Benzpyrene brings about neoplastic changes in rabbit epidermis much sooner than has been supposed. The long interval that elapses before visible growths appear is due in the main to the relatively slight power of the carcinogen to encourage multiplication of the cells it renders neoplastic. Yet some slight power of this sort it has. Methylcholanthrene has somewhat more but not nearly so much as tar. It may initiate neoplastic changes within less than 17 days, as compared with less than 10 days for tar, but tumors due to it do not ordinarily appear until months after those called forth by tarring. All three agents give rise to growths of essentially the same kinds, but most of those due to benzpyrene and methlycholanthrene remain for a long while small, dry, and indolent whereas many of the tar tumors are fleshy, vigorous, and rapidly enlarging,—differences wholly consequent on differences in the ability to promote growth. Such ability is an important element in the effectiveness of carcinogens. Tar and the polycyclic hydrocarbons cause many more cells to become tumor cells than give rise to visible growths. Benzpyrene is as effective in initiating neoplastic changes when dissolved in mineral oil as when in benzene, yet no tumors result from it until months after the benzene solution has given rise to them, the reason being that when in oil it is almost devoid of influence to encourage cell proliferation. Benzene itself has a very slight influence of the sort. Solvents may determine not only whether carcinogens initiate neoplastic change but may condition to a crucial degree the influence of these agents to encourage tumor formation. Rabbit epidermis is much more responsive to carcinogenic influences than that of the mouse, as measured in terms of time taken to elicit benign neoplasms. Even benzene will call forth these growths from rabbit skin. In appraising the relative responsiveness to carcinogens of various animal species it is essential to reckon in terms of cells of identical type.

Rabbit skin can be rendered abnormally susceptible to papilloma virus infection by preliminary treatments with a variety of agents. The most effective agents thus far found are 0.3 per cent methylcholanthrene in benzene and a mixture in equal parts of turpentine and acetone, applied four or five times at 2 day intervals. When virus is inoculated into skin altered by these agents, either intradermally or by inunction after scarification, papillomas appear earlier and in greater number than on normal skin, and much higher dilutions give rise to growths. The method provides a means of detecting amounts of virus which cause no papillomas upon inoculation into normal skin. Papilloma virus material which is rubbed into scarified normal or hyperplastic skin is largely lost in the scabbing which ensues, and nearly all of it fails to reach susceptible cells. The preparatory agents which increase the effectiveness of the virus bring about marked epidermal hyperplasia, and the hyperplastic tissue regenerates with greater rapidity when scarified. The agents evidently act in large part by providing young epidermal cells in quantity to the virus, as also by inducing a richer vascularization than ordinary in support of the papillomatous proliferation. It is possible that they also act by providing especially susceptible cells. The implications of the findings are discussed.

Continued serological investigations of the sedimentable constituents of normal and neoplastic tissues have shown that the blood serum of normal rabbits will fix complement in mixture with saline extracts of normal rabbit tissues. The phenomenon has proved referable, not to anticomplementary effects of serum or antigen nor to so called non-specific complement fixation, but to a naturally occurring serum principle, hitherto unrecognized, which reacts specifically in vitro with a sedimentable constituent of normal tissue cells. The principle exists in the blood of nearly all adult rabbits but is absent from that of rabbits less than 1 month old. It can be salted out from serum with ammonium sulfate and is destroyed when heated at 65°C. for 20 to 30 minutes. Its titer was found to run parallel in general with that of two natural antibodies also present in normal rabbit's blood (natural Wassermann reagin, natural anti-sheep hemolysin); but absorption tests showed it to be distinct from these. Because of its properties, the serum principle has been termed the natural tissue antibody. The substance with which the natural tissue antibody reacts is regularly present in saline extracts of many normal tissues,—those of rabbits and of other species as well. Kidney and liver tissues always yield it in abundance, while spleen, brain, and testicle provide somewhat less; heart and voluntary muscle extracts contain relatively little, and non-nucleated erythrocytes and skin are practically devoid of it. The results of affinity and absorption tests indicate that it is nearly or quite the same from whatever tissue or species derived. It is readily sedimentable in the high-speed centrifuge, little or none remaining in the supernatant liquid of potent suspensions spun at 25,000 R.P.M. (45,400 g) for 1 hour. It either does not come away into alcohol or is inactivated thereby, is readily destroyed by heat (56–70°C. for 30 minutes), and diminishes notably in antigenic potency upon standing overnight in saline suspension or when the tissues containing it are kept in glycerol. Its properties suggest that it may be a protein. The implications of the findings are discussed in relation to the formation of the natural antibody and its place amongst serological phenomena, to so called "non-specific" fixation of complement and other serological complexities, and with particular reference to the character of the sedimentable constituents of normal and neoplastic tissue cells.

The property of influenza virus suspensions to agglutinate red blood ceils appears to be closely associated with the infective activity of the virus (1). The complement-fixing antigen, on the other hand, has been shown to consist of two distinct fractions: one is intimately associated with the hemagglutinin and infective virus particle, while the other has a smaller size and is readily separable from the virus particle by centrifugation or adsorption with red blood ceils (2). A preliminary report (3) showed that the infectivity and hemagglutinin of the PR8 and Lee strains of influenza virus were associated with particles which were at least 60 m/z in diameter. Subsequent papers by other workers (4, 5) have confirmed this latter finding and have set the size of the PR8 virus particle at 80 m# and that of the Lee virus particle at 100 m#.The present paper presents a further study of the various properties of influenza virus and their interrelationships, by means of experiments in the angle centrifuge and the optical ultracentrifuge, and by filtration with graded collodion membranes. It will be shown that the application of a synthetic density gradient, already described (6), to influenza virus suspensions minimizes convective disturbances in the angle centrifuge and permits the demonstration of sedimenting virus boundaries which can be identified by infectivity, agglutination, and complement fixation tests. MethodsPreparation of Virus Suspension.--The PR8 strain of influenza A virus (7) and the Lee strain of influenza B virus (8) were used in these experiments. Mlantoic fluid suspensions of each virus strain were prepared by a technique already described (9). Briefly, it consisted of inoculating the allantoic sac of 11 day old white Leghorn embryos with 0.2 cc. of a 10 -s dilution of allantoic fluid known to contain virus in high titer. After incubation at 37°C. for 48 hours, the eggs were chilled overnight at about 4°C. The blood-free allantoic fluids were then removed, pooled, and cleared by centrifugation at about 1800 R.1,.~r. for 5 minutes. Without further treatment the fresh allantoie fluids were subjected to the experimental procedures to be described, and all tests were made on the same day.Infectivity Tests.--The virus suspensions were diluted in steps of 10 -°'5, and 0.05 cc. samples were instilled intranasally into albino Swiss mice under light ether anesthesia. 301

By preliminary preparation of the skin in ways that render it hyperplastic the presence of infective virus can be demonstrated in extracts of domestic rabbit papillomas which yield no growths when inoculated by the ordinary methods and which for this reason have been supposed to contain no virus. The amount of virus recovered by the method outlined in the present work, however, is small when compared with the yield obtained in most instances from comparable cottontail rabbit papillomas. The yield is greatly influenced not only by the virus strain used to produce the growths but by the individual rabbit host. Although virus has been obtained from papillomas produced in domestic rabbits by all of the virus strains tested, a total of 21 thus far, only about one-fourth of these strains are readily to be procured again from the growths they cause and the others are demonstrable only in hosts in which the conditions are favorable for reasons unknown. An experimental comparison of the capacity of suspensions of papilloma tissue from domestic and cottontail rabbits to elicit specific antibodies has shown that the titers attained are approximately proportional to the amount of infective virus demonstrable in the suspensions. The findings as a whole indicate that far less virus exists in infective or antigenic form in the papillomas of domestic rabbits than in those of cottontail rabbits.