Page 9 - Invasive_rats_2009

Invasive rats and seabirds after 2,000 years 1639

seabird species at the scale of the entire Western than on large islands. Balearic shearwaters breed only
Mediterranean Basin. This pattern is particularly on limestone islands, and when breeding on rat-
clear for the three shearwater species that virtually infested islands, breeding sites are generally located
breed only on rat-infested islands, sometimes with in inaccessible sites (i.e., deep limestone caves in
high rat densities (e.g., Zembra Island, Tunisia, high cliffs) where rat densities are low or even nil
M. Pascal, pers. comm.; Molara Island, Sardinia, (Oro et al. 2004). In turn, storm petrels preferentially
G. Spano, pers. comm.; records up to 50 individuals/ha select safe breeding areas on limestone islands.
were found in Cabrera Archipelago, Toro Island in Moreover, they breed almost exclusively on small
Corsica, and Petit Congloue ´ Islet in French Provence; islands to avoid human activities, and terrestrial and
see Alcover 1993 and Cheylan 1999). Only the avian predators on large islands (Erwin et al. 1995;
abundance of the smallest species, the storm petrel, Borg et al. 1992–1994; Oro et al. 2005).
was related to rat presence on islands. Our findings However, at a more local scale, most studies
are consistent with those of Jones et al. (2008) in their dealing with rat impact on Mediterranean Procellar-
review on rat impact on seabirds, which points out iiformes reported benefits of ship rat absence or rat
that seabirds of the Hydrobatidae family (i.e., storm removal on shearwater productivity (Table 5). More-
petrels) meet all the criteria for susceptibility to rats. over, as confirmed by our GLM analyses, the
Storm petrels are small-sized seabirds, are burrow or breeding success of the largest burrowing seabird
cavity nesters, and have all life stages (i.e., eggs, species, the Cory’s shearwater, has frequently been
chicks, adults) likely to be preyed upon. Although rat shown to be negatively affected by ship rat impact
presence on islands was not a significant component (e.g., Amengual and Aguilar 1998; Thibault 1995;
of storm petrel distribution at the scale of the Western Igual et al. 2006, 2007; Pascal et al. 2008). From a
Mediterranean, at the archipelago scale, storm petrels demographic point of view, population dynamics of
only breed on rat-free islands (Massa and Sultana Procellariform seabirds have a low sensitivity to
1990–1991; Martin et al. 2000). This is why it is changes in fecundity, while changes in adult survival
imperative to consider the scale under study when translate into sharp variations of population growth
inferences are made (Lomolino 1999, 2000; Wittaker rates (Cuthbert et al. 2004; Oro et al. 2004; Louzao
2000). et al. 2006). This low sensitivity, together with some
Seabird presence and abundance were mainly demographic buffer capacities (e.g., age at recruit-
related to island’s physical characteristics, such as ment, skip breeding, immigration), could explain how
elevation, area, and substrate. Island cliffs may shearwater populations have persisted despite centu-
constitute key intra-island refuges for seabird breed- ries of rat impact. Unfortunately, there are not enough
ing since they may be less covered by vegetation and reliable data on Yelkouan and Balearic shearwater
less accessible to mammal predators (Igual et al. breeding success and cases of breeding in rat-free
2006). On New Island, Falklands, Quillfeldt et al. islands to allow rigorous inter-species comparison of
(2008) showed that thin-billed prions preferred shearwater sensitivity to rats. This review failed to
breeding in areas with little vegetation and seemed find an overall significant difference between the
to avoid areas covered by native tussock grass where effect of rat absence (either naturally absent or
ship rat densities were consistently higher. Unlike eradicated) and local rat control on the breeding
Cory’s shearwaters, which are able to breed in success of the Cory’s shearwater. This apparent non-
exposed cavities as well as under vegetation without significance may be a result of insufficient data
substrate protection (Borg 2000), Yelkouan shearwa- available on the long-term monitoring of Cory’s
ters are more selective in their habitat requirements shearwater breeding success after rat eradications,
and breeding cavity selection (i.e., selecting deep and since rat eradication has been proven to enhance
winding cavities that may be less accessible to rats; long-term ecological benefits compared to local
Bourgeois and Vidal 2007; Ruffino et al. 2008). This control (Pascal et al. 2008). This also suggested that
may explain why Yelkouan shearwaters only rarely both rat control and rat eradication may be efficient
breed on small islands, where the probability of methods to recover shearwater breeding success.
finding suitable and safe breeding sites may be lower Unfortunately, rat densities and control efforts were

123