Cacti (Opuntia spp.) are perennial, evergreen, succulent plants native to arid areas of the Americas. Because of their aesthetic appearance, many cacti have been cultivated and introduced to other parts of the world as ornamentals. Cacti are susceptible to phytoplasma infections and develop witches'-broom (WB) disease. Currently, all reported cactus WB cases are associated with infections by phytoplasmas in the peanut witches'-broom group (16SrII) (1,2,4). During a phytoplasma diversity survey carried out during 2004 in Yunnan, China, we collected 29 malformed and 14 healthy-looking naturally occurring cactus plants from 14 locations representing five geographical regions. Each of the 29 malformed plants exhibited stunted growth and possessed clusters of highly proliferating cladodia, typical symptoms of cactus WB disease. Nested-PCR was carried out on the DNA samples extracted from young cladodia of these plants using phytoplasma-universal 16S rDNA primers P1A/P7A and R16F2n/R16R2 (3). Results revealed that all 29 diseased plants that were examined were infected by phytoplasmas, whereas all 14 healthy-looking plants were negative for phytoplasmas. Subsequent restriction fragment length polymorphism (RFLP) analysis of the PCR-amplified 1.25-kb 16S rDNA fragments indicated that 28 diseased plants were infected by a phytoplasma of group 16SrII, whereas one plant (from Suan Village) was infected by a 'Candidatus Phytoplasma asteris'-related (group 16SrI) phytoplasma designated as strain YN26. Nucleotide sequence analysis of the strain YN26 partial rRNA operon (GenBank Accession No. EF190970), covering a near full-length 16S rRNA gene, a 16S-23S rRNA intergenic spacer, a tRNA-Ile gene, and a partial 23S rRNA gene, suggested that this phytoplasma is most closely related to an ash witches'-broom phytoplasma (GenBank Accession No. AY566302, 99.7% identity) and an epilobium phyllody phytoplasma (GenBank Accession No. AY101386, 99.7% identity), both members of subgroup16SrI-B. This YN26-infected cactus plant was transferred to a greenhouse and maintained for more than 2 years, during which time DNA samples were extracted and tested two additional times. The same 16S rDNA RFLP pattern type was consistently obtained in these tests, confirming that the plant remained infected by the 16SrI phytoplasma. To our knowledge, this is the first report of a natural infection of a cactus species by a group 16SrI phytoplasma. Since this 16SrI-cactus WB phytoplasma was found in the same geographical location where 16SrII-cactus WB phytoplasma was detected both in this and a previous study (1), the findings raised the question whether 16SrI- and 16SrII-cactus WB phytoplasmas have overlapping geo- and bioecological niches.