Page 7 - Invasive_rats_2009

Invasive rats and seabirds after 2,000 years 1637

Table 4 Statistical significance of explanatory variables in GLMs on seabird abundance (b = estimate; SE = standard error of the
estimate; W = Wald statistic)
Calonectris diomedea Puffinus yelkouan Puffinus mauretanicus Hydrobates pelagicus
Parameters b SE W P- b SE W P- b SE W P- b SE W P-
value value value value

Intercept -0.43 0.43 1.00 ns -0.85 0.66 1.66 ns -0.47 1.03 0.21 ns -2.30 2.90 6.51 *
log 10 (area) 0.43 0.08 29.93 *** 0.36 0.14 6.62 * -0.00 0.00 0.28 ns -0.19 0.20 0.85 ns
log 10 (elevation) 0.22 0.22 0.99 ns 0.56 0.38 2.28 ns 0.71 0.34 4.39 * 1.25 0.49 6.42 *
Substrate 0.09 0.06 0.29 ns 0.28 0.11 6.02 * / a / a / a / a 0.31 0.19 2.53 ns
(limestone)
Rats (presence) -0.13 0.10 1.55 ns 0.01 0.29 0.00 ns 0.10 0.24 0.18 ns -0.44 0.16 7.70 **
log 10 (Dist. Near. –0.05 0.10 0.25 ns 0.17 0.15 1.34 ns 0.32 0.39 1.24 ns 0.62 0.25 6.35 *
Rats)
N 79 30 69 101
See Table 1 for more thorough information on explanatory variables
Significance levels: *** (P \ 0.001); ** (P \ 0.01); * (P \ 0.05); ns not significant
a
The substrate parameter was not tested for P. mauretanicus abundance since it exclusively breeds on limestone islands

2
source of rats was positively related to storm petrel (scaled Pearson v = 13.0; df = 8; P = 0.88) and
2
presence and abundance. storm petrel presence (v = 75.4; df = 94; P = 0.08)
2
There was no evidence of model inadequacy for and abundance (v = 23.0; df = 17; P = 0.85).
2
Cory’s shearwater presence (scaled Pearson v =
2
194.3; df = 174; P = 0.86) and abundance (v =
Evidence of rat impact on Mediterranean
78.0; df = 72; P = 0.71), Yelkouan shearwater pres-
2
ence (v = 131.8; df = 129; P = 0.58) and abundance Procellariiformes
2
(v = 30.0; df = 24; P = 0.82), Balearic shearwater
2
presence(v =50.6;df = 62;P = 0.15) andabundance The presence of rats had a negative effect on the
breeding success of Cory’s shearwaters (b =-0.25;
SE = 0.03; W = 55.35; P \ 0.001; n = 219). The
breeding success of Cory’s shearwaters was also
influenced by the year of study (b =-0.01;
SE = 0.00; W = 7.67; P \0.01) and the geographical
region (b = 0.06; SE = 0.06; W = 7.27; P \ 0.01).
Breeding success was significantly higher on rat-free
islands (rats absent or eradicated) or when rat popu-
lations were controlled within shearwater colonies,
compared to rat-infested islands where no conserva-
216 = 35.2;
tion measure was carried out (F 2,
P \ 0.001; Fig. 4). The breeding success of Cory’s
shearwaters was not significantly different between
rat-controlled and rat-free islands. Balearic shearwater
breeding success was not significantly influenced
either by rat control or by rat absence (H 2, 12 = 1.9;
Fig. 3 Proportion of rat-infested and rat-free islands among P = 0.38). Chi-square tests for goodness of fit found
the islands where the four Procellariiform species (Calonectris no evidence of inadequacy for the Cory’s shearwater
diomedea, Puffinus yelkouan, P. mauretanicus, Hydrobates 2
pelagicus) are known to breed in the Western Mediterranean. breeding success model (scaled Pearson v = 219.0;
The number of islands used for each species is indicated df = 214; P = 0.61).

123