solution and insist that the problem be solved by the general managers rather than through the agreed structure.This, however, would have undermined the principles on which the structure had been based.Four major difficulties still remain: * The demands on the time ofthe chiefexecutive * The demands on the time of the clinical directors and their specialty colleagues * The perceived interruption of the career paths of the general managers * The potential tensions arising from functional directors, such as finance, delegating responsibility to clinicians and not traditional managers.The organisation is now far better placed to under- take major projects and initiatives than it was before: decisions are multidimensional and largely consensual, and managers and clinicians work in harmony together.The clinical directors are not disenfranchised but are an essential part of the decision making process in the trust.They are readily accepted by their consultant colleagues as acting in the clinical interests of patients.We believe that the strengths of our clinical director- ate model of management, with its direct access to the chief executive, lead to realistic and robust decisions which minimise disruption to the services, make for faster decision making, and maximise benefits to patient care.There are several lessons to be learnt from the experiences outlined above as they affected Frenchay Healthcare Trust.We believe that this system of management has allowed decisiols concerning the future of the services Key point summary * Strategic direction needs to be established early, determining what is to be provided and how this is to be achieved * Flexibility is important: the identification of strengths and weaknesses, and threats and opportunities for the directorate needs to be part ofthe business plan * A clinical directorate model of management, with direct access to the chief executive, mini- mises disruption to service provision, while maximising patient care * The effectiveness of a clinical directorate depends on being given real responsibility and on good support systems provided and their current performance and potential to be taken by the clinical staff and managers working together to achieve a common goal.It has allowed a synergy to prosper between the aims and objectives of these two groups of staff that is changing the culture of the organisation.It has focused on the importance of decisions being taken by staff with a shared perspective, particularly in times of change.

This lecture commemorates Arthur Tudor Edwards, one of the very small group of surgeons who, in the decade following the first world war, established the surgery of intrathoracic organs as a practicable procedure.For one year, 1932-3, I was house officer at the Brompton Hospital to him and his colleague, J. E. H. Roberts, the two leading thoracic surgeons in London.Seeking a possible link between the time when as a young man I was privileged to work for Tudor Edwards and my present interest in the nosology of broncho-pulinonary disease, I propose in this lecture to consider the contribution of surgical biopsy to the diagnosis of widespread lung disease.Extrinsic allergic alveolitis ... 1 Sarcoidosis ...15 Eosinophilic granuloma (histiocytosis X) 5 Malignant disease..4Prinmary emphysema... 1 Categorization doubtful... 8 Alveolar fibrosis (? extrinsic allergic, ?cryptogenic) 3 Necrobiotic nodules .. .4 No evidence of rheumatoid arthritis 3 Later developed rheumatoid arthritis 1 Honeycomb lung, unclassifiable .. . 1 Biopsy not contributory. . . 5 Final diagnosis (after further observation): Probable cryptogenic fibrosing alveolitis (1 with coal miner's pneumoconiosis) . . 3Gas transfer defect of undetermined cause: death from

Philip Ellman, whose work this lecture is intended to commemorate, died just over eight years ago. He was above all a clinician, interested in the problems of individual patients. His contributions to the science of medicine were those of an astute clinical observer, endowed with the gift of describing his observations clearly, both in speech and in writing. His principal interests were in two fields: the chronic respiratory diseases, including tuberculosis, and the rheumatic diseases. He participated actively in the meetings of this Section and of the Clinical Section of this Society. It was at a joint meeting of this Section and the Section of Physical Medicine in 1947 that he referred for the first time to the possibility that chronic lung changes might be associated in some specific way with rheumatoid arthritis,'mentioning briefly a case of rheumatoid arthritis with splenomegaly, hepatomegaly, lymphadenopathy and leukopenia, and widespread radiographic changes in the lungs which at necropsy were found to be due to 'a curious chronic fibrosing broncho-pneumonic lesion' (Ellman 1947). Subsequently, he published further clinical and pathological contributions to knowledge at this meeting-point of his two main interests. There is now a very large literature on the associations between rheumatoid arthritis and changes in the lung and pleura; and many of the authors who have written on this difficult subject have started by referring to these early observations of Philip Ellman's. It therefore seems appropriate to devote this lecture to this subject. I propose to consider, first, the evidence which suggests that certain types of morbid change in the lungs and pleura may be related in some way to rheumatoid arthritis; and, second, the possible nature of any such relationship.

treatment with intramuscular penicillin and Cibalgin suppositories produced relief. The sore throat recurred in June 1967 and she was given Cibalgin suppositories but became febrile. Penicillin was given once more together with further Cibalgin suppositories. The fever persisted and painful ulceration of the anal margin was noticed. A white cell count revealed no granulocytes and 500 lymphocytes per mm3. She was admitted to the H6pital Cantonal de Lausanne and was treated with prednisone 50 mg and ampicillin 8 g daily. The white cell count gradually rose and the granulocyte count improved. With the increase in granulocyte count a phlegmonous pararectal mass developed which fistulated into the rectum, destroying the anal canal, sphincters and lower rectum. On 14.7.67 examination of the perineum revealed a large cavity with necrotic skin over the buttocks, the rectum hanging freely in the sacral concavity. Debridement was performed and a left iliac sigmoid colostomy was made. The cavity healed slowly and the buttock area became covered with granulation tissue. She was transferred to Westminster Hospital on 10.8.67, when she was afebrile, pale and wasted. The perineal cavity was examined on 11.8.67 and the rectal stump was visible at the vault of the cavity. On 18.8.67 excision of the rectal remnant was performed leaving the existing colostomy in the left lower quadrant. The patient progressed well and on 31.8.67 Thiersch grafts were applied to the raw granulating areas on the buttock. She subsequently progressed well, the perineum being almost healed by the time of her discharge from hospital on 6.10.67.

Intrathoracic changes enlargement of mediastinal lymph nodes, with or without superior vena caval (SVC) obstruction, infiltration of the lungs or rarely of the bronchial wall, or pleural effusion may be the presenting manifestation, without evident extrathoracic changes, and in such cases the problem of diagnosis is that of obtaining material for biopsy by the safest and most convenient method.

In spite of the great interest which has been shown in sarcoidosis in the past 20 years, leading to the recognition of increasing numbers of cases in many parts of the world, controversy concerning its aetiology continues. The validity of a contribution to this controversy depends in large measure upon the definition of sarcoidosis adopted by the person advancing it; for sarcoidosis, lik-e any other disease whose aetiology is not generally agreed, must be so defined that any theory whatever about its aetiology can be discussed intelligibly and with logical propriety (Scadding, 1950, 1956). Any definition which includes the statement, " Sarcoidosis is a disease of unknown aetiology"-for example, that quoted by Ricker and Clark (1949)-is self-stultifying in this respect. An investigator who adheres to such a definition must exclude from the category " sarcoidosis " any case in which a causative agent can be identified, and thus his definition closes his mind to some tenable views concerning aetiology. Important among these is the hypothesis that sarcoidosis may be a reaction to an agent or agents already known, but difficult to demonstrate in this particular manifestation. For if in a given case a probable causative agent has been demonstrated, the definition demands either that the case be excluded from the category " sarcoidosis " or that the presence of this agent be ignored as incidental and unrelated to sarcoidosis. Clearly we must adopt a definition which imposes no a priori limitation on our investigation of aetiology.