Invisible infective agents may be divided into: (1) true, ultramicroscopic, living viruses, which do not arise de novo and, so far as is known, are not ubiquitous; (2) transmissible infective agents which arise de novo and are propagated through living cells, but are not themselves living organisms; (3) stimulants to variation which arise de novo, are not transmissible, and are not living organisms.

Class (1) is not represented in malignant disease. “Bacteriophage” is a representative of class (2); very probably the infective agent of fowl sarcoma comes under the same category, and possibly some important human diseases of doubtful aetiology. There is no satisfactory evidence that mammalian malignant disease is related to class (2); its causation, according to the “chronic irritation” theory, must be attributed to influences comprised within class (3).

The stimulants to variation in class (3) depend for their effectiveness upon the unstable energy of living matter. The changes which they produce are “biological” in the sense that they are changes of chemical constitution which could not be obtained without the aid of vital processes.

Regulation of normal growth in the animal body means regulation of the cell's facilities for obtaining energy. I think it is misleading to regard it as a forceful restraint (or stimulus) upon the cell's inherent capacity for unlimited growth.

The assumption, borrowed from “natural immunity” towards bacteria, that there is in the animal body a natural principle which destroys frequently occurring foci of incipient malignancy is also unsubstantiated and misleading.

During the latent period, certain cells, which subsequently grow into a neoplasm, lose their capacity to respond to inhibitory systemic influences. This change is brought about by local and not by systemic causes.

As regards the special class of tumour derived from cells which have been displaced in foetal life, long residence in an abnormal situation does not appear to be equivalent to the ordinary latent period; but it may have had the effect of increasing their susceptibility, so that, if exposed to chronic irritation, the cells would more readily lapse into the latent period predisposing to malignancy.

It is known that the various tissues of the animal body differ in their degree of susceptibility to the precancerous change. This is a cellular characteristic; so also is the difference in the susceptibility of one animal as compared with another. It is not a question of difference in a hypothetical humoral property of “systemic resistance.”

On the termination of the latent period by a fresh stimulus to proliferation, certain cells commence active growth and are incapable of responding to systemic inhibitory influences. These conditions seem sufficient for the origin of an innocent neoplasm. But something more is required to explain malignancy, because the malignant cell is essentially different from the cells in a benign tumour.

About the actual cause of the change to malignancy one can only offer conjectures. I have suggested a way in which the change may possibly be produced through the agency of the local endothelium and the autogenous formation of antibodies.

On taking a broad view, the change into the malignant variant is not something unique; equally remarkable changes are to be found in the properties of bacteria. In both cases the facts have to be accepted, at present, without satisfactory explanation of the conditions which gave rise to them. One finds with bacteria that degradation or “roughness” may be a phase preparatory to the acquirement of new properties, just as the degradation of cells in the latent period seems to be a requisite preparation for acquiring the new property of malignancy. But the actual steps involved in the change from a bacterial saprophyte to an invasive parasite are as difficult to understand as are the processes involved in the conversion of a normal animal cell into its malignant variant.

Throughout the study of cancer it is very desirable to maintain a clear distinction between cause and effect. For example, the enzymes peculiar to cancer are not the cause of cancer but the effect of the biological change which produced the cancerous cell.