Results of studies on cell ultrastructure permit only one conclusion: Without the presence of the viral particle itself, no cytologic lesion has been revealed by electron microscopy that will allow unequivocal inference of the viral origin of a tumor, at least with the present knowledge. The subject treated here was requested for this symposium for the precise purpose of presenting a comprehensive discussion of possible specific lesions to be found in virus tumors. It is of first importance to realize that the results are negative from the point of view of the host cell but, in contrast, are positive with respect to the agent insofar as experimentally induced tumors are concerned.

The implications of these observations, negative and positive, are already tremendous. It must not be forgotten that, as recently as 2 years ago, some virologists chose to adopt an austere attitude toward the findings by electron microscopy. The best that could be said, as quoted from one authority, was that ". . . certain kinds of particles are seen in one kind of tissue and appear to be absent in the other." A new phase of accomplishment has now been instituted in which the morphologic identification of viral particles with the electron microscope has become a valid method without which no further work on virus-induced tumors may be considered as complete (88). Impeding prejudices have been overcome. Previously, the morphologist, who was usually no virologist, had developed an inferiority complex because of the sad situation in which he had inadvertently found himself, either by discovering particles without being able to prove biologically that they were the causal virus or even virus at all, by not discovering them in highly malignant virus-induced tumor tissue, or still worse, by discovering them where he was not expecting them in control tissues. To each of these enigmas he now has an answer. He knows that some virus-induced tumors may harbor no particles at all and why, as in the Rous sarcoma. He knows, as illustrated for instance by the work on mouse mammary carcinoma, that the particle he observed in the tumors of the animal is the responsible agent. He also knows ways in which he can stimulate the development of particles, so that they will be numerous enough to be easily detected in the electron microscope. He even has a most plausible explanation for the presence of particles in "normal" controls. The virus particle, as it appears in tissue sections, has then regained its rightful place-after quite a battle. It has also acquired a unique value, since no other unequivocal sign of malignancy or viral involvement has been revealed.

In its turn, the absence of special cytologic features in virus-induced tumor cells, together with the frequent absence of particulate elements, is of fundamental importance in relation to the problem of human cancer. It has been shown and emphasized that thus far there is no ultrastructural difference between the cells of virus-induced tumors and the cells of a human cancer. Morphologically, on histologic as well as ultrastructural grounds, the research worker is in excellent position to accept the similarity. Experimentally, he is now in a situation where he can try to test the validity of the viral hypothesis of cancer with new techniques. No electron microscopy of cells from human cancer cultivated in vitro has been carried out extensively up to now, at least not to our knowledge. Attempts in this direction-the uncovering of corpuscular cellular elements-should be coupled with efforts to detect components perhaps present in a noncorpuscular or "molecular" state and comparable to genetic material (88). This other fundamental aspect of the problem cannot be developed in this discussion.

Plate 25

Figure 1. Normal-looking and relatively well-preserved nucleus of human breast cell. X 15,000.

Figure 2. Apparently badly preserved nucleus in cell from Rous sarcoma. Note margination of chromatin, patchy destruction of nucleoplasm, and presence of granules (→) of unknown significance. X 18,000.

Normal looking and relatively well preserved nucleus of human breast cellApparently badly preserved nucleus in cell from Rous sarcoma

Plate 26

Figure 3. Portion of nucleus of human breast carcinoma cell. Note aggregate of small granules and "dense bodies" (→) scattered at random in mottled nucleoplasm. Latter bodies are frequent in nuclei of virus-infected cells. NM-nuclear membrane. X 32,000.

Portion of nucleus of human breast carcinoma cell

Plate 27