610 From the work already completed it seems clear that the general structure of ribonuclease shows no startling differences from those of myoglobin or lysozyme. In all, the polypeptide chain forms a loose ball with hydrophobic groups inside and hydrophilic groups outside-a type of structure becoming familiar in globular proteins. There are indications in both lysozyme and ribonuclease that the regions to which substrates become attached lie in clefts. We are still some way from understanding how this allows enzymic activity to take place, but precise analysis of the spatial relationships of enzyme and substrate is a useful step towards understanding the chemistry underlying the reactions.
CLUSTERING
PSYCHIATRIC epidemiologists would probably agree that, from their viewpoint, there are no entirely satisfactory social classifications; social patterns of behavioural difference, in particular, are only very coarsely reflected in existing social categories. Epidemiologists of the chronic diseases would be likely to agree with the psychiatrists; the extent to which social factors are involved in the chronic diseases is day by day becoming more evident, and existing methods of measuring them are proving less and less adequate. Thus, a technique which might find new paths through this thorny jungle and which, moreover, demands onlypencil and paper is especially attractive. Philip and McCulloch,1 of the M.R.C. unit for research on the epidemiology of psychiatric illness, have described such a technique, the purpose of which is to isolate " clusters " of interrelated factors from large masses of information. In a series of studies in Edinburgh, they needed to compare social characteristics of the electoral wards of the city. They collected eighteen social variables derived from various statutory and voluntary sources (a few, arbitrarily chosen from the eighteen by way of example, are: adolescent psychiatric
referrals, juvenile delinquency, .eviction notices, overcrowding, infant mortality, old-age pensioners). A matrix of correlations between the variables was calculated. Faced with the need to reduce the matrix into a meaningful pattern, Philip and McCulloch applied to it the
technique of elementary linkage analysis, originally developed by L. K. McQuitty. Elementary linkage analysis elicits clusters of closely interrelated items from the initial matrix of correlations, and it does so by a simple method which produces immediately intelligible results; one is spared more elaborate and less direct techniques such as factor analysis. The authors give a brief 5-step series of instructions, with the aid of which the required results were reached within a few minutes. What are the advantages of such a technique ? Anything which reliably makes pathways through the often daunting mazes of correlation matrices is welcome. In this instance the data used were readily available, replicable, and objective, and yet the clusters which were derived from them suggested relationships which might easily have been overlooked: thus, one cluster was concerned with self-poisoning, a subject much studied by the M.R.C. unit, and yet the cluster indicated relationships between self-poisoning rates and three other variablescases coming to the notice of the Royal Scottish Society for the Prevention of Cruelty to Children, children in care, and school absences-which had not been noticed during 1.
Philip, A. E., McCulloch, J. W.
Br. J. prev.
soc.
Med. 1966, 20, 122.
the
previous studies. Such relationships may suggest hypotheses which can be tested, and may lead to the understanding of previously unidentified processes. At the very least, by enabling a representative of a cluster to be used as an index for the whole cluster, the technique can save laborious examination of inessential data.
HEPATECTOMY FOR CANCER OF THE LIVER
SURGICAL removal of benign tumours of the liver (even when this means removing a large part of the liver itself) is not only feasible but also, as the records show, widely practised. The liver has remarkable regenerative capacity. After removing about 90% of the liver in a patient with multiple adenomas, Monaco et aLl found that function had returned to normal within six months. Others have reported full functional recovery where only a fifth of the liver had been left after operations. Emboldened by this kind of result, surgeons have turned to partial or neartotal hepatectomy in primary or secondary malignant growths of the liver. Here, the outcome has been disappointing. Reports of long-term survivors are scanty: only a few patients with primary growths are on record as having survived for some years. In the case of metastatic growths the tally is even more dismal. Nevertheless, Friesen et al. believe that newer methods for early diagnosis, careful selection of patients, and refinements of technique could save at least a significant minority of patients. This, they hold, could be brought about by periodic laparotomies before there are symptoms or signs of recurrence, at a time when small nodules of growth are still amenable to further resection. This is the " second look "-a procedure which could be more aptly termed a series of looks ". One of their patients had the left lobe of the liver, part of the diaphragm, and the spleen excised. After eight subsequent operations over a period of five years, entailing further removal of portions of the liver and segmental lobectomy of a lung, the patient remains " well". A further 5 patients have survived re-explorations for periods ranging from eight years to eighteen months. Earlier diagnosis is essential for success. The smaller the tumour the less is the need for massive resection. Friesen et al. regard scintillographic scan of the liver as a great advance in defining the size of the tumour. Portography may disclose obstruction of the portal veina late sign. Exploratory laparotomy is preferable to percutaneous biopsy, because needling of the liver may favour spread. Induced hypothermia during hepatectomy is helpful in the control of bleeding. Combined with clamping of the blood-vessels in the porta hepatis and ligature of the parenchymatous vessels after their exposure by digital dissection,4hypothermia adds greatly to the safety of the operation. The portal circulation, it seems, can be temporarily occluded for considerable periods without damage to the liver. How far this bold attack on malignant disease (admittedly with an otherwise hopeless prospect) can succeed is debatable. Certainly, it demands courage and endurance from the patient as well as the surgeon. "
Monaco, A. P., Hallgrimsson, J., McDermott, W. V., Jr. Ann. Surg. 1964, 159, 513. 2. Pack, G. T., Islami, A. H., Hubbard, J. C., Brasfield, R. D. Surgery, St. Louis, 1962, 52, 617. 3. Friesen, S. R., Hardin, C. A., Kittle, C. F. ibid. 1967, 61, 203. 4. Tung, T. T., Quang, N. D. Lancet, 1963, i, 192. 1.