others which cannot yet be made evident to the eye, except in their consequences. Such are, particularly (1) the power possessed by some rare carcinomata of producing sarcoma in the connective tissue of healthy mice, and (2) the conditions known as immunity reactions, referred to below.

Sarcoma Development. The remarkable phenomenon was first described by Ehrlich and Apolant.' The process has since been observed in other laboratories. It consists essentially in a transformation of the stroma into sarcoma because of some hitherto undefined influence proceeding from the carcinoma cells. Ehrlich and Apolant assumed that the end of all carcinomatous tumors during propagation would be their replacement by sarcoma. Therefore the loss of power to induce sarcoma is of great theoretical interest.

The loss of the power to effect a sarcomatous transformation of the connective tissues occurred in five carcinomatous substrains derived from a strain which called forth sarcoma development irregularly, and its appearance was promoted rather than hindered by rapid repetition of the transplantations. Another tumor caused the development of sarcoma in practically every animal, from the twelfth to the twenty-fifth transplantation, in all the sub-strains propagated, if the tumor was permitted to grow for about two months; but remained pure carcinoma if transplanted at intervals of a month or less. Four carcinomatous sub-strains which were kept growing presented variations from this behavior.

In the first sub-strain the variation was in the direction of an earlier appearance of the sarcoma. This occurred after the twenty-fifth, and continued during the next twenty transplantations, until, finally, the early onset of sarcoma development, combined with slow growth of the carcinoma proper, rendered it impossible to retain the carcinoma in propagation, and the strain became pure sarcoma. In the second sub-strain the change began to appear at a later period, occurring only after one hundred days or more, instead of after sixty days. Subsequently, even after two hundred days, the tumors remained pure carcinoma. This condition continued during fourteen transplantations, or fifteen months, and the power of producing sarcoma was ultimately lost. Sub-strains 3 and 4 began by causing sarcoma about the sixtieth or seventieth day, continuing to do so up to the fifty-third transplantation, but thereafter the appearance of sarcoma was delayed, so that the complete loss of this power seemed probable.

Ehrlich and Apolant.-"Beobachtungen über maligne mäuse tumoren," Berl. klin. Woch., 1905, 28, 871.

The power to induce sarcoma development is characteristic of only a small minority of transplantable tumors. When it occurs it is almost certain that it is a characteristic of the primary tumor. The same phenomena have been observed in transplantations of spontaneous tumors into the mice from which the primary tumors were obtained. This condition has also been demonstrated in man.

At the Rockefeller Institute growths in the fowl have been discovered which, under the microscope, it is impossible to distinguish from sarcoma, and which are transmissible by means of a virus contained in a cell-free filtrate obtained through a porcelain filter and by a dried powder. Cell-free filtrates prepared from carcinomatous tumor, and tumor of mixed carcinomatous and sarcomatous structure, as well as pure sarcoma of the mouse, did not lead to tumor formation after inoculation. It has been impossible, therefore, to separate the property of producing sarcoma from the vital activity of mouse tumor cells. The same negative results followed the inoculation of these tumors when dried in vacuo. The differences between the "sarcoma" of the fowl and the mouse in these respects are not fully cleared up, and the sarcomatous nature of the fowl tumor is received with skepticism by some authorities. The nature of this activity of some mouse carcinomata is still as obscure as that of the origin of malignant new growths in general. The attempt to explain sarcoma development by the transference of a virus from the carcinoma cell to the connective tissue is unsatisfactory, since it will not explain the subsequent loss of this property and at the same time the continued growth of the carcinoma cells.

The Immunity Reactions-The employment of the term "immunity" with reference to cancer signifies nothing more than the exemption of the individual-under clearly defined circumstances from the consequences of transplanting a tumor from another individual of the same species.

Resistance to the transplantation of cancer includes a large number of phenomena which are of the nature of hindrances to the continued growth of the already fully developed cancercell. The analysis of these phenomena has permitted the discovery of some subtle properties of the cancer-cell, as well as of some of the relations existing between it and the animal in which it grows, both in the case of the natural and in that of the transplanted cancer-cell.

Artificially induced resistance to the continued growth of grafts does not create any exemption from the liability to the development of cancer. It is certain, too, that cancer is rarely,

if ever, communicated naturally or spontaneously from one individual to another by transplantation, and that its great frequency cannot be explained in this way. Therefore the mechanism of its natural development differs fundamentally from transplantation, and the use of the term "immunity" may be justified only by convenience.

The study of such artificial resistance, however, has thrown light upon the nature of the tumor-cells and their relation to the individual serving as food purveyor. But sharp distinctions must be drawn between heterologous, homologous, and autologous1 inoculations, whether employed for the purpose of inducing or of testing resistance.

Resistance to Heterologous Inoculation.-Heterologous inoculation may be practiced in two ways. One species of animal, say a rat, may be inoculated with the tumor of another animal, say a mouse, and the rat may then be tested as to resistance to an inoculation of its own kind of tumor, or for resistance to mouse tumor. In the first instance no resistance is induced; in the second there is.

There is general agreement that the normal tissue of tumors of one species of animal is incapable of progressive growth, or even of continued existence, in another species. Very extensive investigations have been made into this subject, and the few positive results cannot be set up against the enormous preponderance of negative results obtained. By far the greater mass of accumulated data shows that tumors of one species are incapable both of progressive growth in another species and of power to induce resistance to tumors of the species inoculated. These two lines of inquiry, therefore, agree in bringing out general exemptions which can be explained only on the basis that tumors have a specificity analogous to that distinguishing the organisms from one another as zoological species. Of themselves these two groups of phenomena have no bearing upon the resistance which may be induced against the inoculation of tumors into other individuals of the same species. In conjunction with other studies they do have a bearing upon the question of whether an immunity is induced analogous to that against infective diseases, by showing that, when induced, it does not depend on a virus common to cancer in whatever species of animal it occurs.

Induced resistance to a repetition of the heterologous inocu

1 Heterologous inoculation = from one species to another.

lation has been demonstrated to occur and to give reactions identical in nature, both in vivo and in vitro, to the well known hemolytic and cytotoxic reactions induced by immunizing with antigens from strange species. It is possible that the tumors and tissues of nearly related species-for example, the blood of the rat and of the mouse-do have a slight reciprocal influence; but, if so, its existence only serves to emphasize the importance of blood-relationship rather than the intervention of a cancer parasite common to the different zoological species.

The tissues need not be alive in order to induce the hemolytic, cytolytic, and heterologous cancer immune reactions, but retain this power, for example, after mechanical disintegration.

Heterologous Immune Sera. One of the hopes awakened by the knowledge of the cytotoxins was that of a further development of a rational organotherapy. In the promotion of more accurate diagnosis, and in forensic medicine, the methods relied on have yielded advances of prime importance; but the hope of fresh therapeutic triumphs has not been fulfilled. The sharp specific distinctions which at times may be drawn in vitro between normal and cancer-immune sera do not necessarily obtain also in vivo. In the test-tube they present, in varying degree, hemolytic precipitin and cytolytic reactions which as yet have not been shown to have any distinctive characters. "Cancerimmune" sera must to-day be regarded as devoid of all action in the living body. As vehicles for the communication of passive immunity, such heterologous sera have proved valueless, and likewise they have not been demonstrated to possess qualities which may be relied upon for purposes of diagnosis.

Under this category it is convenient to refer to the results which have been interpreted as toxic or anaphylactic phenomena following a repetition of an inoculation. On the basis of careful and numerous repetitions of the method on which such observations have been advanced, the conclusion has been reached that they are due, not to anything specific to cancer tissue, but rather to accidental bacterial contamination.

Homologous Immunization and Resistance.-Homologous resistance may be considered from the standpoint of whether it is preëxistent and natural or induced by active changes following various procedures.

Natural Resistance.-Natural resistance played a much more prominent part at the beginning of the investigation of immunity to cancer than it does to-day. During the period when an appropriate technic was being developed, many observations could not be otherwise than tentative. For example, with one form of technic the primary transplantation of certain hemor

rhagic mammary growths of mice succeeded not more than once in 500 attempts. With a different technic these growths were as easy to transplant as other mammary tumors. Naturally the appearance of 499 resistant mice out of 500 was calculated to give great importance to natural resistance. The real explanation, however, was found in the employment of such large doses for inoculation that auto-immunization was induced and growth stopped. Perhaps also the greater suitability for growth of young than of old mice was not fully recognized at that time. The existence of a certain amount of natural variability in power of reaction in different animals cannot be denied, and account must also be taken of age, site, dose, and time interval. Although some tumor-strains are relatively indifferent to age, no strain has been found which grows better in old than in young animals. Young animals, as a rule, are more susceptible than old ones, whether they come of normal or of cancerous stock, the latter not offering a more favorable soil.

Since young animals are more suitable for inoculation and continued growth than old animals, the rarity of cancer in the young is not due to constitutional resistance to growth, and its frequency in the old is not due to a constitutional change, occurring with senescence, favorable to the growth of cancer in general. This conclusion suggested experiments to determine whether animals naturally suffering from the disease offered a more suitable soil than others for the growth of cancer in general. It was found that they did not do so. Whereas the reinoculation of an animal with its own tumor is almost always successful, the success of the primary transplantation on inoculating other cancer-animals is on a level with that for normal animals. The strict individuality of spontaneous tumors has been demonstrated by the resistance which the majority of strange hosts offer to the introduced cancer-cell, and by the fact that it is impossible to immunize a mouse against an inoculation of its own tumor.

Active Resistance Induced by Tumors.-Active resistance may be induced in a variety of ways-by inoculating with spontaneous or propagated tumor without obvious growth following; by inoculation with propagated tumors which, after a considerable amount of transitory growth, become absorbed; or by inoculating with normal tissues.

If the dose of spontaneous tumor inoculated without giving rise to daughter-tumors in normal mice has been sufficiently large, they will exhibit resistance to a second inoculation of a propagated tumor which takes in maximum percentage and possesses extreme power of growth. The extent of the resistance is