Sweet cherry (Prunus auium) and sour cherry (Prunus cerasus) trees from orchards in the Kootenay and Okanagan Valleys of British Columbia, Canada were assayed for the presence of little cherry disease by three different methods: Northern blot analysis of double-stranded RNA, woody indexing for fruit symptoms on sweet cherry cv. Lambert, and woody indexing for foliar symptoms on cv. Canindex 1. Results of the three methods were in agreement for 85% of the samples. Of the 78 orchard trees tested, double-stranded RNA isolated from 48 trees hybridised with a radiolabelled cloned probe specific for little cherry disease. When the 48 trees were tested by woody indexing, buds from 41 trees induced fruit symptoms on cv. Lambert, but only 32 yielded foliar symptoms on cv. Canindex 1 under the conditions of the experiment. Of the 30 orchard trees that did not yield a positive response to the Northern blot analysis, 26 samples were negative on cv. Lambert and 26 were negative on cv. Canindex 1. Northern blot analysis of the 78 cv. Lambert indicator trees revealed that there was an absolute correlation between the presence of little cherry disease-associated double-stranded R N A and the development of typical little cherry disease symptoms on the indicator trees. Reliability of woody indexing of orchard samples was impaired by poor transmission of the disease from the inoculating bud to the indicator tree. Woody indexing with cv. Canindex 1 was particularly prone to a large number of apparently erroneous negative results. Of the three protocols used, diagnosis of little cherry disease by Northern blot analysis was found to be the most reliable and offered a greatly accelerated means of diagnosis.