Rippey & Hobbs: Effects of fire and quokkas on Rottnest Island vegetation 
o 
o 
m 
c 
0) 
O 
U-. 
<D 
O 
CN 
O . 
burnt 
fenced 
burnt 
unfenced 
unburnt 
unfenced 
Figure 8. % cover by the introduced Trachyandra divaricata in the 
three areas (burnt fenced, burnt unfenced and unburned 
un fenced). 
deposited some 60 years previously. Germination 
triggered by the 1997 fire in areas outside enclosures 
presumably succumbed to grazing. 
The relationship between the distribution of Acacia 
within the enclosures and topography, with growth 
restricted to swales and east-facing slopes could be 
associated with the fact that the senescent parent Acacias 
of the 1940s and 1950s are likely to have survived longest 
in favourable sites. 
Acanthocarpus heath 
This was the most widespread vegetation community 
throughout the burnt area at the time of the survey, 
dominating except in areas of Acacia rostellifera. 
Fire and marsupial grazing/browsing 
The basic model of Noble & Slatyer (1980) illustrates 
the development of the vegetation communities of 
Rottnest Island after a single disturbance event. However 
if the vegetation dynamics are considered in terms of 
two major factors, fire and marsupial grazing, then a 
second model of alternative stable states becomes useful. 
Charcoal deposits indicate that infrequent fires have 
occurred on Rottnest Island over the 6500 years since 
separation from the mainland (Backhouse 1993; Groves 
2001). Presumably these would have been initiated by 
lightning, concurring with summer and drought years. 
Fires occurred about every 200 years in the vicinity of 
Barker Swamp alone (Groves 2001). Hopkins et al. (1985) 
suggested that infrequent major conflagrations take place 
on islands where there is no Aboriginal presence. Fires 
are known to be started by lightning on islands in the 
Recherche Archipelago (Weston 1985; A Hopkins, 
Department of Conservation & Land Management, 
Woodvale Research Centre, personal communication; M 
Fitzgerald, Department of Conservation & Land 
Management, Esperance, personal communication). Fires 
caused by lightning strikes are not uncommon on the 
Perth coastal plain, and it is likely that these could have 
occurred more than once per century on the islands 
(R Smith, Wildfire Prevention, Fire and Emergency 
Services Authority of Western Australia, personal 
communication). 
These fires could have been fierce and widespread in 
woodland areas, as a considerable amount of flammable 
material presumably had accumulated. Total available 
fuel in mature Melaleuca forest has been estimated at 
15-25 tonnes ha 1 , and in Acacia rostellifera scrub 30 t ha 1 
(McArthur 1996a). This is a heavy fuel load in view of 
the fact that the maximum loading that can be handled 
by a bush fire fighting brigade on a normal summer day 
is 8 t ha' 1 (M Cronstedt, Fire and Emergency Services 
Authority, personal communication). 
There is also doubt as to whether communities of 
Callitris/Melaleuca or Acacia rostellifera could have 
persisted indefinitely in the absence of fire. McArthur 
(1996b) suggested that Callitris and Melaleuca very rarely 
regenerate other than in the vicinity of burnt parent trees. 
In the Research Site, germination of Melaleuca was 
restricted to the vicinity of a burnt Melaleuca within a 
fenced plot. Seed-bearing cones of Callitris preissii were 
scattered in one fenced plot but failed to produce any 
seedlings (C Hansen, Environmental Officer, Rottnest 
Island Authority, personal communication). However, 
Callitris and Melaleuca can sometimes reproduce without 
fire; in 2000 they germinated on Rottnest Island at Oliver 
Hill, where seed-bearing branches of the two species had 
been laid down to stabilize a small blow-out. Some 
Callitris preissii also germinated in the absence of fire on 
the mainland coast at Woodman Point (Powell & 
Emberson 1981). 
McArthur (1998) reported that Acacia may regenerate 
without fire, mainly from root suckers, and can even be 
rejuvenated if seasonal conditions and the insect load 
permit. Acacia rostellifera can regenerate from rootstock 
after damage to the above ground parts along footpaths 
or after destruction by nesting cormorants, but there is 
not the massive germination and regeneration that 
follows fire. Fire appears to trigger germination and 
possibly reduces insect numbers. 
Melaleuca , Callitris and Acacia are palatable to quokkas, 
and in the presence of a large quokka population 
regeneration rarely occurs (Storr 1962). Planted 
specimens do not survive unless fenced. Palatable species 
can become established however after a widespread fire 
as macropod grazing can be reduced for a time (Whelan 
& Main 1979). Baird (1958) found virtually no signs of 
grazing for two years along the southern margin of the 
huge 1956 fire on Garden Island. After the 1955 fire on 
Rottnest Island, some sites were almost completely 
deserted by quokkas (Storr 1963). This may be because 
resident quokkas are killed by fire (McArthur 1996a) or 
because there is no shelter for too great a distance (Storr 
1963). The influence of grazing by macropods on the 
survival of seedlings is not fully understood. The 
occurrence of highly palatable plant species in an area or 
strong territorial instincts on the part of the animals can 
affect the ranges of grazing animals after fire (Whelan & 
Main 1979). 
Acanthocarpus heath is self-maintaining under a regime 
of heavy grazing, although at a suppressed level. In those 
57 
