The suggested relationship of Klebsiella species to the pathogenesis of ankylosing spondylitis reflects evidence that there was an increase in fecal Klebsiella carriage in patients with active AS when compared to controls, that B27-positive lymphocytes from AS patients could be distinguished from normal B27-positive lymphocytes by an antiserum, and that a nitrogenase enzyme found in some species of Klebsiella had a sequence of six amino acids identical to a sequence seen in B*2705. It is the authors' view that the superficial similarities between these observations has been the chief factor leading to their support but that on close observation none are attractive either on pragmatic or on intuitive grounds.