Search

As top predators, seabirds can be indirectly impacted by climate variability and commercial fishing activities through changes in marine communities. However, high mobility and foraging behaviour enables seabirds to exploit prey distributed patchily in time and space. This capacity to adapt to environmental change can be described through the study of their diet. Traditionally, the diet of seabirds is assessed through the morphological identification of prey remains in regurgitates. This sampling method is invasive for the bird and limited in terms of taxonomic resolution. However, the recent progress in DNA-based approaches is now providing a non-invasive means to more comprehensively and accurately characterize animal diets. Here, we used a non-invasive metabarcoding approach to characterize the diet of the Westland petrel (Procellaria westlandica), which is an endangered burrowing species, endemic to the South Island of New Zealand. We collected 99 fresh faecal samples at two different seasons and in two different sub-colonies. Our aims were to describe the diet of the Westland petrel, investigate seasonal and spatial variation in the petrels’ diet, and assess potential impacts of the New Zealand fishery industry. We found that amphipods were the most common prey, followed by cephalopods and fish. Our results could be the result of natural foraging behaviour, but also suggest a close link between the composition of prey items and New Zealand’s commercial fishing activities. In particular, the high abundance of amphipods could be the result of Westland petrels feeding on discarded fisheries waste (fish guts). Our results also showed significant differences in diet between seasons (before hatching vs chick-rearing season) and between sampling sites (two sub-colonies 1.5 km apart), indicating plasticity in the foraging strategy of the Westland petrel. Due to its non-invasive nature, metabarcoding of faecal samples can be applied to large numbers of samples to help describe dietary variation in seabirds and indicate their ecological requirements. In our example, dietary DNA (dDNA) provided valuable information regarding the dietary preferences of an iconic species in New Zealand’s unique biodiversity. dDNA can thus inform the conservation of endangered or at-risk species that have elusive foraging behaviours.

Many seabird populations are of conservation significance, either because they are rare and threatened with extinction, or because they are abundant and play an integral role in island-ocean systems. Procellariiform petrels in the families Procellariidae, Hydrobatidae and Oceanitidae are especially vulnerable to the impacts of alien invasive species on islands and as a result 45 of 97 species globally are threatened with extinction, a higher proportion than comparable bird groups. In their natural state petrels are also the most abundant seabirds, consuming prey volumes commensurate with commercial fisheries, and transferring nutrients from pelagic to terrestrial and coastal ecosystems on a scale equivalent to other global geochemical fluxes. Seabird nutrient inputs ramify to influence whole-island food-webs. To prevent seabird extinctions, and to restore nutrient pathways the whole-island eradication of invasive species has been developed as an effective conservation tool. In this thesis I explore the conservation ecology of a diverse group of petrels following the World’s largest multi-species eradication to date—the staged removals of Wekas Gallirallus australis, Feral Cats Felis catus, European Rabbits Oryctolagus cuniculus, Black Rats Rattus rattus and House Mice Mus musculus from subantarctic Macquarie Island.Petrels are particularly challenging to study, and few quantitative data on their responses to invasive species eradication are available. Petrels nest discontinuously across rugged terrain on remote islands. Their nests are in under-ground burrows, the entrances to which are often obscured under dense vegetation. Birds are not present year-round, and only return to or leave from colonies at night. Given these inherent challenges, I review in Chapter 1 what motivates people to collect and report population estimates for petrels. Because these species are cryptic and challenging to study I undertook a meta-analysis of uncertainty in published estimates to assess overall levels and drivers of uncertainty. I concluded that the primary motivation for estimating populations is to inform species status and trend assessments, but uncertainty in generated estimates is an impediment to reliable trend detection.The findings of the review provide the rationale for testing field methods that measure two key metrics that underpin population estimates: the proportion of burrows occupied by breeding pairs (Chapter 2), and the total number of burrows on an island (Chapter 3). Chapter 2 compares traditional techniques for estimating burrow occupancy and breeding status with my novel application of camera traps positioned at burrow entrances. I found that (i) camera traps are effective at collecting season-long activity patterns, with evidence that these can be used to distinguish breeding from non-breeding burrows, and (ii) cameras provide higher resolution data on breeding success than repeat visits to inspect burrows manually and are a low impact approach.To understand the current size and distributions of a diverse assemblage of petrels that includes common widespread species (Antarctic Prion Pachyptila desolata) and rare localised species (Blue Petrel Halobaena caerulea and Grey Petrel Procellaria cinereal) a multi-method approach to surveys and analysis is required (Chapter 3). I found that a model-based analysis of stratified randomised whole-island survey data was most effective for estimating Antarctic Prions, less effective for scarcer White-headed Petrels Pterodroma lessonii, but did not yield useful data for Blue or Grey Petrels, which required a targeted approach.In the final chapters I evaluate the outcomes that invasive species management has achieved for petrels to date (Chapter 4) and projected into the future (Chapter 5). The current distribution of petrels on Macquarie Island reflects the legacy of invasive species impacts and the removal of this threat. The two established species, Antarctic Prion and White-headed Petrel are confined largely to refugial habitat, while the recolonising species Blue and Grey Petrel are occupying optimal habitat on the island’s coastal slopes. While all four species occur in superficially similar habitats in terms of vegetation type, their realized niches defined by other environmental variables including slope, elevation, and aspect are starkly different. Each species is now increasing, but the two recolonising species are increasing much more rapidly than the two established ones, at rates suggesting immigration is an important mechanism facilitating initial population recovery.The eradication of invasive predators from Macquarie Island has achieved a major short-term goal: to prevent further seabird extinctions and facilitate threatened species recovery. I present a case for downlisting three species from state or federal threatened species lists. However, suitable habitat is currently under-utilised. Using habitat suitability modelling and inter-island inference I illustrate how the current size and distribution of petrel populations is well short of plausible carrying capacities. Full recovery of populations, and restoration of the ecological and societal benefits that they bring is decades away, but no long-term targets for population recovery exist. I present a future scenario based upon plausible model-derived population estimates to highlight the potential advantages of establishing long-term targets. The thesis concludes with my own narrative vision of a future Macquarie Island when the full post-eradication recovery has occurred in coming decades.

We report the breeding success of four species of burrow-nesting petrels at sub-Antarctic Marion Island where house miceMus musculusare the sole introduced mammal. Feral catsFelis catuswere present on Marion for four decades from 1949, killing millions of seabirds and greatly reducing petrel populations. Cats were eradicated by 1991, but petrel populations have shown only marginal recoveries. We hypothesize that mice are suppressing their recovery through depredation of petrel eggs and chicks. Breeding success for winter breeders (grey petrelsProcellaria cinerea(34±21%) and great-winged petrelsPterodroma macroptera(52±7%)) were lower than for summer breeders (blue petrelsHalobaena caerulea(61±6%) and white-chinned petrelsProcellaria aequinoctialis(59±6%)) and among winter breeders most chick fatalities were of small chicks up to 14 days old. We assessed the extent of mouse predation by monitoring the inside of 55 burrow chambers with video surveillance cameras (4024 film days from 2012–16) and recorded fatal attacks on grey (3/18 nests filmed, 17%) and great-winged petrel chicks (1/19, 5%). Our results show that burrow-nesting petrels are at risk from mouse predation, providing further motivation for the eradication of mice from Marion Island.

The Grey Petrel (Procellaria cinerea) is listed as Near Threatened globally owing to incidental mortality on long-line fishing gear and reduced breeding success on islands caused by the introduction of alien predators. However, there are few studies of its foraging ecology and none of its diving behaviour. We obtained data from temperature–depth recorders (n = 7 birds) and global positioning satellite trackers (n = 15) deployed on Grey Petrels breeding on Gough Island, South Atlantic Ocean. Most birds foraged in the productive oceanic waters west or north-west of South Georgia. Average maximum dive-depth was 3.2 ± 2.2 m with most dives

The white-chinned petrel Procellaria aequinoctialis is one of the most frequently observed seabird species captured in fisheries bycatch, yet some populations remain virtually unstudied. The size of the breeding population on the sub-Antarctic Auckland Islands, New Zealand, is unknown. Disappointment Island is thought to be the main white-chinned petrel breeding site in the Auckland Islands, and maybe also in the New Zealand region, and has never had introduced mammalian predators. We estimated the white-chinned petrel breeding population size taking into account the detection probability of burrows via distance sampling and the burrow occupancy rate. Eighty line transects were distributed over the island, with a total line length of 1600 m. Burrows were patchily distributed and most abundant in dense megaherb communities. White-chinned petrel burrow density $$ \hat{D } $$ was 654 burrows/ha (95 % CI 528–809 burrows/ha), with burrow detection probability $$ \hat{p} $$ varying among vegetation communities from 0.28 ± 0.02 to 0.43 ± 0.02 (±SE). Mean burrow occupancy was 0.73 ± 0.03. We document an estimated total of 155,500 (125,600–192,500) breeding pairs of white-chinned petrels on Disappointment Island during mid incubation in early January 2015. The relatively high occupancy and density of burrows suggest that Disappointment Island is a key breeding site for white-chinned petrels.

Since 2004 there has been mounting evidence of the severe impact of introduced house mice (Mus musculusL.) killing chicks of burrow-nesting petrels at Gough Island. We monitored seven species of burrow-nesting petrels in 2014 using a combination of infra-red video cameras augmented by burrowscope nest inspections. All seven camera-monitored Atlantic petrel (Pterodroma incertaSchlegel) chicks were killed by mice within hours of hatching (average 7.2±4.0 hours) with an 87% chick failure rate (n=83 hatchlings). Several grey petrel (Procellaria cinereaGmelin) chicks were found with mouse wounds and 60% of chicks failed (n=35 hatchlings). Video surveillance revealed one (of seven nests filmed) fatal attack on a great shearwater (Puffinus gravisO’Reilly) chick and two (of nine) on soft-plumaged petrel (Pterodroma mollisGould) chicks. Mice killed the chicks of the recently discovered summer-breeding MacGillivray’s prion (Pachyptila macgillivrayiMathews), with a chick mortality rate of 82% in 2013/14 and 100% in 2014/15. The closely-related broad-billed prion (P. vittataForster) breeds in late winter and also had a chick mortality rate of 100% in 2014. The results provide further evidence of the dire situation for seabirds nesting on Gough Island and the urgent need for mouse eradication.

The use of long-term ecological datasets to explore the importance of the effects of environmental variability on higher predator populations has been focused mainly on high-latitude areas. We modeled the population dynamics of the Westland Petrel (Procellaria westlandica), which spends its time mostly in subtropical waters during both breeding and the interbreeding migration across the Pacific Ocean. We found that the population has slowly increased since the early 1970s, a result of high adult survival, high fecundity (0.6 of all eggs laid survived to fledge) and moderate mean age at first return to the colony (7.7 yr; a recruitment age typical for this genus), strong recruitment rate of juveniles, and negligible emigration. The modeled population trends were supported by similar rates of increase in nest occupancy since 2001 and nest density since 2007. Annual adult survival for breeders was the same for both sexes (0.954, 95% CI: 0.918–0.975) and constant across years. However, nonbreeders ha...

Primary, tail and body moult of three seabirds from Tristan da Cunha archipelago were studied by castnetting offshore south Brazil from February 2006 to August 2007. Timing, duration and synchronization of primary and tail moult are described relative to the annual calendar. Body moult overlapped breeding in Atlantic yellow-nosed albatrosses (Thalassarche chlororhynchos), but tail and primary moult did not. Spectacled petrels (Procellaria conspicillata) had protracted body moult, whereas primary and tail moult were completed by August. We documented onset of primary moult during chick-rearing in spectacled petrels and great shearwaters (Puffinus gravis) of unknown breeding status, and suggest that the south-west Atlantic Ocean holds important numbers of moulting birds of both species during the summer–early autumn. The albatrosses and the spectacled petrels replaced rectrices alternately. Great shearwaters replaced rectrices outward, starting at the central pair. Primary, tail and body moult largely overlap in all three species, suggesting that the metabolic costs of primary moult may not be overly restrictive. Metabolic and nutritional ability to afford simultaneous moult of different feather tracts support the idea that impaired flight caused by wing moult is a strong factor driving no overlap of primary moult and breeding.

The effectiveness of tori-lines combined with double-weighted branch lines in reducing seabird bycatch was evaluated in a pelagic longline fishery. Seabird bait attack behaviour, bycatch number, and sinking rate of baited hooks were examined on double-weighted and unweighted branch lines deployed on the same longline with a single tori-line. Comparisons were conducted from July to October in 2011 during two cruises on a chartered longline vessel in the eastern South Pacific Ocean outside the Chilean and Peruvian exclusive economic zones. Cape petrels (Daption capense), white-chinned petrels (Procellaria aequinoctialis) and Westland petrels (Procellaria westlandica) were abundant during line setting. There were 275 primary bait-taking attacks by these species, but there were only eight primary attacks by albatrosses. Nevertheless, six albatrosses and six diving seabirds were incidentally caught as bycatch. Of the primary attacks by Cape petrels, white-chinned petrels and Westland petrels, 153 led to secondary attacks. These results suggest that off Chile and Peru there is frequent secondary bycatch of albatrosses as a result of their stealing bait from Cape petrels and diving seabirds. Best-fit models for the number of primary attacks and of bycatch included the weighted branch line; the use of weighted branch lines resulted in a lower number of primary attacks. Hooks on unweighted branch lines did not reach any of the benchmark depths (3, 5, and 10 m) within the aerial extent of the tori-lines (the tori-line remaining above the water surface), hooks on weighted branch lines reached 5 m depth within the aerial extent. These results suggest that, for the pelagic longline fishery off Chile and Peru, combining double-weighted branch lines and tori-lines reduces the bycatch more effectively than tori-lines with unweighted branch lines. Copyright © 2014 John Wiley & Sons, Ltd.

We comprehensively tested combinations of three primary mitigation measures in a pelagic longline fishery with one of the highest rates of interaction with what may be the world's most challenging seabird assemblage (dominated by Procellaria genus petrels), aboard fishing vessels typical of the Asian distant water fleet. Multiple measures were used to compare the performance of weighted vs. unweighted branch lines set with two bird-scaring lines – hybrid lines with long and short streamers – during daytime and nighttime. The weights used were a novel double-weight configuration. Secondary attacks on baits brought to the surface by white-chinned petrels drove albatross mortality. Regardless of time of day, weighted branch lines with two bird-scaring lines, deployed and maintained with an aerial extent of 100 m, reduced bird attacks by a factor of four, and secondary attacks and seabird mortality by a factor of seven, compared to unweighted branch lines, with little effect on fish catch rates and with no injuries to crew. This combination yielded zero bird mortalities when gear was set at night. We conclude that the simultaneous use of two bird-scaring lines, weighted branch lines and night setting meet our criteria for best-practice seabird bycatch mitigation for the joint-venture fleet targeting tuna and related species in the South African EEZ. To be successful, the aerial extent of bird-scaring lines should be aligned with the distance astern that baited hooks sink beyond the foraging depth of the dominant seabird – in this case white-chinned petrels to a depth near 5 m. Given that these measures were successful in one of the most challenging pelagic longline fisheries, they are likely to be widely applicable to pelagic longline fisheries using similar gear.

The large midden site of Offing 2 (southern Chile) has yielded a series of human occupations dating back to 4200–2500 bp. From a very large collection of bird specimens, we sampled over 45 000 bird bones from four taxa (Diomedea sp., Procellaria sp., Phalacrocorax sp. and Sterna hirundinacea). The distributions of skeletal parts do not contradict the bone density and functional anatomy hypotheses for cormorants and terns, but the observed data for albatross and petrel are less convincing. The presence of most or all elements (except for the terns) indicates that the birds were brought back to the site as complete carcasses. The deficit of some of the long wing bones observed for albatross and petrel is remarkable and may indicate a particular removal for later use as raw material for tools or trade goods. Copyright © 2013 John Wiley & Sons, Ltd.

We compared the performance of two bird-scaring line designs (light lines with short streamers vs. hybrid lines with a mix of long and short streamers) deployed in pairs with unweighted branch lines on two joint venture tuna vessels typical of distant-water tuna fisheries in the South Africa Exclusive Economic Zone (EEZ). We also added weight to a subset of branch lines, and compared the effects of line weighting and night vs. day setting in combination with bird-scaring lines on bird and fish catch rates. White-chinned petrels (Procellaria aequinoctialis), a diving seabird, dominated the seabird assemblage; they were the most numerous seabird and they attacked baited hooks and were killed at the highest rates. Secondary attacks – surface foraging albatrosses stealing baits from white-chinned petrels – drove albatross mortality. With hybrid scaring lines deployed, both diving and surface foraging seabirds made fewer attacks (1.5 and 2 times, respectively) within the lines’ 100 m aerial extent, where hooks are closest to the surface and birds are most vulnerable to hooking, than with light lines. However, all metrics of comparison between hybrid and light lines were not statistically conclusive, primarily because birds could access baited hooks in areas beyond the protection afforded by bird-scaring lines (aft and to port of their aerial extent). Seabird bycatch rates were 4.6 times higher during daylight hours (2.00 birds/1000 hooks; 52 birds) than at night (0.439 birds/1000 hooks; 28 birds) and night catch rates near the full moon doubled. Bird catch rates were 18 times higher on unweighted branch lines (1.07/1000 hooks; 79 birds) than on weighted branch lines (0.06/1000 hooks; 1 bird) with no detectable effect on fish catch. With respect to streamer lines, our results suggest that in Procellaria petrel dominated systems the aerial extent of bird-scaring lines (of any design) should span the distance that baited hooks are within 10 m of the surface to effectively prevent bird attacks on baits. Overall, results suggest that night setting, adequate branch line weighting, and proper deployment of two bird-scaring lines have the potential to reduce seabird bycatch in pelagic longline fisheries without reducing the catch rates of target fishes. Further development of safe, fast-sinking branch line weighting configurations and bird-scaring lines less prone to tangling on fishing gear is critical to this effort.

The Westland petrel (Procellaria westlandica) is an endemic New Zealand species and one of the very few burrowing seabird species still breeding on mainland New Zealand. It nests only on a series of coastal ridgelines near to Punakaiki on the West Coast of the South Island. Between 2002 and 2005, surveys were undertaken at 28 of the 29 known colonies. The area occupied by the colonies was 73 ha; most colonies had fewer than 50 burrows, but six colonies had 201–500 burrows and four colonies had more than 1000 burrows. We find that the current breeding range of Westland petrel and the location of individual colonies are similar to those reported in both the 1950s and 1970s. Based on total burrow counts at 28 colonies and burrow occupancy rates determined by annual monitoring, the annual breeding population is estimated to be between 2954 and 5137 breeding pairs.

Satellite transmitters were attached to eight adult spectacled petrels Procellaria conspicillata Gould captured during the early incubation period at their breeding grounds on Inaccessible Island, one of the Tristan da Cunha Islands in the central South Atlantic Ocean. Data on their at-sea distribution was obtained for up to six months. All birds remained within the South Atlantic from 24–44°S, with most between 25 and 40°S. Breeding birds mainly foraged in oceanic waters, but failed breeders or non-breeders concentrated their foraging activity over the Rio Grande Rise and the Walvis Ridge and along the shelf break off the east coast of South America. Little foraging occurred along the Benguela shelf break off southern Africa. Non-breeders favoured relatively warm water with low chlorophyll concentrations, reducing the risk of bycatch in fisheries. Tracked birds spent 16% of their time in areas with high levels of tuna longline fishing activity, with overlap greater for non-breeding birds (22%) than breeding birds (3%). Birds in this study foraged in shallower waters along the continental shelf edge off South America than spectacled petrels tracked in this area in winter, potentially increasing their risk of exposure to demersal longline fisheries in this area in summer.

SummaryThe predatory behaviour of introduced house mice Mus musculus at Gough Island is known to impact on albatross and petrels, resulting in the Tristan Albatross Diomedea dabbenena and Atlantic Petrel Pterodroma incerta being listed as “Critically Endangered” and “Endangered”, respectively. Although predation has been documented for two burrowing petrels and one albatross species, the impact of house mice on other burrowing petrels on Gough Island is unknown. We report burrow occupancy and breeding success of Atlantic Petrels, Soft-plumaged Petrels Pterodroma mollis, Broad-billed Prions Pachyptila vittata, Grey Petrels Procellaria cinerea and Great Shearwaters Puffinus gravis. With the exception of the Great Shearwater, breeding parameters of burrowing petrels at Gough Island were very poor, with low burrow occupancy (range 4–42%) and low breeding success (0–44%) for four species, and high rates of chick mortality in Atlantic Petrel burrows. Breeding success decreased with mass, suggesting that smaller species are hardest hit, and winter-breeding species had lower breeding success than summer breeders. The results indicate that introduced house mice are having a detrimental impact on a wider range of species than previously recorded and are likely to be causing population declines among most burrowing petrels on Gough Island. The very low values of burrow occupancy recorded for Soft-plumaged Petrels and Broad-billed Prions and greatly reduced abundance of burrowing petrels in comparison to earlier decades indicate that Gough Island’s formerly abundant petrel populations are greatly threatened by the impact of predatory house mice which can only be halted by the eradication of this species from the island.

Seabirds deposit large quantities of marine detritus on land, but little is known of the soil arthropods processing this material. Burrow-nesting seabirds concentrate their activities within their burrows, so we tested the hypothesis that burrow arthropod fauna is more marine-like in its isotopic enrichment (13C/12C, 15N/14N; expressed as δ13C and δ15N) than the arthropods on the adjacent forest floor. Results from a Westland petrel (Procellaria westlandica) colony on the South Island of New Zealand did not support the hypothesis. Instead, δ15N was universally marine (13–22‰). While δ13C separated into two clusters, the distribution was not according to arthropod provenance. Most taxa had a terrestrial δ13C; only two taxa (a leiodid beetle and the mesostigmatic mite Ayersacarus woodi) incorporated marine C. The leiodid beetle occurs both in burrows and on the forest floor; beetles from both habitats had a marine δ13C. Ayersacarus woodi is found only in burrows. We conclude that, in this system, m...

Endemic Westland petrels (Procellaria westlandica) are a remnant of extensive seabird populations that occupied the forested hill country of prehuman New Zealand. Because seabird guano is rich in Se, an often-deficient essential element, we proposed that Westland petrels enhance Se concentrations in ecosystems associated with their breeding grounds. We sampled terrestrial (soil, plants, riparian spiders) and freshwater (benthic invertebrates, fish) components from Westland petrel-enriched and non-seabird forests on the western coast of New Zealand's South Island, an area characterised by highly leached, nutrient-poor soils. Median seabird soil Se was an order of magnitude higher than soil from non-seabird sites (2.2 mg kg− 1 compared to 0.2 mg kg− 1), but corresponding plant foliage concentrations (0.06 mg kg− 1; 0.05 mg kg− 1) showed no difference between seabird and non-seabird sites. In streams, Se ranged from 0.05 mg kg− 1 (riparian foliage) to 3.1 mg kg− 1 (riparian spiders and freshwater mussels). However, there was no difference between seabird and non-seabird streams. Stoichiometric ratios (N:Se, P:Se) showed Se loss across all ecosystem components relative to seabird guano, except in seabird colony soil where N was lost preferentially. Seabirds therefore did not enrich the terrestrial plants and associated stream ecosystems in Se. We conclude that incorporation of trace elements brought ashore by seabirds cannot be assumed, even though seabirds are a significant source of marine-derived nutrients and trace elements to coastal ecosystems world-wide.

During seabird censuses performed as part of scientific research looking into seabird bycatch onboard industrial pelagic longline vessels targeting swordfish (Xiphias gladius) we made the first documented records of black petrel (Procellaria parkinsoni) in the pelagic waters of northern Chile, considerably extending the species’ range southward. These observations were made during hauling operations in the austral winters of 2008, 2009 and 2010 between 23°00'S and 32°49'S. Black petrels were observed in 9.6% of censuses and a total of 10 birds were recorded. Despite the fact that this species fed upon discards and wastes generated during fishing, no incidental mortality was observed. Our results are relevant to the conservation of the black petrel in the south-eastern Pacific marine ecosystem, as they provide new information on species range and interactions with pelagic fisheries.

Mites are a diverse and important component in the soils of the Southern Ocean islands, but for many groups, their taxonomy and biogeography is little studied. This paper reports the Algophagidae, a poorly known family, from the New Zealand region for the first time. The study used soil samples from seabird burrows, brackish algal wrack samples and a museum collection to recover specimens. Terraphagus antipodus gen. n., sp. n. is described from grey petrel Procellaria cinerea burrow soil on the Antipodes Islands, New Zealand. The axillary organ is confined to the dorsum; the epigynal apodeme ends are fused with the ends of coxal apodeme II; seta 2a and ω2 are absent, and the famulus is bilobed. Males have the sternum fused to coxal apodemes II; the tarsal setae of legs I, III and IV of the male are modified for mate guarding. The short stout legs I and II end in huge spines in females as e, d, q, s, wa and hT. From brackish algal wrack, Lake Forsyth/Wairewa, Canterbury, New Zealand, another algophagid Neohyadesia minor sp. n. is described and illustrated. The new species is the smallest known algophagid. A neohyadesid collected in 1961 is newly reported here from Eudyptes sp. penguin rookery mud from North Head, Macquarie Island. These records further extend the known microhabitats of the family to the allochthonous marine seabird nutrient flow into islands in the New Zealand region and make Neohyadesia Hughes and Goodman, 1969 present in all Southern Ocean provinces.

Species of petrels and shearwaters with high diving ability could facilitate the catch of albatrosses in pelagic longline fisheries, because they retrieve bait to the surface from depths that albatrosses cannot reach. Once on the surface, large seabirds such as albatrosses can easily displace smaller species thus gaining access to baited hooks which increases their likelihood of getting caught. This paper evaluates the extent to which diving species (i.e. Procellaria aequinoctialis, Procellaria conspicillata and Puffinus gravis) increase the susceptibility of albatross to bycatch in pelagic longline fisheries. In 48 sets, attacks on baits were quantified. When more than one bird (of the same or different species) tried to attack the same bait this was defined as a multiple attack. There were 384 attacks on baits, of which 260 were made by a single individual and 124 by more than one. Multiple attacks were the largest source of bycatch of albatrosses (22 of 27 albatrosses whose attacks were observed). Of the baits attacked by albatrosses (n = 244), 17% were indirectly facilitated by diving medium-sized petrels. Considering only the multiple attacks in which albatrosses participated (n = 114), 36% were initiated by these medium-sized petrels. Eleven (41%) of the albatrosses captured, and whose attacks were observed, resulted from a diving medium-sized seabird species first having contacted the terminal tackle. This paper shows that medium-sized petrels, with a strong ability to dive, increase considerably the access to bait, and indirectly, the incidental bycatch of albatrosses. Observations made in the absence of mitigation measures also provide useful information to improve the performance of tori lines. Based on the seabird behaviour, it is recommended that tori lines should have a minimum aerial coverage of 50 m. Copyright © 2012 John Wiley & Sons, Ltd.

Rapid classification of a study subject's gender is critical for many ecological, behavioral, and conservation projects. We evaluated sexual-size dimorphism of adult Westland Petrels (Procellaria westlandica), a large nocturnal colony-nesting seabird, using linear discriminant function analysis and compared our results to birds classified using standard DNA gender identification techniques. The results revealed a strong pattern of sexual dimorphism (Wilks' Lambda = 0.43, F7,29 = 5.6, P < 0.001) in the standard discriminant function analysis despite an unbalanced sex ratio in our sample of adults captured at the breeding colony. Minimum bill depth and head length, of the seven morphometric characters we measured, successfully assigned the correct gender to 95% of all individuals sampled (n = 37). We provide a canonical classification function of morphological traits that may be used in the field to rapidly differentiate adult females and males of this rare petrel species.

Ayersacarus woodi sp. n. (Acari: Leptolaelapidae) is described from Westland petrel (Procellaria westlandica Falla) burrows at Punakaiki, South Island, New Zealand. The genus has previously been reported only from sub-Antarctic islands. The new species is most like A. plumapilus Hunter (type species) but differs from it in females in the size and shape of the epigynal shield; much larger metapodal shields; and a wider than long anal shield with lateral pores. The ecology of the new species was explored using stable isotope (13C/12C; 15N/14N) analysis of females alongside contextual data from the site. From the 13C/12C results, the new species is dependent on marine C from petrels rather than terrestrial photosynthetic C. In terms of trophic level, the isotopic data are consistent with consumption of guano decomposers or their eggs.

Until recently, the spectacled petrel Procellaria conspicillata Gould was listed as Critically Endangered due to its small population size and ongoing incidental mortality on fishing gear. Surveys at its sole breeding locality, Inaccessible Island in the central South Atlantic Ocean, indicated that the population increased from 1999–2004, resulting in the species being down-listed to Vulnerable. We repeated the census of breeding spectacled petrels during the early incubation period in October–November 2009. Numbers of burrows increased by 55% from 2004–09, with increases in all count zones, and the greatest changes in peripheral populations. Burrow occupancy estimates remained high, averaging 81% during one-off checks. Our best estimate of the population in 2009 was 14 400 pairs, continuing the c. 7% per year increase inferred since the 1930s following the disappearance of introduced pigs. This confirms the rapid recovery of this species despite ongoing mortality on fishing gear. Our results suggest that at least some procellariiforms are able to sustain strong growth rates in the face of fishing mortality when colony based threats are removed.

Knowledge of the dynamics of long-distance migrations of pelagic seabirds is limited. Recent advances in tracking technology have yielded detailed, continuous accounts of the movements of individual seabirds over large spatial and temporal scales. We studied the timing of migration and year-round distribution of the Westland Petrel (Procellaria westlandica), listed by the IUCN as vulnerable, with miniature archival light loggers (geolocators) deployed on 10 incubating birds breeding in 2007 at Westland, New Zealand. We retrieved data from eight Westland Petrels, indicating the birds migrated in November directly east ∼7000 km from the coast of New Zealand to South American waters in 6 days (range 4–7), then returned the following April in 10 days (range 8–13). The durations of an individual's outward and return flights and the dates of its outward and return migrations were positively correlated. During their journeys east and west, birds spent on average (±SD) 9.9% (±9.7) and 17.2% (±12.0), res...

Seabird species within a community are expected to have distinct trophic niches according to foraging methods and body size, but some seabirds exploit fishery waste which can be quite distinct from natural foods. Even in the presence of fishery discards we hypothesize that a pelagic seabird community is structured according to body size, feeding methods, and access to discards. We measured carbon and nitrogen stable isotopes in whole blood of seabirds from the offshore wintering community of Procellariiformes (albatrosses, petrels, shearwaters and storm-petrels) in the SW Atlantic Ocean. We compared this with isotope values of potential prey items and fishery discards, to investigate the importance of discards from the tuna-shark pelagic longline fishery in the diet of these birds. Despite contrasting body masses and feeding techniques, there was extensive overlap in the range of isotopic ratios for different species. Carbon isotope values were typical of the subtropical offshore region. Nitrogen values also showed a high degree of overlap among species and clearly signify that the birds are feeding on fishery waste (especially shark liver). Recently arrived migrants from the Antarctic and sub-Antarctic, and species still rearing chicks when sampled (e.g. wandering albatross Diomedea exulans) had carbon isotopic values indicative of Antarctic and sub-Antarctic regions. All species breeding at more southerly latitudes underwent a marked shift (increase) in carbon and nitrogen isotopes, indicating a change in diet between breeding and wintering seasons. Cory's shearwater (Calonectris diomedea), the only sampled species not attending vessels, had low nitrogen values reflecting a diet of flying fish which occur naturally in the area and showed no change in isotope values between breeding and wintering grounds. Mixing models demonstrate the need to include an external food source (Antarctic krill) to recreate the isotope values obtained from seabird blood: this corresponds to the blood of wintering migrants retaining a minor component from their diet in the southern breeding grounds. The stable isotope results suggest that the availability of discards from pelagic fisheries in the SW Atlantic Ocean lead to an artificially and poorly structured seabird community, with most species utilizing the same food resource. The balance between population benefits obtained from feeding upon otherwise unavailable discards, and the costs from incidental mortality on longline fishing hooks is likely to differ among species, with positive effects for small-sized species and negative effects on albatrosses and Procellaria petrels.

A recent analysis of archaeological bird remains from Harataonga Bay, Great Barrier Island is reported and compared with a prior study conducted in the 1960s. The assemblages come from two Maori occupations, one dating to ca. fourteenth century AD and the other to late prehistory. The new study identifies several previously unreported species for Harataonga, and the first record of prehistoric Maori use of Black Petrel or Taiko (Procellaria parkinsoni), and possibly Pycroft's Petrel (Pterodroma pycrofti). The assemblages are dominated by seabirds, a common pattern for the South Island but unusual for more northern areas, where forest birds are typically better represented. The limited forest birds are concentrated in the early occupation. Combining the two studies broadens our understanding of past species distributions on Great Barrier Island, and Maori use of these resources over time. Analytically, comparison of the two studies, one conducted some time ago, demonstrates the impact of fine-mesh...

Stable isotope studies have repeatedly shown marine nutrient incorporation from seabirds, anadromous fish and tidal wrack into terrestrial and freshwater ecosystems. However, little is known about the physiological consequences of marine-derived nutrient subsidies. Protein content and lipid storage are important physiologically, and the C:N ratio is a widely used proxy that reflects changes in these quantities. In this study we tested the response of C:N ratios in stream biota to the presence of marine-derived nutrients from Westland petrels, a forest breeding seabird. Samples of different stream invertebrate functional feeding groups, predatory freshwater fish, coarse particulate organic matter and riparian soil and vegetation came from four reference streams and four streams with a wide range of seabird densities. Samples were analysed for percentage C, percentage N and δ15N using isotope ratio mass spectrometry. The data were tested against petrel colony size, stream size and distance from colony to sampling site. Despite increased δ15N accompanying petrel presence (as reported previously), Westland petrels had no effect on stream biota C:N ratios, regardless of colony presence or absence or colony size. Despite the nutrients provided by petrels, we conclude that petrel N replaced rather than enhanced non-marine N in these stream ecosystems.

The genus Procellaria traditionally consists of four species, two restricted to New Zealand and two widespread in the Southern Ocean. All four are threatened because of incidental mortality on longlines and other fishing gear. The White-chinned Petrel P. aequinoctialis is the seabird killed in largest numbers by fisheries in the Southern Ocean. A spectacled form recently has been elevated to species status, Spectacled Petrel P. conspicillata, based on differences in morphometrics, vocalisations and breeding phenology. Cytochrome b sequences support species status for the Spectacled Petrel and show that the White-chinned Petrel has two regional populations, one around New Zealand and one throughout the South Atlantic and Indian Oceans. White-chinned and Spectacled Petrels segregated approximately 0.90 million years ago by allopatric fragmentation, and the two populations within White-chinned Petrels diverged approximately 0.35 million years ago. Climate changes and corresponding changes in ocean currents are most likely responsible for these patterns.

White-chinned petrels Procellaria aequinoctialis and grey petrels Procellaria cinerea are among the most frequently killed seabird species by accidental bycatch, and both species have received strong conservation concern. Data on population size are required to evaluate the impact of bycatch and to establish management plans. We estimated the population size of both species at Kerguelen, Southern Indian Ocean, from 2004 to 2006 by explicitly taking into account detection probability of burrows using distance sampling and burrow occupancy. A total of 31 line-transects were distributed across the eastern part of Kerguelen, representing a total length of 566 km. Detectability was low (from 0.19 to 0.54 for whitechinned petrels, 0.58 for grey petrels). Burrow densities varied from 1.37 � 0.67 to 25.77 � 5.23 burrows ha � 1 for white-chinned petrels and was 2.78 � 0.79 burrows ha � 1 for grey petrels. For white-chinned petrels, these densities were extrapolated to the entire surface area of vegetation and there were 234 000 (186 000–297 000) active burrows on Kerguelen. For grey petrels, the number of active burrows for the eastern part of Kerguelen was 3400 (1900–5600). Based on these estimates, the potential biological removal method suggests that the additional mortality on birds caused by the fisheries operating around Kerguelen can be considered a serious threat for the species at least at the regional scale of the Southern Indian Ocean, especially for grey petrels.

Satellite transmitters were used to determine the marine habitat utilization and fishery overlap of wintering spectacled petrels Procellaria conspicillata in the southwestern Atlantic Ocean in 2006 and 2007. Kernel density analysis of tracked birds demonstrated intense use of waters in the Brazilian Exclusive Economic Zone, from 26 to 31° S, mainly over the continental shelf break and off- shore waters (mean depth in the < 20% kernel density areas = 1043 ± 794 m), over warm tropical and subtropical waters (mean SST = 22°C and 21°C in 2006 and 2007, respectively) and mesotrophic/ oligotrophic waters (chlorophyll a density = 0.301 and 0.281 mg m -3 in 2006 and 2007, respectively). These habitats used by spectacled petrels are remarkably different from those used by their sister species, the white-chinned petrel P. aequinoctialis, which occurs mainly over the continental shelf on sub-Antarctic and productive waters. A close association between spectacled petrels and the pelagic longline fishery was demonstrated through comparison of the main kernel areas used by spectacled petrels and the areas used by the pelagic longline fleet determined by fishing effort at resolution of 1 × 1° quadrants. Travel speeds and distances during night and daytime periods did not differ. The present study demonstrates the importance of high resolution fishing effort data to address relation- ships between bird marine habitat use and specific fishing fleets, and to determine marine habitats and investigate at sea segregation between closely related species.

Most studies of seabirds that involve trapping and marking birds are carried out at breeding colonies. This bias toward the breeding period and colony-based research is partially caused by difficulties in capturing birds at sea. From 2005 to 2007, we used a cast net thrown by hand from a fishing boat to capture albatrosses and petrels at sea in the southwestern Atlantic Ocean. About 500 birds of 13 species were captured, ranging in size from the 30-g Wilson's Storm-petrel {Oceanites oceanicus) to the 10-kg Wandering Albatross {Diomedea exulans). Cast nets are potentially useful for capturing any seabird that can be attracted close to fishing vessels by bait, such as sardines, squid, offal, or shark liver, thrown into the water. Our method was most effective for capturing bold species, such as Cape {Daption capense), Spectacled {Procellaria conspicillata), and White-chinned {Procellaria aequinoctialis) petrels and Great Shearwaters {Puff nus gravis), but was not effective for capturing shy species, such as Cory's {Calonectris diomedea), Cape Verde {Calonectris edwardsii) and Manx {Puff nus puff nus) shearwaters, species that rarely sit on the water, such as Wilson's Storm-petrels, Thin-billed {Pachyptila belcheri) and Antarctic {Pachyptila desolata) prions and Atlantic Petrels {Pterodroma incerta), and species with excellent diving abilities, such as Sooty Shearwaters {Puff nus griseus). For many species of seabirds, cast nets would likely be more efficient for capturing large numbers of birds than other methods, such as hoop nets.

Flocks of seabirds attending commercial bottom long-line fishing operations on the coastal shelf off southeastern Brazil show a greater species diversity during the summer than the winter (16 v 9 species), although the number of birds per flock tended to be greater during the cold season. During the summer, the Spectacled Petrel Procellaria aequinoctialis conspi-cillata was the commonest species, followed by the Great Shearwater Puffinus gravis, Cory's Shearwater Calonectris diomedea, skuas Stercorarius spp. and the Yellow-nosed AlbatrossDiomedea chlororhynchus. During the winter, the White-chinned Petrel Procellaria aequin-octialis, Yellow-nosed Albatross and Black-browed Albatross Diomedea melanophrys were jointly the commonest species. Marked differences in the relative abundance of species were observed between the different sampling periods, probably because of migratory movements but also because of seasonal shifts of the sea currents and the influence of cold fronts. Waters off southeastern Brazil are important feeding areas for some seabird populations nesting in the Tristan da Cunha and Gough group, especially for nonbreeding Spectacled Petrels and post-breeding Yellow-nosed Albatrosses.

SUMMARY The food of L. argentatus on Skomer and Skokholm Islands (Wales) was studied by the analysis of stomach contents, pellets and direct observation. L. argentatus is mainly a scavenger and exists for the most part on fish-waste and garbage discarded by man. At Ncwborough Warren argentatus feeds more on arable land and on the shore than do the Skokholm and Skomer birds. Although argentatus frequently eats limpets Patella spp., there is little competition for food between them and Oystercatchers Haematopus ostralegus because the two species eat different sized limpets. A summary of previous studies into the food of argentatus is given. L. marinus is mainly a predator feeding on the young of other gulls, sea-birds and Rabbits but it also eats carrion, offal and fish-waste when available and during the winter. The numbers of Manx Shearwaters Procellaria puffinus killed on Skokholm and Skomer are discussed. L. fuscus obtains most of its food on the shore and on arable land. There appears to be little competition, at least in the areas studied, between .fuscus and argentatus for food.

Taxonomic characters of gadfly petrels (Procellariidae: Pterodroma spp.) are reviewed and the genus is redefined. The structure of the upper intestines, which have become helicoidally twisted to varying degrees in most species, is an important character not hitherto used. Results of a phylogenetic study of the genus based mainly on the development of helicoidal intestines agree substantially with findings from studies of the Mallophaga parasitizing these petrels. One species, the Kerguelen Petrel Lugensa brevirostris, previously classified in Pterodroma, is shown to have closer affinity with some fulmar genera. The Tahiti Petrel Pseudobulweria rostrata and its rare or extinct relatives, formerly placed in a superspecies of Pterodroma, are more closely related to the Procellaria and Bulweria petrels, and may be derived from the link between the subfamilies Procellariinae and Fulmarinae. From consideration of the fossil record, anatomical characters and Mallophaga, a phylogeny of the Procellariidae is proposed which supports its monophyletic origin. Gadfly petrels apparently descended from the fulmars through the ancestral lines of Snow Petrels Pagodroma, the Kerguelen Petrel and finally the Blue Petrel Halobaena, which may have given rise separately both to gadfly petrels and to the prions Pachyptila. A late Pliocene origin of Pterodroma in the vicinity of southern New Zealand is possible. The genus Pterodroma is divided into four subgenera, representing four connected radiations, and 29 species are recognised. In probable chronological order, Proaestrelata (new subgenus) contains five species and is restricted to the Pacific Ocean, Cookilaria comprises six species restricted as breeders to the South Pacific, Hallstroma has seven species mainly in the tropics of the Pacific, Indian and South Atlantic Oceans, whereas Pterodroma includes 11 species which extend as breeders from the southwest Pacific west to the South Atlantic and into the North Atlantic. Ecological, physical and physiological adaptations in the evolution of gadfly petrels are discussed.

Summary. 1. Large-scale excavations and observations during homing experiments provided incidental data on incubation. 2. The main period (median 90%) of egg-laying is from 25 April to 20 May, of hatching from 17 June to 12 July. The median dates for these processes are 6 May and 28 June. The average incubation period is 53 days. 3. Incubation shifts are long (average 5 days) and variable, 2 to 16 days if unrelieved. Temporarily deserted eggs were only found in about 1% of undisturbed nests. 4. Nevertheless many well-developed eggs hatched after being chilled in the burrow for up to 7 days. Others remained viable for up to 13 days in the laboratory. The length of the chilling period and the stage of development at which it occurs have little effect on the proportion surviving. 5. The ecological and evolutionary significance of the faculty is considered, with a brief survey of chilling resistance in the embryos of other birds.

Summary. The breeding distribution of the Sooty Shearwater and the dates of its breeding season are briefly reviewed. Records from the Southern Ocean are summarised. It is concluded that they are accounted for by non-breeding birds feeding off the ice-edge in the Australian sector of the Antarctic during the latter part of the breeding season. The great majority of Sooty Shearwaters spend the southern winter in the Pacific Ocean. Birds from the New Zealand colonies are probably mostly found on the feeding grounds off Japan: some may make a “circular” migration around the Pacific. Birds from the Cape Horn colonies spend the southern winter in the rich areas off the western coast of North America. Some Sooty Shearwaters spend the southern winter off the South African coasts. A number of non-breeding birds remain there throughout the summer. The Sooty Shearwaters “wintering” in the North Atlantic are probably numbered in tens of thousands only. The distribution of records is reviewed month by month and the records from Atlantic transects are summarised. The migrations in the North Atlantic are discussed in relation to prevailing meteorological and oceanographic conditions, and food supply. It is concluded that most Sooty Shearwaters in the North Atlantic make a “circular” migration, crossing the North Atlantic from west to east in June and July, and re-crossing about September further south in the northeast trades.

The report by Titian Ramsay Peale on birds encountered during the Wilkes Expedition was withdrawn for inaccuracy when few copies had been distributed, and re-written by John Cassin. A survey of the accounts of the petrels shows that this was not an improvement. Two important type localities for Procellaria brevipes and Thalassidroma lineata are probably wrong, and could be exchanged.

The decline of populations of certain seabirds and sea turtles around the world is partly related to their incidental capture in large-scale fisheries. However, the impacts of small-scale fisheries on endangered seabirds and sea turtles, being carried out in many places around the world, have been largely neglected by scientists and governments. We monitored 178 fishing days and described a range of poorly known hook-and-line commercial fisheries carried out by the Itaipava fleet, southeastern Brazil, composed by 497 vessels and deploying hooks from 18°S to 35°S. Seven fisheries were defined: fast trolling for tuna and tuna-like species, slow trolling for Bigeye tuna, handlining, surface longline for Dolphinfish, pelagic longline for Swordfish, bottom dropline, and pole-and-line with live bait. We observed bycatch of 47 seabirds of six species and 45 turtles of four species. Capture rates were higher for the surface longline for Dolphinfish (0.15 birds/1000 hooks and 1.08 turtles/1000 hooks), slow trolling for Bigeye tuna (0.41 birds/day) and handlining targeting Yellowfin tuna (0.61 birds/day). Endangered Spectacled petrel (Procellaria conspicillata), Atlantic Yellow-nosed (Thalassarche chlororhynchos), and Black-browed (T. melanophris) albatrosses were the main seabirds caught. Immature Loggerhead turtles (Caretta caretta) and immature or adult Leatherback turtles (Dermochelys coriacea) were the main sea turtles affected by the surface longline for Dolphinfish. Monitoring the fleet and bycatch levels, development of mitigation measures, establishment of educational programs, government control over the fleet, and enforcement, are urgently required for the hook-and-line fisheries described in the present study.

The southwestern Atlantic Ocean is an important foraging ground throughout the year for several albatross and petrel species. Longline fishing fleets in the region currently pose the main threat for this group of seabirds at sea, and conservation measures are urgently required. We present information on bycatch rates of seabirds in the Brazilian domestic pelagic longline fishery from 2001 to 2007, and review bycatch rates reported for the demersal and pelagic longline fisheries in the southwestern Atlantic Ocean. Overall seabird capture rate for the Brazilian pelagic longline fleet during 63 cruises (656 sets and 788 446 hooks) was 0.229 birds per 1000 hooks, varying from 0 to 0.542 according to season. Capture rates were higher between June and November (cold season) and affected mainly the black-browed albatross Thalassarche melanophris (55% of birds captured), the white-chinned petrel Procellaria aequinoctialis, the spectacled petrel Procellaria conspicillata and the Atlantic yellow-nosed albatross Thalassarche chlororhynchos. Capture rates previously reported in the SW Atlantic varied from 0 to 5.03 birds per 1000 hooks, with those reports based on logbooks or fishermen interviews tending to underestimate capture rates, whereas those based on small num- bers of hooks or short time periods tend to greatly overestimate rates in both pelagic and demersal longline fisheries. Previous studies have played an important role in delimiting the seabird bycatch problem, forming a baseline for mitigation actions and serving as a guide for improvements in data collection. However, data collected by onboard seabird-dedicated observers are more reliable, pro- vide a greater range of information relating to bycatch, and form a baseline for more robust analysis and addressing further questions. The current study highlights the stochastic nature of seabird fatal- ities in longline fisheries and the need for extensive sampling to obtain realistic estimates of capture rates covering different years, seasons, vessels, and the range of fishing gear and practices.

We studied the diet of feral cats (Felis catus) on New Island, Falkland Islands, through the analysis of 373 scats collected during the austral summers of 2004/2005 and 2005/2006. The most frequent prey were three introduced mammals (house mice Mus musculus, ship rats Rattus rattus and rabbits Sylvilagus sp.) and the thin-billed prion Pachyptila belcheri (each season present on ca. 21% of the analysed scats). These represent the first systematic data on feral cat diet for the Falklands. A simple bioenergetics model suggests that cats could be eating in the region of 1,500–11,000 prions per season, representing

Marine nitrogen (N) has been traced into terrestrial and freshwater food chains at petrel (Aves: Procellariiformes) colonies on the South Island of New Zealand, but the effects of N or other marine‐derived nutrients on the productivity of mainland forests are unknown. Mean ring widths in matai (Prumnopitys taxifolia; n = 3) and rimu (Dacrydium cupressinum; n = 2) trees growing in a Westland petrel (Procellaria westlandica) colony were 0.9–1.4 mm yr–1, similar to ring widths reported elsewhere. However, series of much wider rings in cores showed that trees in a petrel colony can experience periods of accelerated growth. δ15N values of whole wood were, at 3.2–11.1%, significantly enriched in comparison to other temperate forest trees. The highest δ15N values were at the cambial and pith ends of different cores, so isotopic fractionation during N remobilisation is unlikely to have been significant. The radial extent of 15N enrichment suggests that petrels have bred at the site since at least the mid...

In a previous study, we reported that deploying small amounts of biogenic oil derived from the livers of school sharks (Galeorhinus galeus) was effective in reducing the numbers of seabirds attending fishing vessels and seabird diving activity in the Hauraki Gulf, New Zealand. This seabird assemblage was numerically dominated by Flesh-footed Shearwaters (Puffinus carneipes) and the globally vulnerable Black Petrel (Procellaria parkinsoni). The results suggested that the oil deterrent operated via an olfactory or chemesthetic mechanism. Here we compare the deterrent effects of shark-liver oil with three other fish oils, in the Flesh-footed Shearwater–Black Petrel assemblage. We also report on trials assessing the deterrent efficacy of shark-liver oil and one other fish oil in an assemblage including four species of albatrosses (Diomedea spp., Thalasarcche spp.), Southern and Northern Giant-Petrels (Macronectes giganteus and M. halli), and Cape Petrels (Daption capense). Shark-liver oil was the most effective deterrent of the four fish oils tested in the Flesh-footed Shearwater–Black Petrel assemblage, but it did not deter albatrosses, giant-petrels or Cape Petrels from attending vessels. Thus, the deterrent was effective only on burrow-nesting seabird species. These results may provide further support for our speculation that the deterrent operates via an olfactory mechanism, while corroborating other work linking life-history characteristics of seabirds to their sensory ecology. We recommend research to investigate the mechanism and active ingredients by which shark-liver oil deters some species of seabirds. We also suggest testing fish oils on White-chinned (Procellaria aequinoctialis) and Grey (Procellaria cinerea) Petrels, and Sooty Shearwaters (Puffinus griseus), as well as in other areas occupied by Flesh-footed Shearwaters, to assess further potential for this natural oil or its ingredients to reduce by-catch of seabird species globally.

The spectacled petrel Procellaria conspicillata is listed as critically endangered due to its small population size and ongoing mortality on long-lines. Spectacled petrels were counted in 2004, repeating a census made in 1999 at their sole breeding locality, Inaccessible Island. The 2004 survey took place earlier in the breeding season than the previous count, allowing for more robust estimates of burrow occupancy. During early incubation, birds responded to call playback at 69% of burrow entrances, but birds in at least 8% of burrows remained silent. Birds in shallow burrows were less likely to respond to playback than were those in deep burrows. Two repeat trials at 100 marked nests showed that at least 61% of apparently ‘unoccupied’ burrows were occupied on subsequent checks, resulting in an overall occupancy estimate of 91%. Occupancy was equally high in peripheral colonies. The apparent spatial extent of colonies increased slightly from 1999, and the estimate of total burrow numbers increased by 50%, from 5900 burrows in 1999 to 8900 in 2004. Validation surveys indicated that burrow numbers were underestimated (84 ± 3%) to the same extent as that in 1999 (85 ± 4%), and repeat checks of one colony where all nests were marked showed that even careful counts underestimated actual numbers of burrows by up to 10%. This suggests there are some 11–12,000 burrows, and assuming 90% occupancy, the adult population is likely to be at least 20,000 birds. The population has increased over the last five years, continuing the apparent recovery from a very small population size in the early 20th century. Despite this increase, demographic models indicate that the population remains at risk from relatively small increases in mortality, if mortality is determined primarily by fishing effort. Mitigation of long-line mortality remains the key conservation goal for this species.

Longline fisheries throughout the world have frequent and often fatal interactions with seabirds. We experimentally tested one possible solution to seabird–fisheries interactions that was proposed by a New Zealand longline fisherman. This involved dripping school shark Galeorhinus galeus liver oil on the ocean surface behind fishing vessels. We tested the efficacy of shark liver oil in reducing the numbers of seabirds attending fishing vessels and the number of dives seabirds executed in pursuit of pilchard Sardinops neopilchardus baits. We conducted trials in northern New Zealand where seabird assemblages include the globally vulnerable black petrel Procellaria parkinsoni . Shark liver oil was effective in reducing both seabird numbers and dives on baits, compared to canola oil and seawater control treatments. Comparisons of seabird responses to shark liver oil and vegetable oil suggest that shark liver oil acts as an olfactory or chemesthetic deterrent for seabirds. Further work should include testing the oil with additional seabird species and investigating active ingredients and habituation of seabirds to the oil in order to assess wider opportunities for long-term use of shark liver oil to reduce seabird bycatch.

Breeding of the Grey Petrel (Procellaria cinerea) was confirmed on Macquarie Island, with 26 colonies located, and unconfirmed nesting activity found at two additional localities. Colonies comprised from one to 12 burrows in which nesting was confirmed, with the total number of burrows in which nesting confirmed for all colonies combined ranging from eight in 2000 to 59 in 2003. The total number of active burrows with breeding not confirmed in all colonies combined ranged from 58 in 2001 to 54 in 2002 and 21 in 2003. The total number of fledged chicks ranged from six (2000) to 20 (2002), with a success rate of 75% in 2000 and 80% in 2002. Breeding success declined in 2003 to 32%. Factors accounting for nest failure are discussed. Colonies were located in Poa foliosa tussock grassland on steep seaward-facing (n = 26) or near-coastal slopes (n = 2) in deep soil with little surface rock. Five environmental variables determined colony location. Numbers of burrows increased with increasing depth of soils and on distinct spur-crests with a westerly aspect. Increased breeding success at higher altitude was related to decreased Black Rat activity. Uncertainty in our estimation of the size of the breeding population on Macquarie Island was caused by the high proportion of burrows in which breeding could not be confirmed. However, based on numbers of burrows for which nesting confirmed, a minimum breeding population of 59 pairs was present in 2003. If unconfirmed nesting burrows are included, the breeding population of Grey Petrels on Macquarie Island was probably 79 and 80 pairs in 2002 and 2003 respectively.

Only a small proportion of the marine nutrients brought ashore by breeding colonial seabirds is retained in colony soils. However, means by which these marine nutrients can be dispersed to the wider terrestrial ecosystem have rarely been identified. In this study, moulted feathers from Hemiphaga novaeseelandiae (an endemic pigeon; n=9) and Gallirallus australis (an endemic flightless rail; n=3) and from a Procellaria westlandica (Westland petrel) colony were analysed for 13C and 15N enrichment. Potential diet items from both the colony and a control location were also analysed, and their contributions to the diets of H. novaeseelandiae and G. australis calculated. Hemiphaga novaeseelandiae assimilated marine N but not marine C, and showed a wide range in the significance of different diet sources. Food for most H. novaeseelandiae came from within the colony, implying minimal transport of marine nutrients to the wider terrestrial environment. All G. australis food came from within the colony, but 29–39% was from petrel tissue, or invertebrates feeding on petrel material. Thus, both H. novaeseelandiae and G. australis assimilated nutrients brought ashore by P. westlandica, but neither was shown to disperse the nutrients much beyond the petrel colony. However, birds that are now locally extinct may have dispersed marine nutrients more widely. Future research should focus on the consequences of restoring these species to petrel colonies as part of attempts to re-establish the functioning of pre-human ecosystems in New Zealand.

SUMMARY 1. Stable isotope ratios of aquatic invertebrates, aquatic mosses and leaves of riparian plants were used to determine whether marine-derived nutrients from breeding colonies of the Westland petrel (Procellaria westlandica) were incorporated into the food webs of small streams in New Zealand. 2. The d 15 N signatures of all plants and animals examined were higher by 3.6–4.6& in small streams draining catchments with petrel colonies than in nearby streams where petrels were absent. d 13 C values of leaves from terrestrial plants were also enriched by about 2& where petrels were present, but the carbon ratios of aquatic species were depleted in 13 C, rather than enriched, suggesting that any marine signal was over-ridden by isotopic shifts related to photosynthetic fractionation. 3. A high marine-nitrogen signal was maintained along the 3 km length of Scotchman Creek with the d 15 N values of leptophlebiid mayflies and predatory insects ranging from 7.4–9.5 and 9.2–11.9&, respectively. 4. Most nutrients derived from petrels are likely to be translocated to streams via the soil, which they enter in the form of excreta, spilled food, feathers, dead chicks, and abandoned eggs. However, because changes in d 15 N values are brought about by soil processes such as volatilisation of ammonia, nitrification and denitrification, it is difficult to predict the exact isotope signature of nitrogen entering a stream. Tentative estimates of the proportion of marine-derived nitrogen in stream biota, calculated using a mass-balance approach, ranged from 28–38%. 5. Our findings indicate that marine nutrients transported inland by seabirds can be incorporated into the food webs of streams. In pre-human times when there were many more seabird colonies on mainland New Zealand than exist today, marine-derived nutrients introduced by birds may have had significant effects on nitrogen cycling and the productivity of New Zealand streams.

Longline fisheries have expanded throughout the world's oceans since major commercial distant-water pelagic fleets began fishing for tuna and tuna-like species in the early 1950s. Along with the more recent development and expansion of demersal longline fleets for species such as Patagonian toothfish, these vessels are a major source of mortality to several species of seabird. Vessels can set many thousands of baited hooks in a day across many kilometres of water. These waters are often used as foraging areas by wide-ranging seabirds. Attracted by baits and offal, the birds can be caught on the baited hooks and subsequently drown. To provide a greater understanding of the potential impact of the Southern Ocean's longline fleets on seabird populations, this paper describes the trends in longline effort of the major pelagic and demersal fisheries in southern waters. The total reported effort from all longline fleets south of 30°S has been well over 250 million hooks per year since the early 1990s. However the spatial and temporal distribution of this effort has not been constant. While effort from the Japanese pelagic distant-water longline fleet declined through the 1990s, the Taiwanese fleet expanded dramatically. Likewise demersal fishing for toothfish increased markedly during the mid-1990s. These fisheries, along with substantial illegal longline fisheries, may be placing the long-term viability of many Southern Ocean species of seabird in jeopardy.

Pelagic longline fisheries worldwide incidentally take long-lived seabird species. This mortality has led to fisheries restrictions to protect seabirds at risk, including Wandering (Diomedea exulans) and Amsterdam Albatross (D. amsterdamensis) in the South Pacific and Spectacled Petrel (Procellaria conspicillata) in the South Atlantic. Because pelagic longline fisheries involve multinational fleets operating in vast ocean regions, assessing total bycatch levels for a seabird is challenging. Here we present a case study of quantifying bycatch from a basin-wide pelagic longline fishery and assessing the population-level impact on a vulnerable seabird, the Black-footed Albatross (Phoebastria nigripes) in the central North Pacific. We develop an assessment method that uses observer data to estimate bycatch for one fleet and then uses scenario analysis to estimate bycatch for remaining fleets. Our method generates a bounded estimate of bycatch within an ocean region, ranging from the worst-case to the best-case bycatch scenario. We find that Black-footed Albatross mortality across all fleets in the central North Pacific may total as much as 10 000 individuals/yr. At this level of mortality, population declines are likely. However, even at the best-case bycatch estimate (5200 individuals/yr), population declines may occur over the next three generations (60 years). Although this analysis requires extensive estimation and extrapolation from existing data, it is critical to provide fisheries managers with bounded estimates of likely population-level effects of current fishing activity. Corresponding Editor: Paul Dayton