Sixty-five samples of blood from patients with rheumatoid arthritis and 94 from control subjects were examined by multiple culture and monthly subculture for periods of up to 33 months to detect the reversion stages of Bacillus licheniformis var. endoparasiticus (BLE), which exists as L-forms in blood. Isolates of BLE were obtained more often from the blood of control subjects than from rheumatoid patients during the first 6 months of incubation, when there was clustering of positive cultures within samples. Thereafter, the isolation rate was similar for the two groups and positive cultures were distributed randomly between samples. Isolation of diphtheroid intermediate reversion stages (phases A and B) occurred mainly during the first year of incubation, but isolations of the fully reverted sporing bacillus (phase C) increased in frequency with incubation time, particularly from cultures with a high rate of desiccation during prolonged incubation. The proportion of different phases of BLE amongst the isolates and the distribution of the phases with incubation time were similar for the rheumatoid-arthritis patients and the normal subjects. 98 ROBERTA BARTLETT AND K. A. BISSET rheumatoid arthritis patients than from those of patients with non-rheumatoid joint conditions. Bartholomew and Nelson (1972) have reported a higher isolation rate of the anaerobic or microaerophilic "Corynebacterium acnes" from the blood of rheumatoid patients than from non-rheumatoids. Previous work in this laboratory (Pease, 1974; Bisset and Bartlett, 1978) has produced evidence that the organisms resembling Mycoplasma and those described as diphtheroid bacteria, isolated from the blood of rheumatoid patients and from normal persons (Benedek, 1955; Pease, 1969,1970; Duthie et al., 1976; Markham and Myers, 1976; Tedeschi et al., 1978), are L-forms and reversion stages of B. licheniformis var. endoparasiticus, hereafter referred to as BLE. MATERIALS AND METHODS Blood samples Collection of samples. A total of 164 blood samples were examined, 65 were from rheuma-toid arthritis patients, five from patients with other arthritic conditions, and 94 from controls. Fifty-five samples of blood from healthy donors were taken by courtesy of the West Midland Regional Blood Transfusion Service; 1 O-ml samples were collected in 1-ml sterile acid-citrate-dextrose anticoagulant at the end of the donation from the tube that remained connected to the donor's vein, by the method and with the precautions previously described (Bisset and Bartlett, 1978). Twenty-one samples of blood from laboratory staff were taken by Dr M. Synnott of the Department of Medical Microbiology, University of Birmingham. The donor's skin was disinfected with 70% isopropyl alcohol (Sterets Injection Swab, Prebbles Medical Ltd, Mersey-side); 5 ml of blood was taken into disposable hypodermic syringes and transferred to sterile screw-capped bottles containing 0.25 ml of 0.4% (w/v) aqueous heparin. The same technique was used for the collection of two samples of blood from laboratory st...