cell, or else the cancer cell itself becomes robbed of its power of inciting a specific reaction.

A consideration of the specific nature of the stroma reaction to various tumours, which can only be determined by the parenchyma cells themselves, as well as of the specific nature of the phenomena of immunity, as recorded on other pages of this Report, leads one to regard the immunity as being directed against the cancer cell itself, the most probable explanation being, at the present time, that it is directed against the chemiotactic influences exerted by cancer cells on the connective-tissues of the host, as suggested by Bashford, Murray and Cramer, in the preceding paper, p. 340.

Professor Starling's experiments upon the development of the mammæ of virgin rabbits, produced by the inoculation of an extract of embryo-emulsion, have also an indirect bearing upon this question. These inoculations stimulated the secreting cells of the mamma, and led to an increase in the bulk of that organ, which could only take place along with an increase in the vascular supply. Tumour 27 is so typically an adeno-carcinoma of the mamma as to resemble closely at times the structure of the normal lactating organ, and yet it has been found that the previous inoculation of embryo-emulsion prevents the provision of a vascular supply for any subsequent inoculation of this tumour. This may be the expression of a fundamental difference between the normal cells of the mamma, and those of an adeno-carcinoma of the same organ, although the fact that in Starling's experiments we are dealing with an organ already laid down cannot be neglected.

Ehrlich found that a hæmorrhagic chondroma, when inoculated subcutaneously, grew with a white instead of a blue colour under two conditions, either where he used immune mice, or where the vitality of the tumour-cells had been previously injured by heat. Gierke has recorded on p. 115 how hæmorrhagic mammary tumours alternate naturally between hæmorrhagic and non-hæmorrhagic phases, and we have not been able to influence the angioblastic chemiotaxis for them in the way described by Ehrlich for his chondroma, any more than Murray has keratinisation in tumour 32.

It therefore seems justifiable to ascribe the absence of "the specific stroma reaction" in immune animals, to the suppression of the chemiotactic properties of the cancer cells. The suppression of this chemiotactic power robs them at the same time of their powers of assimilation and growth, so that they lie inert when inoculated subcutaneously, and die.

There must be present in the resistant animals, either in the circulating fluids or in the tissues, something which inhibits this chemiotaxis. Owing to the absence of any evidence which would justify an analogy with the antitoxines or anti-bodies to infective organisms and their products, we refrain from the use of these terms. The resistance induced to cancer is in nature sui generis and requires to be studied as such. All attempts made by other workers to demonstrate the presence of an active cell poison in vitro have been inconclusive as yet. The extensive investigations made on this subject in the Imperial Cancer Research Laboratory have been entirely negative. It is not impossible that the positive results claimed by others are fallacious. That the induced resistance to inoculation is not due to a very active cell poison may be seen from the power which the cancer cells retain, of continuing their proliferation for 7-10 days in this unfavourable medium, provided that they can obtain sufficient nourishment. Further, the cells at the periphery of the graft are those which go on growing, and these are the very cells which are most exposed to the influence of any supposed poison, whereas the cells towards the centre of the graft, which are not so exposed to a free supply of this inimical substance, are the ones which die rapidly because of the interference with their food-supply.

The action of the cyto-toxins and cytolysins, which are produced by the inoculation of epithelial tissues into a strange species, as was shown by von Dungern, is much more powerful than this, leading as it does to a rapid disintegration of the cells tested. The results obtained in the early stages in immune rats as described above, help us in reconciling the differences. In the rats previously treated with 0·1 c.c. of tumour 27, there was a rapid degeneration of the cells of the new graft; this degeneration involved the cells throughout the whole graft, and led, in about three days time, to their total necrosis. If any difference could be detected here in the location of the process, that is, between the cells at the periphery and those centrally placed, it was rather in favour of the latter retaining their vitality for a longer period, but the difference was not marked enough to lay any weight upon it. The rate of degeneration in immune rats was, however, much more rapid than in immune mice, and this can be interpreted as being due to a more active production of the inimical factors in the rat, and, in addition, to qualitatitive differences in their nature when induced in a strange species. Considering that the rats had been immunised with tissues from another species viz., the mouse, this result is only what would have been expected. Further experiments are being made to determine whether the pheno

mena in the case of rats lend themselves more readily to investigation in vitro than in the case of mice.

The influence exerted by the body-fluids in the immune mouse cannot be referred to an active cell-poison, since it allows the implanted cells to continue proliferating for about 8-10 days, but eventually it injures them to such an extent that they can no longer withstand the encroaching fibroblasts of the new host. These latter penetrate between the cells and producing scar tissue instead of the connective tissue scaffolding required by the parenchyma cells, they compress the latter, leading to their final atrophy and degeneration.

The minute details of the final process are similar to those described by Orth1 in human epitheliomata, by M. B. Schmidt, Lubarsch and Gierke3 in carcinomatous emboli in the lungs, and by Bashford, Murray and Cramer1, and later by Gaylord and Clowes, in spontaneous healing of mouse tumours. The probability is great, therefore, that in all these cases, local or general conditions interfering with the requisite re-organisation of the affected groups of cells, are responsible for the progression of the reactive processes towards encapsulation of the cancer cells, and spontaneous healing.

In conclusion it is necessary to express thanks to the Executive Committee of the Imperial Cancer Research Fund for permission to work in the laboratory, and to Dr. Bashford and his colleagues for their encouragement and assistance, and for the opportunities which have permitted the conduct of this investigation, and my fulfilment of the conditions of tenure of the Georgina McRobert Cancer Research Scholarship at the University of Aberdeen.

1 Zeitschr. f. Krebsforschung, Band i. Heft. v. 1904–55, § 399.

2 Die Verbreitungswege der Karzinome. Jena, Fischer, 1903.

3 Page 131 of this Report.

4 Discussion on paper on "The Growth of Cancer." Transactions of the Medical Society of London, 1905, also Second Scientific Report of the Imperial Cancer Research Fund, 1905.

* Surgery, Medicine, and Gynecology, June 1906.

RESISTANCE AND SUSCEPTIBILITY TO

INOCULATED CANCER.

BY E. F. BASHFORD, M.D., J. A. MURRAY, M.B., AND
M. HAALAND.

IT has been pointed out repeatedly in the preceding pages and elsewhere, that with the achievement of the experimental propagation of cancer, opportunities are afforded for studying the conditions of growth of malignant tumours. Experimentally produced conditions unfavourable to growth, have received particular attention from practically all investigators of experimental cancer, because of the aspiration to further the rational treatment of malignant disease in

man.

In this connection more almost than in any other, the distinction between the continued growth of cancerous cells either in the animal spontaneously attacked, or in normal animals to which they have been transferred experimentally, and the inception of malignant proliferation must be ever present in the mind of the investigator and of the reader. This consideration has dictated the wording of the title of this paper, and although in the sequel, the terms immunity," "immune mice," etc., will be frequently used, this is done entirely as a matter of convenience, and in no case must be taken to imply a diminished liability to the development of spontaneous cancer, but only such an alteration in the animals, that on introducing intact cancer cells into them, no tumour is produced.

66

The description of the various procedures by which this effect can be produced, and the precautions which must be observed in carrying out the experiments or in interpreting the results, will involve some repetition of observations already described. This is rendered necessary by the intricacy and delicacy of the processes, and by our desire that others may be enabled to repeat our experiments.

Since the publication of the paper, reprinted at p. 315, on "Natural and Induced Resistance of Mice to the Growth of Cancer," in which the earlier literature is discussed, this subject has been dealt with in communications by Haaland, Borrel, Bridré, Schöne, Lewin, Flexner, and Jobling.

Haaland was able to show, that mice might be resistant to one strain of transplanted cancer and yet susceptible to another, and that the absorption of even considerable quantities of Jensen's carcinoma did not necessarily protect against subsequent inoculation of the strain of Ehrlich's experimental sarcoma with which he worked. Borrel and Bridré showed, confirming earlier observations, that resistance could be induced by the absorption of tumour material without intervening tumour formation. That a certain considerable quantity of tumour tissue was necessary, and that similar results could be achieved by preliminary treatment with normal mouse tissues (liver, spleen, but not testis), and not by tissues of other species. They devoted considerable attention to the results of successive inoculation with the same and other tumours, and came to the conclusion, that where small doses were used, the second inoculation generally gave a result parallel to that of the first inoculation.

In the account of the general features of the propagation of mouse tumours on p. 262, attention is again drawn to the importance of technical minutiæ in modifying the results of experiment. All that is there laid down, applies with increased force to experiments designed to show differences in growth in animals subjected to preliminary treatment. The greatest care must be exercised that the animals used as control in all such experiments should conform closely in race, age, weight, and general health, to those with which they are compared. Further, the material for inoculation must be distributed among the animals as uniformly as possible. possible. To this end, the site of inoculation and the dose of material inoculated must be identical, and the latter must be accurately measured. When inoculation is carried out with the hollow needle and plunger, uniformity of dosage is attained by paying the greatest heed to the equal subdivision of the material and to the choice of the fragments. When the syringe method is used, the tumour emulsion must be as uniform as possible, prepared by sharp cutting instruments and not by crushing. The dosage in the syringe method is by volume, and when this is small, it is important to use the finest hypodermic needle which will transmit the fragments making up the emulsion of tumour.