreme effects that large seabird breeding colonies have at the landscape and regional levels, and the importance of their ongoing population declines, extirpations and extinctions cannot be 07-Turvey-Chap07.indd 165 Hiller et al. (1988); Ainley et al. (2006) Gummer (2003) Gummer (2003) Byrd et al. (1984) Gummer (2003) Priddel et al. (2006) Podolsky and Kress (1992) Gummer (2003) Gummer (2003); Bell et al. (2005) Miskelly et al. (2004) Podolsky and Kress (1989) overestimated. Seabird fisheries by-catch can and is being addressed within the territorial waters of most countries, and is also being tackled by international initiatives and agreements outside national Exclusive Economic Zones. Recent work has demonstrated that compensatory mitigation can facilitate high-value uses of biological resources and cost-effective conservation gains for species of concern (Wilcox and Donlan 2007). For example, by levying fishers for their by-catch, money can be made available to remove invasive mammals from breeding islands. However, pollution can only be prevented from affecting populations through effective international agreements and monitoring. It has been shown that removal of invasive predators is more than 20 times more effective in giving percentage increases in population growth per dollar invested than fisheries closures, and is also more socio-politically feasible (Wilcox and Donlan 2007). Rat populations can be reduced by poisoning (Igual et al. 2006), but ultimately the only long-term solution on islands is eradication. Work in New Zealand has shown that large-scale rat 12/5/2008 12:21:49 PM 166 HOLOCENE EX TINCTIONS eradication can be achieved even on comparatively large islands such as Campbell Island (115 km2) in New Zealand’s sub-Antarctic (Towns and Broome 2003). Techniques have also been developed to remove feral cat populations from islands. Over the past two decades, these conservation techniques have prevented the extinction of many island species and restored many island ecosystems (Nogales 2004). Habitat restoration and active 07-Turvey-Chap07.indd 166 transfer of petrel species to islands is further helping to redress the balance (Table 7.4). Ultimately, the procellariiforms are a largely secretive group of birds that remain difficult to study, and their precarious status is also relatively poorly known. Ironically, their survival may ultimately depend on publicity and public awareness, and not the secretiveness that has enabled them to survive for so long. 12/5/2008 12:21:50 PM