Tumor viruses as well as infectious disease viruses, although more or less specific for certain tissues in a given species, may have a different pattern of tissue selectivity in another host species. For example, herpes simplex virus is predominantly dermotropic in man, but relatively more neurotropic in rabbits. Viruses of the Coxsackie group are another example. The polyoma virus is perhaps the most remarkable among the tumor viruses, since it produces tumors of the salivary epithelium, mammary epithelium, hair-follicle epithelium, thymic epithelium, and in several other sites in the mouse, including some connective tissues (5). In the rat, only connective-tissue tumors are produced (0). Such species differences suggest an analogy with morphologic control mechanisms recognized in experimental embryology. For example, in the grass frog (Ranafusca) the eye vesicle is necessary for induction of lens development from the overlying epidermis (7). In the water frog (Rana esculenta) the lens will develop in the absence of the influence of the eye vesicle [see Grobstein (8) for discussion of time and quantitative factors]. Furthermore, "strange" ectoderm from the belly of the water frog fails to form lens in response to eye vesicles that do induce lens either in lens-forming or belly-ectoderm epithelium of R.jusca (9). From this it would appear that lens epithelium, although known to be similar antigenically Over a wide species range, as well as ectodermal epithelium in other regions of R. Jusca, is biologically a different tissue from comparable epithelium of certain other frogs, and it might be expected to respond differently to a single hypothetical virus.

A similar example is to be found in the induction of balancers in larvae of newts and salamanders. Larvae (axolotls) of the salamander Amblys-toma tigrinum do not have organs known as balancers, whereas larvae of the newts (Triton) do have them. However, it was shown by Mangold (10) that the proper area in the head of the axolotl is capable of inducing balancers in grafted belly ectoderm of the newt. Thus a certain area of ectoderm from rather closely related species responds differently to a given morphogenetic influence. If a virus or any other tumor-inducing agent were to be cast in the role of the morphogenetic influence, the situation would be analogous to positive and negative reactivity to that agent by homologous tissue of different species. A third example illustrates a slightly different type of analogy. A specific region in the mouth endoderm of the urodeles has been shown to be responsible for the development of teeth. When the prospective tooth-developing primordium of the urodeles is substituted by the same region from anurans, a horny type of jaw characteristic of the anurans develops instead of teeth (11). Here, homologous areas of ectoderm respond to the same morphogenetic influence, but in a different way. Experimental embryologists (12) have referred to an "intrinsic latent repertoire of potencies" characteristic of each species, and ascribe this repertoire to limitations placed upon cell reactivity by genetic factors. Analogous situations for virus infection in mammals are difficult to find, possibly because in animals infected at birth or later, the variety of ways in which most cells can react has already been realized with reference to differentiative direction. Perhaps an example is to be found in the mixed tumors of fat and bone (18, 14) in mice receiving polyoma virus. Such tumor response with bivectoral differentiative capacity has not yet been reported in rats or other animals infected with this agent, though simple bone tumors of the skull in rats have been noted by Fogel and Sachs (15).

Adding to thqse illustrations some experimental evidence for a contact infection-like behavior of certain developmental organizers (16-18) [see also Grobstein on viroid transmission of inducers (8)], we are provided with enough resemblances between virus systems and developmental induction systems to warrant a search for possible interlocking areas.

Before one can take advantage of species differences in devising experiments to investigate possible relationships between morphogenetic and differentiative influences and viral response, it is necessary to have information regarding each of these two systems centered about the same tissue. How, then, do different species respond to polyoma virus?

The pitfalls introduced by dose-response effects and viral strain d'ffer-ences must always be kept in mind when this question is answered. It is regrettable that there is not available a carefully quantitated study to answer it fully. At the all-or-none response level, however, it was established in the act of discovery of this agent (19-21) that mice respond to it with tumor development in the salivary glands, adrenal glands, and subcutaneous connective tissue. Later Eddy et al. (22) reported an oncogenic response by golden hamsters, and then by rats (6). Eddy and Stewart (28-25) next reported the development of tumors, followed by regression, in rabbits. More recently, Rabson et al. (26) have found that the multimammate African rodent, Rattus (Mastomys) natalensis, also develops tumors in response to polyoma virus. This is the sum of present positive knowledge on species susceptibility in vivo. Numerous reports confirm the tumor-inducing activity of the virus in mice, hamsters, and rats, and only a few are cited here (13, 16, 27-29).

Negative data are of almost equal importance. Rowe (SO) has failed to find serologic evidence of polyoma-virus infection in man, and isolation of this virus from man or from human tumors has not been positively demonstrated. Rowe (SO) has also found opossums, porcupines, chickens, dogs, and other wild as well as domestic animals to be free of virus, and to have negative serologic reactions. Although Gross (31) mentioned induction of parotid tumors in a few mice receiving extracts of normal guinea-pig tissues, guinea pigs have not been reported to develop tumors in response to polyoma virus. Fogel and Sachs (16) reported no tumor response by guinea pigs, chickens, and dogs after observation period up to 4 months. Eddy and Stewart (32) are observing monkeys inoculated with polyoma virus, but thus far have not noted tumors attributable to the effects of this agent.