Journal of the Royal Society of Western Australia, 86(2), June 2003 
Table 2 
Number of native species and percentage cover of various types, with 95% confidence intervals (n = 20 quadrats, over 4 transects). 
Burnt fenced 
Burnt unfenced 
Unbumed unfenced 
Number of native species 
5.6 (4.9-6.3) 
4.3 (4-4.6) 
3.6 (3.3-3.9) 
% cover of native species 
91 (89-93) % 
57 (51-63) % 
68 (61-75) % 
% cover of Trachyandra 
1 (0-2) % 
11 (7-15) % 
13.7 (11.7-15.7%) 
% cover of introduced species 
(including Trachyandra) 
4.5 (3-6) % 
20 (15-25) % 
22 (15-29) % 
% Bare earth 
4.5 (2.5-7) % 
23 (19-27) % 
10.5 (7.5-13.5) % 
o 
oo 
§ 2 
o ^ 
"So 
oo 
o 
burnt 
fenced 
burnt 
unfenced 
unburnt 
unfenced 
Figure 5. A: Number of native species in the three areas (burnt fenced, burnt unfenced 
species in the three areas (burnt fenced, burnt unfenced and unburnt unfenced). 
and unburnt unfenced). B: % cover by native 
increased the number of native species present, 
particularly if the area was fenced against quokkas (Table 
2; Fig 5A). The ANOVA for log of species count was 
highly significant for different groups (P < 0.001). The 
percentage cover of native species in the three areas also 
showed significant differences (P < 0.001; Table 2). In 
unfenced and unbumed heath, percentage cover of native 
plants was variable (Fig 5B). Where the heath had been 
burnt but not fenced, cover was low indicating that 
where quokkas could graze the new growth the 
vegetation had not recovered even after 4.5 years. By 
contrast the cover of native species in burnt areas fenced 
against grazing was high. 
Melaleuca/Callitris 
Melaleuca lanceolata and Callitris preissii trees growing in 
plantations adjacent to the burnt area were not burnt but 
two isolated Melaleuca trees and one Pittosporum 
ligustrifolium were burnt. One of the Melaleucas was 
included in the quokka-proof fencing subsequently 
erected, and numerous Melaleuca seedlings germinated to 
the west side of this tree after the fire (although not within 
the sampled quadrats). These self-seeded Melaleucas grew 
much more rapidly than the planted Melaleuca seedlings. 
The second burnt Melaleuca tree was not fenced and no 
seedlings were observed growing in its vicinity. The 
Pittosporum tree that was burnt was just outside one of the 
fenced plots. This species is capable of sprouting from 
damaged roots after fire (McArthur 1996a) but in this case 
did not do so. No seedlings were observed outside the 
fence, but 76 young Pittosporum 30-40 cm high were 
growing within the fence to the west of the burnt tree in 
August 2001 (C Hansen, Environmental Officer, Rottnest 
Island, personal communication). Melaleuca lanceolata and 
Callitris preissi planted in the four years following the fire 
were noted in all transects within fenced areas, and 
formed some 7% of cover in those transects (Table 1). 
Acacia rostellifera 
The most striking contrast between the vegetation of 
the three areas was the dense germination and growth of 
the native Acacia rostellifera within enclosures, reaching 
5 m height on sheltered slopes. It had not grown in the 
research site for over 40 years prior to the 1997 fire. 
However, patches of acacia scrub had grown there 
between 1942 and 1955, decreasing in extent until 
eliminated after the 1955 fire (Figs 6A,B). The distribution 
of Acacia rostellifera in 2000 (Fig 6C) corresponded with 
that 60 years earlier, in 1942. Comparison of the 
distribution of acacia after the 1997 fire with a contour 
map showed that the species had germinated in swales 
and on east-facing slopes (Fig 7). 
54 
