466 The intravascular electrode readings correlated extremely well (r-.1) with the arterial blood-samples analysed in vitro but, a comparison of POz on sixty-four paired samples of arter- ial and oxygenator blood gave a poor correlation (r—0- 36). In two additional cases the P02 was measured continuously in the radial artery and the oxygenator with the Roche elec- trode (figure). An analysis of twenty-nine paired points, taken at five-minute intervals from the continuous measurement, again gave a poor correlation (r=0-$l). On several occasions a large difference between the two P02 values was seen. The wide discrepancy between arterial and orxygenator values is probably due to inadequate mixing of blood at the sampling port and the presence of microbubbles of oxygen, which equilibrate with arterial blood after leaving the oxy- genator. The conclusion we reach, therefore, is that in this particular bubble oxytenator, the use of samples taken from the arterial reservoir for assessment of PaOz is misleading. Since similar criticisms may apply to other oxygenators the problem merits further investigation. Intensive Therapy Unit and Departments of Anaesthesia and Cardiology, Middlesex Hospital, London W1N 8AA SHYAM V. S. RITHALIA PATRICK J. BENNETT MARVIN F. STURRIDGE JACK TINKER MYCOPLASMA CONTAMINATION IN CELL CULTURE OF CROHN’S DISEASE MATERIAL SIR,-Whorwell, Phillips et al.’ have detected agents induc- ing a cytopathic effect (CPE) in WI38 (human embryonic lung) cell cultures inoculated with 220 nm filtrates prepared from resected intestinal sections from 6 of 10 patients with Crohn’s disease. Three isolates which grew well were found to be antigenically related to Nebraska calf diarrhoea virus, a rotavirus. At our request Dr C. A. Phillips kindly sent the fol- lowing specimens from (or derived from) this study: 1. Whorwell PJ, Phillips CA, Beeken WL, Little PK, Roessner KD. Isolation of reovirus-like agents from patients with Crohn’s disease. Lancet 1977; i: 1169-71. (a) Cell-culture passages of the three agents (137, 159, 168) charac- terised as rotavirus. (b) A ’Genetron’ treated 220 nm filtered 10% homogenate of resected intestinal tissue from patient 137 which represented the tissue inoculum that yielded putative rotavirus 137. (c) A similarly prepared intestinal homogenate from another patient (171-1) which had also yielded a CPE-inducing agent not identified by the Vermont group. (d) Guineapig antiserum to the putative rotavirus raised with two inoculations of second W138 passage of strain 137 with incomplete Freund’s adjuvant. ’ Dr Phillips also kindly sent: (e) A 14th passage in W138 cells of an explant of ileal tissue from Crohn’s disease patient 146, which had been established as an explant tissue culture and after trypsinisation, inoculated onto W138 cells, inducing a CPE which was described as being identical to changes observed when W138 cells were inoculated with virus-positive tissue from patients with Crohn’s or other inflammatory bowel diseases. The latter ’infected W138 cultures were described as having reovirus-like particles in the cell cytoplasm by electronmicroscopy which were further characterised. (Data on strain 1462 were not in the Lancet article1 cited above; explant strain 146 had been passaged 12 times in W138 cells in the abstract report. [CA Phillips, unpublished]). (f) 14th passage of harvests of W138 cells as control for (e). We examined most intensively strains 137 and 146. Certain of the passage materials of 137 and the passage material of explant 146 induced a definite CPE in W138 and MA7 (human embryonic lung) cell cultures, which could be pas- saged. However, we could not detect the presence of rotavirus in tested passage materials prepared in our laboratory, rou- tinely utilising for rotavirus detection enzyme-linked immuno- sorbent assay, and direct immunofluorescence (IF) (fluores- cein-conjugated rabbit antibovine rotavirus globulin kindly supplied by Dr C. A. Mebus and Mr M. B. Rhodes). These assays are sensitive for rotavirus detection.’ The 137 tissue homogenate and a cell-culture passage of 137 and of 146 received from Vermont were examined by immune electron- 2. Phillips CA, Little PE, Roessner K, Moehring J, Chapple PJ. Virus carrier culture derived from the ileum of a patient with Crohn’s Disease. In Vitro 1977; 13: 158 (abstract). 3. Kapikian AZ, Yolken RH, Greenberg HB, Wyatt RG, Kalica AR, Chanock RM, Kim HW. Gastroenteritis Viruses. In: Lennette EH, Schmidt NJ, eds. Diagnostic Procedures for Viral, Rickettsial and Chlamydial Infections. Washington, D.C.; American Public Health Ass, 1979: (in press). 4. Theil KW, Bohl EH, Agnes AG. Cell culture propagation of porcine rota- virus (reovirus-like agent). Am J Vet Res 1977; 38: 1765-68. Immunofluorescence studies for M. hyorhinis. (A) Vero (African green monkey kidney) cell culture inoculated with strain 137 W 18-2 and examined after incubation with fluorescein-conju- gated reference M. hyorkinis (strain GDL) mule antiserum. Note numerous small fluorescent bodies (M. hyorhinis) m close association with m- fected cells. (B) Control uninoculated Vero cell culture incubated with same fluorescein-conjugated reference M. hyorhinis antiserum as used in (A). Note absence of mycoplasma. (C) Vero cell culture inoculated with strain 137 W138-2 and examined by DNA fluorescence staining procedure. Note numerous small fluorescent bodies in close association with infected cells. (D) Control uninoculated Vero cell culture examined by same DNA fluorescence staining procedure as used in (C). Note absence of small fluores- cent bodies.