same association, for in it the higher the percentage of takes the lower is the interval within which the mice die. We are, therefore, at a loss to understand how Calkins comes to a different conclusion, and do not find either in our own experience or in the data adduced by him, sufficient evidence to warrant the assumption that the fluctuations in percentage of successes, are due to recurring cyclical changes in an associated intracellular parasite. To designate percentages of successes infectivity, is merely to give another name to the facts, and to obscure their significance by introducing a doubtful analogy. While it is inadvisable at present to speculate on the nature of the mechanism of these fluctuations and their relation to the metabolism of cancer, their ubiquity in practically every strain we have studied, and in most of those studied by others, points in our opinion to their importance. After making every allowance for the effects of technical details on the facies of the curves, the general result seems to be elicited that the amplitude and frequency of the fluctuations are greatest in those tumour-strains which grow most rapidly. It is extremely difficult at present to devise experiments, which will enable us to penetrate with confidence more closely into the nature of the processes producing these effects, but the opinion may be hazarded that when we are able to do so, energy of growth will be found to depend on the rapidity with which these phases succeed each other in the life of the cells, and therewith an important indication will be obtained of the nature of the cellular transformation which takes place in the initiation of cancer, and how it maintains its apparently continuous (vegetative) growth. [Reprinted by permission of the Council from the PROCEEDINGS OF THE ROYAL SOCIETY, B. Vol. 78, 1906.] THE EXPERIMENTAL ANALYSIS OF THE GROWTH OF CANCER. By E. F. BASHFORD, M.D., J. A. MURRAY, M.B., B.Sc., [Communicated by Professor J. ROSE BRADFORD, F.R.S. Received May 30; Read June 14, 1906.] IN the present paper we shall attempt to analyse the growth of cancer when propagated artificially in mice, mainly on the basis of 25,000 inoculations of Jensen's tumour performed in conjunction with Dr. W. Cramer on behalf of the Imperial Cancer Research Fund; but also with reference to inoculations made with 32 other mouse tumours during the past three years. Although the question of the continuous or interrupted nature of cancerous proliferation is of fundamental importance, both from the standpoint of the ultimate explanation of the nature of the disease, and from the standpoint of its treatment, such an analysis has never been attempted before. It has been assumed that the growth of cancer is vegetative, as inexplicable as any other form of growth, only to be partially understood by an elucidation of the processes by which normal tissues become cancerous. Artificial propagation enabled us to submit this assumption to the test of experiment, and imposed the necessity of determining by direct observation whether propagated cancer exhibited a mode of growth throwing light on the nature of the disease and the apparently continuous proliferation of sporadic tumours. While the experimental propagation of cancer may reveal new facts with a bearing on the nature of the disease it also affords opportunities for rational and empirical therapeutic experiments, and adequate opportunity for controlling the results. These two purposes have been constantly kept in view in our investigations. When a number of animals are inoculated with a transplantable mouse tumour, all do not develop tumours and the tumours which do develop are not all of the same size after the same interval. In order that propagated cancer might be available for the second of these purposes it was necessary to find out what influence the variable conditions of experiment exerted on the proliferation of the cells. In the course of these preliminary studies facts bearing on the nature of cancer have also been ascertained. Irregularities in the rate and amount of growth are introduced by (1) Transference from one race of mouse to another even when nearly allied; (2) Transference from young to old mice of the same race or vice versá; (3) Variations in the site of implantation of the cancerous. tissue; (4) Variations in the amount of the tissue implanted and in the manner of introducing it ; (5) Variations in the character of the tumour cells themselves. Any one of these factors may cause a very great deviation from the rate and amount of growth observed previous to the subinoculations introducing it, and invalidate the results of experiments of which information as to possible modification of growth was the object. The variations depending on the first four factors mentioned must be eliminated before variations can be referred to the tumour cells themselves. We have taken the following precautions in studying the fluctuations which we believe depend on differences in the tumour cells: 1. The same race of mice has been used throughout. We have observed differences in the suitability of animals of different colours even among the ordinary English tame mice; and we have avoided the use of those varieties prized by mouse-fanciers. The wild mouse probably offers more uniform conditions than the tame mouse, but a sufficient stock of uniform age is difficult to obtain, keep and supervise. Jensen's tumour rarely yields a number of successful subinoculations in wild mice equal to that obtained in a control batch of tame mice, and this result when once obtained has not been maintained, but is followed by an increasing difficulty of propagation. The experiments in wild mice may be looked on as control observations to those recorded in tame mice. 2. The tame mice used have been of uniform age, and from five to seven weeks old. We showed that young animals provide conditions more favourable for the artificial propagation of cancerous tissue than old animals. This conclusion has been amply confirmed by our later experience, and in one of its aspects also by the work of Ehrlich and Apolant*, who state that the age of the animals is without importance and, especially, that old females are not more suitable than young animals for the propagation of mammary tumours. We have found that the greater suitability of young animals is even more marked than we at first suspected. The inoculation of a tumour into young and old animals respectively may occasionally give similar results in the two cases, or even a less favourable result in young mice, still such results are exceptional in our experience. As a rule, a much higher percentage of tumours develops in young animals, and they attain large dimensions in a shorter time after inoculation. The tumours which have developed most rapidly, e. g., attaining a weight of 1:05 grammes in a mouse of 9 grammes, within five days, and those ultimately attaining the largest dimensions as compared with the size of the host, have always occurred in young mice, although tumours of 7 or 8 grammes also develop rapidly in adult animals. Slow growing tumours, which remain of relatively small dimensions, occur both in old and in young animals. The extent to which the youth of the animals usually favours the continuation of growth after transplantation may be illustrated by the results of 18 series of inoculations, in which portions of the same parent tumours were transplanted simultaneously into young and adult animals respectively; 214 implantations into adult animals three to six months old yielded 62 tumours, or 29 per cent. were successful; 363 implantations into young animals five to seven weeks old gave 172 tumours, or 47 per cent. were successful. This result is by no means an extreme case, either as regards proportion of successes or as regards difference in age of the inoculated animals. 3. When the precautions above indicated are observed, the individual variations in the general suitability of different mice of the same race and age are negligible if implantation be performed in the same site, provided sufficiently large numbers are used. We have preferred the subcutaneous tissue of the back. The attempt to perform collateral series of intra-peritoneal inoculations was abandoned, owing to the frequency with which growth within the peritoneum had occurred secondarily by extension from tissue implanted in the abdominal muscles. 4. We have endeavoured to transplant pieces of healthy-looking tissue of uniform size by means of hypodermic needles, and have obtained more satisfactory results by this method than by breaking tumours down into an emulsion and injecting larger quantities of tissue suspended in physiological salt solution. With certain reservations the rate of development and the size the daughter tumours will attain within 10 days is directly proportionate to the amount of healthy tumour tissue implanted; 0·02 to 0.03 gramme of tissue usually gives larger tumours within a given time interval than 0·005 to 0·01 gramme. 5. When the conditions referred to in the four preceding paragraphs are maintained uniform, fluctuations independent of them appear, and we shall endeavour to show that they are, in all probability, natural features of proliferation. The detailed study of these fluctuations has been undertaken with the tumour which has proved readily capable of transmission during the longest period yet attained, viz., that of Jensen. This tumour has now been propagated for four and a-half years, without permanent alteration in its histological characters or its behaviour. We have obtained success in from 5 to 90 per cent., occasionally even in 100 per cent., of the animals inoculated, the percentages being based on data obtained from those mice which were still alive* 10 days after the inoculations were made. The amount of tissue transplanted in each animal varied between 001 and 0.02 gramme †. The pieces were selected from the whole tumour, and hence their behaviour furnishes an estimate of the proliferate energy of its component parts. The use of a restricted number of random fragments is rendered necessary, because it is impossible to transplant the whole of every tumour; the number of animals required of itself limits the investigation. The method of experimental propagation by implanting minute cellular grafts leads to a progressive subdivision of the parenchyma, and to the distribution over a large number of animals of the descendants of cells previously associated together in one animal. The experimental tumours consist of a parenchyma arranged in alveoli. The study of the early stages after transplantation shows that, at first, single alveoli constitute separate centres of growth, and we may therefore term them the "parent alveoli " of the tumour. Since the cells of different alveoli do not intermingle, the progeny of the discrete growing centres in the transplanted tissue remain separate, and are further separated from one another as the parent alveoli increase in size and bud off daughter It is our practice to kill from time to time a number of mice during the first 10 days after inoculation, in order to examine the site of implantation. Tumours of transplantable size, 0·75 to 1·5 grammes weight, are rare before 8 to 10 days. As the object of these experiments was to estimate the power of continued growth as distinct from mere transitory proliferation, some such time limit was necessary. Estimates of the percentage of success and of the frequency of the spontaneous cessation of growth in tumours which had established themselves must exclude transitory proliferation of the cells introduced and inflammatory swellings at the site of inoculation. The weight of the fragments inoculated into each animal has been arrived at by weighing the mass of tumour used for transplantation and dividing by the number of animals used, e. g., 1 gramme of tumour transplanted into 100 mice gives 0-01 gramme per implantation. |