In some virus-induced tumors, the morphologic observation of virus particles in the tumor tissue itself presents no difficulty. The Shope fibroma of the rabbit is a good example (55). The high infectivity of extracts of the growth is related to the presence of large quantities of virus. The agent belongs to the pox group and is, therefore, of large size (around 200 mµ). The development of the virus particles can be followed in vivo as well as in vitro as early as the 6th hour after infection (56). The same applies, also, to molluscum contagiosum (27, 28, 67), because this disease, also, is caused by a virus of the pox group, the members of which are all of large size. Titration methods have not been developed for the etiologic agent, but the great quantity of particles may be regarded as related to high infectivity. These two examples fall into a special category of probably exceptional importance because of the link they represent between tumors and classical pox infections.

Apart from these conditions, and with the exception of some strains of mammary cancer, there is no virus-induced neoplasm in which detection of a viral particle with the electron microscope has not meant, at the beginning, a tremendous effort and in which tenacious search has not been necessary. In the Rous sarcoma, Claude, Porter, and Pickels (11) succeeded in finding particles immediately, but this could not be repeated for a long time in spite of an arduous search (58). When at last particles were detected, their extreme rarity was striking (59); only 0.38 percent of the cells contained particles. The scarcity of particles has continued to be remarkable, also, in subsequent studies (60, 61). In the Murray-Begg endothelioma, though virus was found (62) in 33 tumors examined, the particles were extremely difficult to detect, and sometimes only 2 or 3 were found after hours of search in many samples. This was also true of Fujinami's sarcoma (63) and of the fowl leukemias when the search was made for particles in circulating blood cells (37).

In the mouse, intracellular particles may be extremely scarce, as in the leukemic organs of mice with Gross' virus-induced leukemia (64) and in some strains of mammary tumors in which no particles were found (48), even though the Bittner agent could be demonstrated biologically (52, 65, 66). Moreover, virus particles were not found at first in mouse polyoma tumors studied in vivo. It was only later that these were observed in hamster kidney tumors induced with a strain of polyoma virus isolated in Toronto (19). The picture of the Shope papilloma virus observed (IS) for the first time in skin growths was revealed in only 1 case out of 19 examined (67).

Other examples could be found to illustrate the difficulty with which particles are sometimes detected. This needs to be stressed repeatedly, because it is by no means apparent when pictures of various tumor viruses, sometimes spectacular, are being shown. The difficulties in this respect are not only of great practical importance but of much theoretical meaning. Since no other ultrastructural sign is characteristic, the presence of the viral particle in a tumor acquires a fundamental significance, if its morphology, as seen, for example, in purified preparations (fig. 21), and the experimental conditions are well defined. It is thus of critical importance to seek strains of animals in which intracellular virus particles are plentiful. Such may represent laboratory "monstrosities," a criticism sometimes raised, but the essence of caricature is but to emphasize and thus better to reveal some traits of the model. In the study of neoplasms in which particles are not readily detectable for one reason or another, the role of the investigator, and particularly of the electron microscopist engaged with the viral hypothesis of cancer, is to attempt to develop the conditions under which virus particles become morphologically detectable.

It is remarkable that this has really been achieved. One has been able to show that it is possible to modify the quantitative aspects of cell-virus relationships in such a manner that virus particles can be detected in the electron microscope relatively easily where previously they had not been apparent. Studies with Rous sarcoma will serve to illustrate this.