The plurality of localizations of viruses also can be related to affinities of mutated agents for different cells. The experiments of Gross (105), which show the possibility of separating out of the same filtrate the agents responsible for salivary gland tumors and for leukemic changes, would indicate that a change in the virus must be also considered. However, the change can be interpreted as a mutation and can be related to the influence exerted by lipids upon the virus. Treatment of virus with lipids has shown the possibility of inducing changes in its behavior. Data showing the influence upon tissular receptivity of such changes will appear in future publications.

With this concept of the role of viruses in the pathogenesis of cancer, it seems possible to explain other peculiarities that have led to confusion in this field.

It has been noted that viruses act as factors determining the change to a cancerous entity which, once induced, can continue to develop without need of further intervention of the virus. This poses the problem of the relationship between viral carcinogenesis and development of the virus in the tumor itself. Even in a tumor, the multiplication of the virus has to be separated from that of the growth of the tumor. Although often interrelated, they must be considered as two different processes. The growth and even direct transmissibility of the tumor can continue, independent of the presence of the agent that originally induced it. When tumors have been induced by a chemical carcinogen, they can be transmitted in continuous generations over many years, producing large tumors each time, a fact which would preclude any possible direct intervention of the agent in these later tumors. Similarly, tumors once induced by virus can be further transmitted by cells, the virus no longer being apparent in the tumors. A tumor induced by a virus often serves as a medium for the multiplication of the virus. However, even while the tumor can continue to grow, it can become an adverse medium for the further multiplication of the virus.

This explains the peculiar fact that tumors induced by a virus can be rich in or can lack an appreciable amount of virus, as often seen in a Shope papilloma (106), or even in tumors in fowl (107), which pass through periods when transmission through cell free filtrates becomes impossible, while transmission through the transplant of tumor cells still continues. The virus multiplication capacity can vary not only with the host but with the virus itself, thus explaining the changes noted above.

The results of the interesting studies of Bryan, Galman and Maloney (108) who have investigated the relationship between richness in virus of an induced Rous sarcoma and the percentage of positive results in induction easily can be interpreted under this view. The chances of inducing tumors increase in cases in which the host is also a favorable medium for the multiplication of the virus, and vice versa. This would explain why these authors found little or no active virus in cases in which the injected material produced less than 50% positive results, but cases originated by a material that induced a large proportion of positive results were rich in active virus. The capacity to multiply after inoculation in the host itself thus increases the ability of the virus to act as a carcinogenic agent, which seems logical. The relative independence of the two processes—the multiplication of the virus and the induction of tumors—appeared clearer in the cases mentioned above, where the virus develops in the entire body of mice, even in successive generations, without inducing tumors.

The presence of viruses in the organism, even without inducing tumors, helps to explain the rather puzzling experiment in which a transmission through filtrates, considered characteristic for viruses, was seen to occur for tumors induced by chemical carcinogens. Carrel has claimed to have transmitted through filtrate passages tumors induced by arsenic, tar preparations and even indoles. These tumors were of the Rous type obtained in fowl. More recently, similar tumors transmitted through filtrates were observed by Mcintosh and Selbie (109), Maisin, Haddon and Haagen (110), and Oberling and Guerin (111) after injection of methylcholanthrene. especially in fowl. The considerations presented above furnish a logical explanation for these observations.

A first factor to consider is the presence, in animals regarded as normal, of a virus able to intervene to produce a neoplastic effect under special circumstances. We have noted that such a virus can be present without inducing tumors. Fowl appear to be especially susceptible to viruses (112), statistics showing that viral lymphomatoses are responsible for 50% of the malignancies in chickens. Even while some species display an inborn resistance to viral infection, others are highly receptive, as seen for the viruses of sarcomas (113) and lymphomatosis. (114) Viruses have been found in as many as 10% and even 20% of chickens, according to some reports. (294)

The number of animals with viruses and no tumors must be considered still higher when presence of viruses is revealed by antibodies. Duran Reynals, in collaboration with the East Lansing Agricultural Experiment Station (115), has shown that, while not one of 23 chickens kept isolated and free of lymphomatosis showed antibodies in the blood, hundreds of chickens taken at random did have the antibodies. This fact makes it highly probable that the presence of the viruses in the chickens used for carcinogenic studies was independent of the administration of the chemical carcinogen. Furthermore, the role of chemical agents acting alone as carcinogens can be discounted because their effects differ widely. In the series of Mcintosh, the tumors appeared far from the site of injection which is very unusual for methylcholanthrene. Furthermore, the agents used by Carrel, except tar extracts, generally have little or no carcinogenic activity.

The two hypotheses—one, that chemical carcinogens alone can induce filtrable tumors; the second, that this is only coincidence and the tumor is entirely of viral origin—can be reconciled under the concept of plural intervention. Thus the chemical carcinogen would induce only part of the process; the remainder, at higher levels, would result from viral intervention. The change that occurs in the cells through the influence of the chemical carcinogen could also favor the change in the virus, making it not only more active but also of neoplastic character. The concept of plural changes needed to induce active carcinogenesis permits us not only to integrate the intervention of viruses in the concept of carcinogenesis in general, as presented above, but also to consider this intervention in relation to other factors.

Hormones can play a part; they are needed to induce the degree of differentiation that is a condition for viral co carcinogenic intervention. Inoculated intraperitoneally, the Bittner milk factor, although active, will rarely induce mammary tumors in virgin females although the virus can be proved to be in the body. The hormonal changes related to pregnancy and lactation influence the mammary gland and cause a differentiation. Since this differentiation represents a condition for viral neoplastic activity in these cases, hormonal intervention can be integrated as an added factor important for the viral carcinogenic activity. The hormonal factor would appear to be an indirect co carcinogen, and it is under this aspect that its role in carcinogenesis has to be studied.

The concept of plural co carcinogenic intervention permits us not only to relate the different pathogenic factors involved in carcinogenesis; in addition, by relating these factors to various levels of organization, it allows us to obtain new insight into the genetic factor.

Genetics And Carcinogenesis

The genetic factor in carcinogenesis can be understood in terms of phylo genetic hierarchic development. Such development results from a series of progressive changes which lead to successive hierarchic levels of organization. As we have seen, for each level of the organization a series of different solutions are available when a new hierarchic entity is to be realized. This results first from the fact that various numbers of entities take part in the constitution of the principal parts. Since different constituents can form the secondary parts of these new hierarchic entities, the number of solutions is increased. The resulting solutions can be considered on a statistical basis. Many of these new entities will die immediately; others will subsist as such; still others will progress. Their fate results largely from their interrelationship and the conditions present in the environment in which they find themselves.

The striking similarity to the resonance process studied in the lowest levels of organization, such as atoms or molecules, has led us to consider that changes at higher levels are of the same fundamental nature. Of all the resonance forms that occur at each of the levels, there are some that, on a statistical basis, persist and develop. These persisting resonance forms make up the normal organism. The favored resonance forms are determined by heredity and also by environmental conditions. While the resonance forms appear on a statistical basis, the environmental condition can vary and new resonance forms will mark the intervention of external factors. As a normal entity is composed of the persistent resonant forms, abnormality occurs when such an entity persists. The characteristics of any individual are provided by the resonance forms which have developed phylogenetically and also ontogenetically. These predominant forms are "isotropic." For didactic purposes, we called the others "allotropic."