Incidental and intentional catch threatens Galápagos waved albatross

As large, long-lived seabirds with delayed and slow reproduction, albatrosses have low intrinsic mortality rates and are especially vulnerable to extinction from extrinsic sources of mortality such as fishery bycatch. Leg-band recovery information for waved albatrosses revealed mortality from both incidental catch and intentional catch for human consumption. Annual adult survival in 1999-2005, estimated from capture-mark-recapture data, was lower than historical estimates. This recent increase in adult mortality probably contributed to recent and dramatic shrinkage of the breeding population; periodic matrix models confirm that population growth rate is most sensitive to changes in adult survival. Banding data and recovery information also suggest that capture by fisheries is male-biased, which should reduce fecundity in this species with obligate bi-parental care. This new documentation of bycatch, harvesting, and associated demographic consequences provides reason for serious concern about the persistence of the single breeding population of the waved albatross.     

Population trends in a community of large Procellariiforms of Indian Ocean: Potential effects of environment and fisheries interactions

Despite the worrying conservation status of several albatross and petrel population, the long-term trends of many populations remain largely unknown and the causes of decline in many cases are known or very strongly suspected to be incidental mortality in fisheries. Here we combine long-term monitoring of population trends, breeding success and band recoveries to examine the past and current status of five species of albatrosses and giant petrels breeding at the same site: sooty albatross (Phoebetria fusca), light-mantled albatross (Phoebetria palpebrata), wandering albatross (Diomedea exulans), northern giant (Macronectes halli) and southern giant petrels (Macronectes giganteus) on Possession Island, Crozet archipelago. We identified three groups of trends over a 25-years period (1980-2005) suggesting common underlying causes for these species in relation to their bioclimatic foraging ranges. The Antarctic species - light-mantled albatross and southern giant petrel - appeared stable and increased recently, the Sub-Antarctic species - wandering albatross and northern giant petrel - declined with intermediate periods of increase, and finally the subtropical species - sooty albatross - declined all over the period. Breeding success, indicative of environmental conditions, showed two kinds of pattern (low and fluctuating versus high and/or increasing) which were consistent with oceanographic variations as found in a previous study. We present the analysis of fisheries-related recoveries, indicative of fisheries bycatch risks showing specific catch rates. No direct relationship between population trends and longline fishing effort was detected, probably because census data alone are not sufficient to capture the potentially complex response of demographic parameters of different life stages to environmental variation. This study highlights the contrasted changes of procellariiform species and the particularly worrying status of the subtropical sooty albatrosses, and in a lesser extent of Sub-Antarctic species.     

The use of artificial breeding chambers as a conservation measure for cavity-nesting procellariiform seabirds: a case study of the Madeiran storm petrel (Oceanodroma castro)

Procellariiform seabirds are amongst the most severely threatened taxa worldwide. Whilst the specific threats vary among species, problems such as introduced predators, loss of suitable nesting habitat and inter- and intra-specific competition for nest sites are common to many situations. This paper reports on an experiment to increase the availability of secure nest sites at a colony of Madeiran storm petrels breeding on an islet (free from introduced predators) in the Azores islands where there is evidence of inter-specific competition. The provision of nest boxes that were designed to exclude larger species led to around a 12% increase in the size of the breeding population in the first year and a 28% increase over the original colony size in the second year. Over three seasons, the breeding success of storm petrels nesting in boxes averaged 2.9 times greater than that of birds at natural sites, despite their likely reproductive inexperience. The techniques presented here should serve as a model for conservation management of other burrowing procellariiform species in situations where colony size or breeding performance is believed to be constrained by nest site quality.     

Expectations for population growth at new breeding locations for the vulnerable New Zealand sea lion (Phocarctos hookeri) using a simulation model

Management plans for threatened or recovering large vertebrate species that are increasing in population size and range focus on the establishment of viable populations within set temporal limits. New Zealand (Hookers) sea lions (Phocarctos hookeri) were declared a threatened species in 1997, and New Zealand legislation requires that threatened species of marine mammals must be managed to reduce human-induced mortality and achieve a non-threatened status within 20 years. The present breeding distribution of P. hookeri is highly localised, with over 95% of total annual pup production located at Auckland Islands and almost all of the remainder at Campbell Island. Breeding elsewhere has been ephemeral or restricted to <10 adult females. The only recorded sustainable breeding at a new location has been at Otago, South Island, New Zealand. This breeding population consisted of a total of four breeding females in 2002 and is derived from one immigrant female that gave birth to her first pup in the 1993/1994 breeding season. The New Zealand Department of Conservation management plan specifies that to achieve a non-threatened status P. hookeri (1) at Otago must increase in the number of breeding females to ?10, and (2) must establish ?two new breeding locations within the 20-year time frame, each with ?10 breeding females. This study 1) projects the population growth trends at a new location (Otago) to see if it will achieve ?10 breeding females within the legislated time frame, and (2) examines the likelihood that other breeding locations will establish elsewhere given the demographic information available for this species. We present 20 deterministic and three stochastic Leslie matrix model scenarios for female population growth for the initial years following the start of breeding at a new location. Our results indicate that (1) a new breeding population derived from one immigrant female is unlikely to reach 10 breeding females in 20 years; this duration is more likely to be 23-41 years (deterministic models) or 23-26 years (stochastic model), (2) the likelihood of two new sites establishing within 20 years is unquantifiable, but the probability is low, and (3) if the legislated outcome and time limit are not revised in the population management plan, the feasibility and effectiveness of re-locating young females could be investigated.     

Estimating the ability of birds to sustain additional human-caused mortalities using a simple decision rule and allometric relationships

Many bird species are subject to human-caused mortality, either through direct harvest (e.g. game birds) or through incidental mortalities (e.g. fisheries-related bycatch of seabirds, impact with vehicles, wind turbines, or power lines). In order to assess the impact of additional mortalities on birds, both the number of birds killed and their ability to sustain those deaths must be estimated. Niel and Lebreton [Niel, C., Lebreton, J.-D., 2005. Using demographic invariants to detect overharvested bird populations from incomplete data. Conservation Biology 19, 826-835] applied a simple decision rule [Wade, P.R., 1998. Calculating limits to the allowable human-caused mortality of cetaceans and pinnipeds. Marine Mammal Science 14, 1-37] to estimate the level of additional human-caused mortality or potential biological removal (PBR) that can be sustained for bird species given only (1) estimates of the population size, adult survival, and age at first breeding, and (2) the current population status and management goals. We provide guidelines for appropriate use of the method and case studies comparing results from this method to other approaches. Particular focus is placed on applying the method to Procellariiformes.PBR limits may then be set without a population model and when monitoring levels are minimal, and in a computationally straightforward manner. While this approach has many advantages, there are limitations. The PBR rule was initially developed for cetaceans and pinnipeds and there have been no adaptations for the unique biology of birds which may need further consideration. Additionally, because this is a simplifying method that ignores differences in life stages, it may not be appropriate for very small populations or for those listed as ‘critically endangered’, and further work is needed for situations where mortalities have large gender or age bias.     

Resolving conflicts between butterfly host resource abundance and genet population size estimates for a vegetatively spreading, threatened grassland legume

Population size estimates are an integral part of rare plant conservation, but common abundance measurements such as cover and ramet number may not accurately index genet population size for vegetatively spreading species. Population monitoring of Kincaid’s lupine (threatened species) populations occurs through genet-anonymous leaf cover and raceme counts despite extensive, non-adventitious rhizome growth. While the current monitoring scheme provides important resource abundance measurements for the endangered Fender’s blue butterfly, whose larvae feed on Kincaid’s lupine leaves, the methods are not appropriate for estimating lupine genet number. Major axis regression revealed well supported statistical relationships between cover, raceme, and plantlet (a measurement of modular plant growth) density within six study patches (n = 3 populations) of Kincaid’s lupine (R² > 0.90) and when all patch data were combined (R² > 0.91). Genet population size estimates from genotype only data with ACE (an estimator used to infer species richness) were similar to estimates derived from a combination of plantlet density and genet to plantlet ratios (genotype derived) in small, more thoroughly genotyped lupine patches. However, genet number estimates from ACE were 3–5-fold greater in less intensely genotyped patches. Genet-anonymous plant abundance measurements, such as cover, can be used to estimate genet number in populations of vegetatively spreading plants provided they are calibrated with a unit of modular plant growth. Calibration of vegetative measurements, consistency of between population relationships, and closer scrutiny of highly supported statistical models may be necessary to develop more pertinent monitoring methods for rare, vegetatively spreading plants.     

Demography of Asian elephants (Elephas maximus) at Uda Walawe National Park, Sri Lanka based on identified individuals

We provide estimates of population size and other demographic variables for individually-identified Asian elephants (Elephas maximus) in Uda Walawe National Park (UWNP), Sri Lanka based on systematic year-round observations. Two hundred and eighty-six adult females and 241 adult males were identified, of which four adults (2% of males) had tusks. Sightings-based demographic models showed seasonal immigration and emigration from the study area. The total population, including non-adults, was between 804 and 1160 individuals. Density ranged from 102 to 116 adult females per 100 km² and remains at this level throughout the year. This large, un-fragmented population of Asian elephants should be of high conservation priority. We find that estimates of survivorship and migration rates should be based on long sampling intervals when possible, but estimates of density and population size can still be made when observations are constrained to shorter intervals, if spatial data are available. We offer suggestions to guide census design for other elephant populations or cryptic species. We urge that other locations be systematically surveyed as well using photographic identification.     

Ingestion of fishing gear and entanglements of seabirds: Monitoring and implications for management

Fisheries are increasingly adopting ecosystem approaches to better manage impacts on non-target species. Although deliberate dumping of plastics at sea is banned, not all fisheries legislation prohibits discarding of gear (hooks and line) in offal, and compliance is often unknown. Analysis of a 16 year dataset collected at South Georgia indicated that the amount of gear found in association with wandering albatrosses was an order of magnitude greater than in any other species, reflecting their wider foraging range and larger gape. Unlike other taxa, most gear associated with grey-headed albatrosses was from squid and not longline fisheries, and mistaken for natural prey rather than the result of direct interaction. Observed rates of foul-hooking (entanglement during line-hauling) were much higher in giant petrels and wandering albatrosses than black-browed albatrosses, and no grey-headed albatross was affected. The index of wandering albatross gear abundance showed two peaks, the most recent corresponding with a substantial increase in the number of multifilament snoods (gangions), suggesting that the widespread adoption of a new longline system (Chilean mixed) may have been responsible. Although all identified gear was demersal, given the widespread use of similar hooks, little could be assigned to a specific fishery. Stomach content analysis suggested that 1300-2048 items of gear are currently consumed per annum by the wandering albatross population. Many hooks are completely digested by chicks, long-term effects of which are entirely unknown. We suggest a number of management approaches for addressing the problem of gear discarding, and guidelines for monitoring schemes elsewhere.     

Demography and conservation of the Tristan albatross Diomedea [exulans] dabbenena

The Tristan albatross Diomedea [exulans] dabbenena is the third rarest albatross species, with a breeding population of around 1500 pairs almost totally restricted to Gough Island in the Tristan da Cunha group, central South Atlantic. During January 2000, the entire breeding population of Gough was surveyed for the first time since 1956, and 2400 incubating pairs were counted. An analysis of the areas that are likely to have been surveyed most accurately in the past suggests that the population has decreased by around 28% over 46 years. The number of large chicks counted over three successive seasons (1999-2001) was highly variable (range 318-1129). The average count over this period (705 chicks) is less than counts made in 1979 (792) and 1982 (798). A total of 656 chicks were counted in September 2001, giving an island breeding success of just 27.3%. However, breeding success varied considerably in different areas of the island, ranging from 17.6 to 68.0%. During the 2001 season most breeding failures were of large chicks, and over 4 years where data were available, 75% of breeding failures occurred during the chick period. Predation by introduced house mice Mus musculus is the most likely cause of chick mortality. In a small study population, birds began breeding at an average age of 9.7 years and annual adult survival from 1985 to 2001 was 92.6% (SE=1.6%). Both breeding success and adult survival estimates are low in comparison with other Diomedea species and population modelling predicts a population decreasing at an annual rate of 2.9-5.3%. Further research is needed urgently to assess whether breeding success is typical, and to confirm that mouse predation is the cause of chick mortality. The low productivity of this species will compound the negative impacts of longline fishing mortality, which are likely to be reducing adult and juvenile survival.     

From both sides: Dire demographic consequences of carnivorous mice and longlining for the Critically Endangered Tristan albatrosses on Gough Island

The IUCN recently uplisted the Tristan albatross (Diomedea dabbenena) to Critically Endangered. Here we present new data indicating negative population trends on Gough Island arising from low adult survival (∼91%, ascribed to accidental mortality on fishing gear) and low breeding success (averaging 32%, due to mouse predation). Fledgling production from 1979 to 2007 and numbers of incubating adults from 1956 to 2007 have both decreased by ∼1% p.a. Consecutive annual counts of incubating adults and a population model permit the first reliable estimates of the Tristan albatross population, presently 5400 breeding adults and 11,300 birds in all age- and stage-classes. Population models explore scenarios of likely demographic trends using combinations of hypothetical best-case estimates vs. observed estimates for two key parameters: adult survival and breeding success. These scenarios highlight the relative benefits to the species of eradicating mice or mitigating bycatch. The model scenario using observed estimates predicts annual growth rate at −2.85%. Adult survival rates have probably decreased in recent years, concomitant with increased longline fishing effort, which might explain the discrepancy between counts and modelled trends. Negative trends cannot be reversed by improving breeding success alone, and adult survival must exceed an improbable 97% to balance the current chick production. A worst-case scenario including a fixed number of adult deaths annually predicted a catastrophic 4.2% p.a. decrease and extinction in ∼30 years. Population growth was most sensitive to adult survival, but even using an adult survival estimate without fishery mortality, current breeding success is insufficient to maintain the population. These findings do not support the ‘compensatory mitigation of bycatch’ model (offsetting bycatch impacts by eradicating invasive species), and the impacts of both fishery mortality and mouse predation must be addressed to improve the conservation status of the Critically Endangered Tristan albatross.     

Use of a population model to assess the impact of longline fishing on wandering albatross Diomedea exulans populations

An age-structured model of a wandering albatross Diomedea exulans population is developed to stimulate population trends over time, using demographic parameters from the population at Possession Island, Crozets, during 1968–1986. The simulation results portray a population decreasing at a rate of 2·29% per year, which concurs with global population trends. Sensitivity analyses of model parameters indicate that both adult and juvenile mortality are contributing to the decrease. Wandering albatross mortality is presumed to have increased as a result of deaths caused by longline fishing vessels; such deaths are likely to be relatively more frequent among young, naive birds. The model is used to investigate the potential impacts of new longline fisheries such as that for Patagonian toothfish Dissostichus eleginoides in Antarctica. Assuming longline fishing operations affect juveniles more than adults, there is a time lag of 5–10 years before further decreases in population numbers are reflected in the breeding population. Also, because wandering albatrosses are long-lived, population growth rates take approximately 30–50 years to stabilize after a perturbation. Consequently, caution must be exercised when interpreting population trends; short-term (<20 year) estimates may not provide good indications of long-term trends.     

At-sea distribution of waved albatrosses and the Galápagos Marine Reserve

Albatross populations worldwide are threatened by incidental takes in longline fishery operations. The recent establishment of the Galápagos Marine Reserve (GMR) is relevant to the longline bycatch issue, as it prohibits industrial longlining in the vicinity of the major nesting site of waved albatrosses (Phoebastria irrorata). However, the legality of the fishing protections is being challenged, highlighting a need for data on use of the GMR by albatrosses. We used satellite tracking over a total of four breeding seasons to determine the distribution of waved albatrosses inside and outside the GMR, and thereby assess the degree of protection that GMR provisions offer to this species. During the incubation period, breeding adults made commuting trips from the nesting island (Isla Española) to the Peruvian upwelling zone, traveling north, east, and south after leaving the nest. During the brooding period, the distribution contracted markedly, and most satellite fixes were within the GMR. During the rearing period, breeders performed both long trips outside the GMR and short trips within. The southeastern portion of the GMR is used throughout the incubation, brooding, and early rearing periods by breeding waved albatrosses. Indirect information from non-breeding adults indicates that they are likely to use the waters of the GMR extensively.     

The conservation status of the spectacled petrel Procellaria conspicillata

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.     

The roseate tern: World distribution and status of a threatened species

The roseate tern Sterna dougallii Montagu (Aves, Laridae) has a nearly cosmopolitan breeding distribution. The North American and European populations, the only ones which have been well-documented, have both declined drastically in the past decade, raising grave concern over the species' future. Literature search and correspondence reveal that the species currently maintains substantial breeding populations in the Indian Ocean, Caribbean, and Australasian regions. Most old reports provide no quantitative population estimates, while even many recent reports provide only vague estimates of population size. Nonetheless it has been possible to estimate regional populations in most cases to within one order of magnitude. The North American and Caribbean populations total about 10 000 pairs, while the European population is about 1500 pairs. The African population, including the Seychelles and the western Indian Ocean, may have 15 000 to 22 000 pairs. The documented world population is between 20 000 and 30 000 pairs, but the breeding populations of many areas are not yet documented. Substantial variation within and between regions in appearance and breeding biology indicates that each of the five named forms, at least, should be studied and managed separately.     

Assessing the impact of fisheries, climate and disease on the dynamics of the Indian yellow-nosed Albatross

Many seabird populations are currently decreasing, especially albatrosses for which the primary threat is recognised to be mortality in fisheries. Introduced predators, climate change and other factors such as diseases can also have large impacts on seabirds. Here, we assessed the relative effect of three potential threats: climate, fisheries and diseases on the demography of an endangered marine predator and modelled its population dynamics to project its size under different scenarios. We based our study on a long-term monitoring of a colony of individually marked Indian yellow-nosed albatrosses at Amsterdam Island, subtropical Indian Ocean, that has declined during the past twenty years. We found no evidence for an impact of legal tuna longlining on demographic parameters. Hatching success was lower during El Niño years but survival (0.902 ± 0.011) was not affected by climatic factors. Avian cholera caused high chick mortality (0.808 ± 0.181) which in turn probably triggered the high emigration rate (0.038 ± 0.011) through dispersal of failed breeders. This colony has a high risk of extinction. However, the rest of the population at Amsterdam Island seemingly not affected to the same extent, declined but stabilised since 1998. Matrix models indicated that lowered adult survival and the very low breeding success, resulting in low recruitment, have both contributed to the decline of the yellow-nosed albatross colony until the mid-1990s, but that more recent decline was primarily caused by low fledging success. Our results highlight that potential threats such as fisheries, diseases or climate have to be considered simultaneously to disentangle their roles when assessing the conservation status of a marine predator species.     

What genetics tell us about the conservation of the critically endangered Balearic Shearwater?

The Balearic shearwater Puffinus mauretanicus is one of the most critically endangered seabirds in the world. The species is endemic to the Balearic archipelago, and conservation concerns are the low number of breeding pairs, the low adult survival, and the possible hybridization with a sibling species, the morphologically smaller Yelkouan shearwater (P. yelkouan). We sampled almost the entire breeding range of the species and analyzed the genetic variation at two mitochondrial DNA regions. No genetic evidence of population decline was found. Despite the observed philopatry, we detected a weak population structure mainly due to connectivity among colonies higher than expected, but also to a Pleistocene demographic expansion. Some colonies showed a high imbalance between immigration and emigration rates, suggesting spatial heterogeneity in patch quality. Genetic evidence of maternal introgression from the sibling species was reinforced, but almost only in a peripheral colony and not followed, at least to date, by the spread of the introgressed mtDNA lineages. Morphometric differences were not correlated with mtDNA haplotypes and introgression is probably due to a secondary contact between the two species several generations ago. Overall, results suggested that the very recent demographic decline in this critically endangered species has not yet decreased its genetic variability, and connectivity found among most colonies should help to reduce species extinction risk. Spreading of introgression should be monitored, but the species is not jeopardized at the moment by genetic factors and the major conservation actions should concentrate at enhancing adult survival.     

Closed mark-recapture models to estimate species richness: An example using data on epigeal spiders

Species richness is a fundamental ecological property. The problem of estimating the number of species is quite similar to that of estimating the population size of a single species. Some authors of mark-recapture statistics have proposed using some of these methods to estimate species richness. This necessitates understanding how the recording probabilities of individuals differ from those of species. In particular, the species of a species pool are likely to exhibit a wide range of recording probabilities. Depending on sampling conditions, temporal or spatial variation in species detection probability may also occur, making model M _th estimators particularly useful. Empirical detection probabilities and estimates of species numbers using three coverage, one point, and two jackknife estimators are presented for series of spatial and temporal trapping occasions of epigeal spiders.     

A global assessment of the impact of fisheries-related mortality on shy and white-capped albatrosses: Conservation implications

Hundreds of thousands of seabirds are killed each year as a result of interacting with longline and trawl fishing operations, and the severity of the impact varies regionally. Shy and white-capped albatrosses, Thalassarche cauta and Thalassarche steadi respectively, are phenotypically similar species known to be incidentally killed by fishing operations. The magnitude of this mortality has not previously been assessed across their range. Here we examine recent effort and bycatch rates in fisheries known to incidentally kill these species and qualitatively evaluate the level of impact of each fishery. Results indicate that over 8500 of these albatrosses may be killed annually, although the reliability of this estimate is low due to the paucity of comprehensive observer data in most fisheries. Of the estimated deaths of all seabird species in the fisheries assessed, trawl and longline fisheries killed birds in approximately equal proportions, but when the mortality levels of shy-type albatrosses were examined, trawl fisheries were responsible for 75% of all deaths. Data suggest most of these birds were killed in South African, Namibian and New Zealand demersal trawl fisheries and the South Africa pelagic longline fishery. Because most adult shy albatrosses are comparatively sedentary and rarely found outside Australian waters, it is primarily juvenile shy albatrosses that regularly encounter fishing fleets known to kill large numbers of albatrosses. In contrast, throughout most of their range juvenile and adult white-capped albatrosses are exposed to fisheries that collectively kill many thousands of these albatrosses each year. These data emphasise the urgent need for robust assessments of the impact of bycatch at a species and population level, and the urgent implementation of effective mitigation measures.     

The Prince Edward Islands: A sanctuary for seabirds in the Southern Ocean

This paper reviews the breeding distribution and numerical status of selected species of seabirds at 23 oceanic island groups in the Southern Ocean (between latitudes 35 and 70°S), based on census data and population estimates which have become available through increased scientific endeavour in the region during the last 15 years or so. The paper focuses on the avifauna of the Prince Edward islands, in the southern Indian Ocean, and points to the importance of these islands as a breeding station and sanctuary for seabirds. Only one oceanic island group, the Crozets, contains more species of breeding seabirds than the Prince Edward islands. Substantial segements (about 10% and greater) of the world breeding populations of the king and macaroni penguins, and the wandering, grey-headed, yellow-nosed and sooty albatrosses are based at the Prince Edward group. We recommend that Prince Edward Island (sensu stricto) should be a wilderness area, and that development at Marion Island should be kept to an absolute minimum.     

Reducing incidental mortality of seabirds with an underwater longline setting funnel

A demersal longline fishery for Patagonian toothfish (Dissostichus eleginoides) that commenced off the Prince Edward Islands during 1996 has killed significant proportions of locally breeding albatrosses and petrels. As one of a suite of mitigation measures, we tested the efficacy of a Mustad underwater setting funnel to reduce incidental mortality of seabirds. The funnel, which deploys the longline 1-2 m beneath the sea surface, was used on 52% of 1714 sets (total effort 5.12 million hooks) over a 2-year period. Used in conjunction with a bird-scaring line, overall seabird bycatch rate was low (0.022 birds per 1000 hooks), and was dominated by white-chinned petrels (Procellaria aequinoctialis) (88% of the 114 birds killed). Bycatch rate was three times lower when the funnel was used both by day and at night. Daytime catch rates with the funnel were less than those attained during night sets without the funnel. In conjunction with other mitigation measures, underwater setting offers a significant reduction in seabird mortality in this fishery and could increase fishing efficiency by allowing daytime setting. However, small numbers of albatrosses were caught during daytime sets with the funnel, and its use for daytime sets should be closely monitored.