Selvagem Grande (30°09′N, 15°52′W) is the largest island of the Portuguese sub-archipelago of the Selvagens and lies between Madeira and the Canary Islands in the north-east Atlantic. The island supports one of the most important Atlantic colonies of Cory's shearwaters Calonectris diomedea borealis (Granadeiro et al., Reference Granadeiro, Dias, Rebelo, Santos and Catry2006) and the birds have been studied intensively since the expedition organised by the Museu Municipal do Funchal (História Natural) in 1963 (Pickering & Maul, Reference Pickering and Maul1965; Zino, Reference Zino1971, Reference Zino1985; Zino et al., Reference Zino, Zino, Maul and Biscoito1987; Mougin et al., Reference Mougin, Jouanin, Roux, Zino and Zino2002a,Reference Mougin, Jouanin, Roux, Zino and Zinob; Granadeiro et al, Reference Granadeiro, Dias, Rebelo, Santos and Catry2006). Two study areas of c. 15 * 100 m (site EFGH) and 16 * 140 m (site ABCD) on the island's cliff slopes have been defined for long-term monitoring (Zino et al., Reference Zino, Zino, Maul and Biscoito1987). These areas are on opposite sides of the island, one facing the north-east trade wind and one in the lee of it. Frequent expeditions were made to the sites in order to conduct the monitoring, with special emphasis on the time of egg laying (late May to mid June) and prior to fledging (late October). The breeding cycle of the birds on the island is remarkably consistent and predictable. After a period of recovery following a devastating illegal cull of birds in 1976 (Zino, Reference Zino1985), demonstrated in the increasing number of nests monitored in the years between 1984 to 1989, the numbers of nests studied did not vary appreciably among the annual study periods (Fig. 1).

Fig. 1 Percentage of monitored Cory's shearwaters nests on Selvagem Grande that produced fledged chicks (means for study areas ABCD and EFGH combined). Stippled bars are before eradication of mice and rabbits, and black bars after eradication. Values at the top of bars are the numbers of nests monitored. Rabbits and mice were eradicated during the 2002 breeding season. In breeding seasons for which no data are presented, logistical problems prevented adequate study. The number of nests studied in 2005 represents c. 1.5% of the total number of breeding pairs estimated to be present on the island in that year (Granadeiro et al., Reference Granadeiro, Dias, Rebelo, Santos and Catry2006.).

Populations of the rabbit Oryctolagus cuniculus and the house mouse Mus musculus inhabited the island, probably the result of ancient introductions by mariners, but were removed in a successful eradication programme carried out by the Parque Natural da Madeira involving the intensive application of baits containing the anticoagulant rodenticide brodifacoum (Oliveira et al., Reference Oliveira, Trout, Menezes, Geraldes, Buckle, Domingues and Rodrigues-Luengo2003). The removal of the majority of mammals occurred during August 2002, coinciding with the period of hatching of Cory's shearwater chicks. Rabbits and house mice were the only alien mammals present on the island.

The removal of rats (Rattus spp.) from breeding colonies of seabirds can result in significant rises in productivity and survival because of their predation of adults, eggs and young (Zino et al., Reference Zino, Oliveira, King, Buckle, Biscoito, Costa Neves and Vasconcelos2001; Courchamp et al., Reference Courchamp, Chapuis and Pascal2003; Buckle, Reference Buckle2004). Mice, mainly house mice, but also Peromyscus spp. and others, also inhabit many oceanic islands and in spite of their small size are able to predate nesting seabirds and their eggs and chicks (Blight et al., Reference Blight, Ryder and Bertram1999: Drever et al., Reference Drever, Blight, Hobson and Bertram2000; Cuthbert & Hilton, Reference Cuthbert and Hilton2004; Wanless et al., Reference Wanless, Angel, Cuthbert, Hilton and Ryan2007). The removal of rabbits has also been shown to be beneficial in terms of island conservation (Courchamp et al., Reference Courchamp, Chapuis and Pascal2003). However, the principal biota that benefit are plants and the major effects on breeding seabirds are probably indirect, mediated by vegetation changes, reduced soil erosion and the removal of the prey base of raptors that may take seabirds as secondary prey items (Priddel et al., Reference Priddel, Carlile and Wheeler2000; Bullock et al., Reference Bullock, North, Dullo, Thorsen, Veitch and Clout2002).

The existence of a long-term data set on the breeding success of Cory's shearwater on Selvagem Grande provided a unique opportunity to study the benefits of the removal of the two alien invasive species, rabbits and house mice. Ideally, in breeding studies of seabirds, nests are monitored throughout the breeding season, the breeders ringed, the date of laying observed, and losses throughout incubation and the dates of hatching and fledging are noted (Zino et al., Reference Zino, Zino, Maul and Biscoito1987). It has not been possible, however, to obtain such complete data sets in every breeding season of Cory's shearwaters on Selvagem Grande due to the difficulty of access to the island. In most years the ringing of the juveniles took place after late September and losses of juveniles after this date and prior to fledging are known to be minimal. Therefore, the numbers of nests monitored and of chicks ringed are used in this study as indicators of annual productivity in the two study areas.

A significant increase in Cory's shearwater productivity was observed in the years following the eradication programme in both study areas. (Table 1, Fig. 1). Fledgling productivity in study area ABCD increased from a mean of 36.5 fledglings per 100 nests monitored in 13 breeding seasons before eradication to one of 53.7 per 100 nests in the 5 subsequent seasons. A similar effect was observed in study area EFGH, with fledgling productivity increasing from a mean of 45.4 to one of 55.7 per 100 nests. These increases represent, respectively, 47 and 23% more young birds in nests immediately prior to fledging after rabbit and mouse eradication.

Table 1 The mean percentage of Cory's shearwater nests that produced fledglings in two study areas of Selvagem Grande before and after the eradication of rabbits and mice, together with two-tailed t-tests for the difference of means with unequal variances, hence the different degrees of freedom.

Until Selvagem Grande was classified as a Nature Reserve in 1971, and permanently wardened from 1980, the island's Cory's shearwaters were intensively, and latterly illegally, exploited for meat, oil and feathers (Mougin & Stahl, Reference Mougin and Stahl1982; Zino, Reference Zino1985). This resulted in a population decline from an estimated 130,000-150,000 breeding pairs to 7,000 pairs in 1980 (Mougin & Mougin, Reference Mougin and Mougin2000). Since the last extensive culls in 1975 and 1976 the population of Cory's shearwaters on Selvagem Grande appears to have been increasing at an annual rate of c. 5% and has now recovered to an estimated 29,540 breeding pairs (Granadeiro et al., Reference Granadeiro, Dias, Rebelo, Santos and Catry2006). It is also clear, however, that a highly significant increase in the number of fledglings produced in the two study areas in the last five consecutive breeding seasons has coincided with the removal of rabbits and mice from Selvagem Grande. If the increase seen in the study plots is typical of the nests on the island as a whole it is likely that the rate of growth of the island's population will also increase. A similar increase in the breeding success of Cory's shearwater was found by Igual et al. (Reference Igual, Forero, Gomez, Orueta and Oro2006) who controlled, but did not eliminate, black rats Rattus rattus at Cory's shearwater breeding sites on the Chafarinas Islands in the Mediterranean.

Our observations do not prove a causative link between the eradication of alien mammals and improved breeding success of Cory's shearwaters. However, competition between rabbits and seabirds for burrows has been noted on some oceanic islands (Courchamp et al., Reference Courchamp, Chapuis and Pascal2003) but, as previously stated, Selvagem Grande once supported, in the presence of rabbits, a much larger population of Cory's shearwater than it does now. Thus, the removal of competition for burrows seems an unlikely explanation for their increased breeding success, although disturbance may have been important. Also, if the eradication of rabbits resulted in a benefit to the birds through reduced soil erosion and vegetation recovery we would have anticipated a more gradual change in breeding success than was apparent. Mice may have direct impact on breeding seabirds through predation (Blight et al., Reference Blight, Ryder and Bertram1999; Cuthbert & Hilton, Reference Cuthbert and Hilton2004; Wanless et al., Reference Wanless, Angel, Cuthbert, Hilton and Ryan2007). However, while very high densities of mice frequently occurred on Selvagem Grande, no obvious sign of their predation of Cory's adults, chicks and eggs was ever seen. Therefore, attributing the increased fledging rate of Cory's shearwater either to mouse or rabbit removal, or to a combination of both, must remain speculative. Nevertheless, the adverse impacts of both rabbits and mice on the breeding of nesting seabirds is well established (Courchamp et al., Reference Courchamp, Chapuis and Pascal2003) and their removal from Selvagem Grande appears to have had an immediate and dramatic effect in increased breeding productivity of Cory's shearwater. The increase was apparent in the 2002 breeding season during which the rabbit and mouse eradication took place. This suggests that the impact of the alien mammals may have been on the birds’ chicks, rather than their eggs, because the mammals were present throughout laying and incubation in 2002 and effective removal occurred only around the time of hatching. In spite of their necessarily speculative nature, these observations will provide encouragement and stimulus to those planning programmes for the removal of alien vertebrates from other oceanic islands.