The "warts" which tar elicits on rabbit skin (papillomas, carcinomatoids, frill horns) are true tumors, benign growths expressive of slight yet irreversible deviations of epidermal cells from the normal. The neoplastic condition gives the cells a superiority over their neighbors when both are submitted to the same encouraging influences, and then they proliferate into tumors. Their state entails such disabilities, though, that they are unable to maintain themselves under ordinary circumstances, and consequently growths composed of them disappear when no longer aided. Often the neoplastic cells resume the normal aspect and habit of life long before the tumor mass is gone; and they may persist as part of an apparently normal epidermis, retaining their neoplastic potentialities for months after all signs of the growth have disappeared. In these instances it can be made to appear again, sometimes repeatedly, by non-carcinogenic stimulation of the skin (wound healing, turpentining). There is reason however to suppose that in the end the tumor cells, unless helped, die or are cast off. It is plain that the neoplastic state does not necessarily connote independence of behavior or success in tumor formation. On the contrary it may render cells unable to survive or endow them with powers which they can exert only under favoring conditions. This is the case with the cells composing the tar warts of rabbits. In the lack of such conditions the cells of these growths do not manifest themselves but remain in a subthreshold neoplastic state, whereas if aided they form neoplasms. The deviations from the normal represented by the benign tar tumors of rabbits are slight and limited in character, but further deviations in larger variety may be superimposed upon them, with result in malignant tumors, growths possessed of a greater, though not always absolute, independence. Tar cancers usually come about in this way, by successive, step-like deviations from the normal, and so also do the cancers which derive from virus-induced papillomas as well as many human carcinomas. After cells have become cancerous they frequently undergo further changes, some apparently step-like in character, and all taking the direction of greater malignancy. The hypotheses that tumors are due to somatic mutations and to viruses respectively are discussed in the light of these phenomena.
