447 - 
ying an 
n are 
if ied 
Table 2. 
A further measure of reliability is the ability to 
detect a source of P. cinnamomi infection. Although errors 
in detecting the exact boundaries of an area of dieback will 
lead to errors in area, it is important to detect the source 
of infection whether or not the boundaries can be accurately 
determined. In the area checked in the field all nine die- 
back patches found were detected on the 1s 3,000 photos, 
while only four of the six patches found in the area covered 
by the 1: 40,000 photos were detected. 
000 
The study has shown that 70 mm 1s 3>000 colour 
photography is a considerable advance on conventional 
1s 40,000 black and white photography for the detection of 
dieback disease in western Australia. Sources of infection 
and the boundaries of incipient dieback are detected with a 
ield 
Pansect 
ne tres) 
high degree of reliability. Advanced dieback is detected 
with a satisfactory degree of reliability. 
On the 1s 40,000 photographs, the ability to 
detect centres of infection and to define boundaries of 
14,000 
; 850 
0 
Type B and associated Type 0 can still be regarded as un 
satisfactory despite the improvements made by the additional 
information from the 70 mm strip lines. So far as new 
infections are concerned, an anomaly exists in the results. 
640 
r 800 
0 
It would appear that the 1s 40,000 photos are as reliable as 
the1s 3>000 photos for this purpose (see Table 1) . However 
this is due to the fact that several infections detected on 
the 70 mm photos were linked along watercourses, as discussed 
earlier. These watercourses were in fact infected but they 
could not have been detected without the aid of the 70 mm 
0 
0 
200 
i 
photographs. Where infections are isolated, Type G cannot 
be detected with any degree of reliability on the 1: 40,000 
photos • 
0 
DISCUSSION 
Although the large-scale phase (70 mm) has proved 
successful, the small-scale phase requires improvement if 
the system is to be fully effective in the southern jarrah 
forest. 
Colour infra-red photographs could be expected to 
provide some improvement in detecting differences in total 
leaf density. Although a trial of small-scale colour 
infra-red photographs has been carried out, difficulties are 
being experienced in obtaining a satisfactory colour balance. 
Further trials with the 70 mm photography are also 
required to determine the smallest scale at which individual 
understorey specimens can be identified. If a smaller scale 
is possible, then the increased coverage should improve 
results by providing a better picture of the pattern of the 
deaths in the understorey. 
Apart from mapping, the project has emphasised