3. Methods
3.1. Quantification of species abundance
Estimates of the abundance of kangaroos and dingoes at each
sub-sitewere conducted using spotlight and track surveys and were
previously published in Letnic et al. (2009). We extracted the data
on the abundances of kangaroos and dingoes from Table S1 in Letnic
et al. (2009) for presentation in graphs. Because previous studies
have shown that kangaroos were more abundant in the absence of
dingoes, we used a one-tailed t-test to test the hypothesis that
kangaroos were more abundant in the absence of dingoes at each
site. Tests were calculated using the mean, standard deviation and
sample size for each sub-site presented in Letnic et al. (2009).
Daytime counts of kangaroo remains (skeletons and carcasses)
within 100 m of the road were undertaken along infrequently
travelled vehicle tracks. Counts were also taken from foot survey in
random directions off vehicle tracks. There was no discernible
difference in carcass abundance between the two methods. Skeletons
were recorded as kangaroos if the characteristic tibia or skull
was intact (Reed, 2001). For each remain, the following characteristics
were recorded: condition (presence of skin, articulation/
disarticulation), the perpendicular distance from the road (metres),
and provenance (latitude and longitude) with a global positioning
system. For each transect we recorded the total distance travelled
(km).We calculated the density of dead kangaroos as the number of
dead kangaroos sighted per hectare.
We used a paired t-test to compare the density of dead kangaroos
between paired sub-sites on either side of the dingo fence. The
correlation between live kangaroo abundance and the density of
dead kangaroos was investigated using Spearman’s rank correlation.
Contingency table analysis (c2) was used to investigate if the
degree of articulation of kangaroo remains differed on either side of
the dingo fence