Editor,—We would like to comment on the study by Arya and colleagues¹ in which they failed to find evidence for Mycoplasma hominis being pathogenic in the vagina, or otherwise contributing to bacterial vaginosis (BV). They mentioned the 21 year old review of Taylor-Robinson and McCormack² who surmised that M hominis might act in symbiosis with other organisms or as a sole pathogen in BV. The latter was referred to as non-specific vaginitis or Gardnerella associated vaginitis at that time, the term BV being used from about 1984. Since then, much has been learned about the vaginal microflora in health and disease, but the question of which bacteria, if indeed any, cause BV remains unanswered. The few M hominis organisms in the healthy vagina appear to behave as commensals. We challenged³ the suggestion of Mårdh and colleagues that M hominis was associated with a number of genital signs and symptoms after BV had been excluded, our assertion being that M hominis organisms outside the context of BV would be present in small numbers and, therefore, unlikely to cause a problem. In contrast, the few M hominis organisms in the healthy vagina increase in number, perhaps by 10 000-fold or more, in the vagina of women with BV. This increase, however, occurs only late in the development of BV.⁴ Indeed, it is rare to find large numbers in the “intermediate” (grade 2) stage between the normal vaginal flora and “full blown” BV (grade 3). Thus, in the study by Arya and colleagues we have difficulty in understanding why only 35 (48%) of the 73 women with M hominis positive BV had large numbers of organisms (>5 × 10⁵). A Gram stain evaluation should have distinguished women with grade 2 flora from those with grade 3. Be this as it may, the authors contend that because the additional presence of M hominis with G vaginalis and strict anaerobes did not seem to increase the likelihood of the women developing BV, M hominis is not involved. It is clear that M hominis organisms are not essential for the development of BV and unlikely that their initial presence in the vagina increases the likelihood of BV developing. However, if they are present in the vagina initially, then they will multiply as indicated and large numbers will ensue. The data of Arya and colleagues do not resolve the issue of whether large numbers contribute to the disease process or are involved in its persistence. Against this, as they point out, is a study⁵ in which metronidazole, inactive in vitro against M hominis, cleared vaginitis, and doxycycline, active against M hominis, did not. However, it should also be remembered that M hominis organisms caused pharyngitis and cervical lymphadenopathy when given orally in large numbers to volunteers,⁶ indicating the pathogenic potential of the organisms. Furthermore, the M hominis species is heterogeneous, some strains having greater epithelial cell adherence properties than others. We do not see any data that point to M hominis being a sole pathogen or co-pathogen in the vagina but, equally, we are not convinced by data that purport to show that it is not.