Systematic surveys were used to determine the isolation frequency of Phytophthora cinnamomi and other Phytophthora spp. from the centre and margin of infested sites (dieback sites) in native vegetation in the south-west of Western Australia. Soil and fine root samples were baited with Eucalyptus sieberi cotyledons, dried for 1 week and then rebaited (double baited), in order to maximise recoveries. Any Phytophthora spp. isolated were identified. P. cinnamomi was recovered from 1.1 and 7.2% of soil samples taken from the centre and margin, respectively, of dieback sites. Double baiting increased the recovery of Phytophthora spp. from 1.9 to 2.5% from the centre, and P. cinnamomi from 6.3 to 7.2% from the margin of dieback sites. The proportion of samples from which Phytophthora was not recovered was used to estimate the number of samples needed from a suspected dieback site which, if all were negative, would indicate that Phytophthora was not present. For P. cinnamomi, with a probability of 0.05, the sample size is 40, taken from within 5 m of a vegetation boundary, whereas the sample size is 271 for samples from the centre of a suspected dieback site.