Predation of <scp>A</scp>tlantic <scp>P</scp>etrel chicks by house mice on <scp>G</scp>ough <scp>I</scp>sland

<jats:title>Abstract</jats:title><jats:p>The impacts of predation by invasive mammals on island fauna are a major driver of insular biodiversity loss. Devastating, hitherto unsuspected impacts of predatory house mice on breeding seabirds have been described recently. We studied the fate of 178 <jats:styled-content style="fixed-case">A</jats:styled-content>tlantic <jats:styled-content style="fixed-case">P</jats:styled-content>etrel <jats:italic><jats:styled-content style="fixed-case">P</jats:styled-content>terodroma incerta</jats:italic> nests at <jats:styled-content style="fixed-case">G</jats:styled-content>ough <jats:styled-content style="fixed-case">I</jats:styled-content>sland, over four seasons, from <jats:styled-content style="fixed-case">O</jats:styled-content>ctober 2003 to <jats:styled-content style="fixed-case">J</jats:styled-content>anuary 2008. Introduced house mice <jats:italic><jats:styled-content style="fixed-case">M</jats:styled-content>us musculus</jats:italic> were found in all study burrows checked for mouse visits. From <jats:styled-content style="fixed-case">O</jats:styled-content>ctober 2003 to <jats:styled-content style="fixed-case">S</jats:styled-content>eptember 2004, we video‐recorded attacks by mice on six (of 13) live, healthy <jats:styled-content style="fixed-case">A</jats:styled-content>tlantic <jats:styled-content style="fixed-case">P</jats:styled-content>etrel chicks and on one (of three) great shearwater <jats:italic><jats:styled-content style="fixed-case">P</jats:styled-content>uffinus gravis</jats:italic> chicks. In all years, chicks died from mouse attacks. Stage‐specific daily nest survival rates were modelled, from which estimates of breeding success were derived that accounted for the variable exposure periods studied among years. Average daily survival rate of eggs was 0.998, and hatching success through the entire incubation period (55.5 days) was 0.924 [95% confidence interval (<jats:styled-content style="fixed-case"><jats:roman>CI</jats:roman></jats:styled-content>) 0.903–0.940]. Daily chick survival rates were 0.990, which gave a modelled fledging success of 0.247 (<jats:styled-content style="fixed-case"><jats:roman>CI</jats:roman></jats:styled-content> 0.165–0.338) over the 138‐day chick period, and average annual breeding success (chicks fledged per breeding attempt) of 0.228 (<jats:styled-content style="fixed-case"><jats:roman>CI</jats:roman></jats:styled-content> 0.150–0.318), which is low compared with congeners. Productivity estimates were used as a parameter in a population simulation model, which predicted a population multiplication rate (λ) of 0.993 (<jats:styled-content style="fixed-case"><jats:roman>CI</jats:roman></jats:styled-content> = 0.966–1.021). However, in the one season studied from laying to fledging (2007), from 58 nests, only one chick fledged (1.7%). This suggests the wide errors on the model results may obscure a more severe reality. More than 60% of model simulations resulted in an <jats:styled-content style="fixed-case">I</jats:styled-content>nternational <jats:styled-content style="fixed-case">U</jats:styled-content>nion for <jats:styled-content style="fixed-case">C</jats:styled-content>onservation of <jats:styled-content style="fixed-case">N</jats:styled-content>ature classification of <jats:styled-content style="fixed-case">E</jats:styled-content>ndangered. Our results add support to calls to eradicate mice from <jats:styled-content style="fixed-case">G</jats:styled-content>ough <jats:styled-content style="fixed-case">I</jats:styled-content>sland. More generally, mice cannot be ignored as a potential threat to island fauna, and island restoration and management plans should routinely include eradication of introduced mice.</jats:p>