GENETIC INTERACTIONS BETWEEN ANIMAL VIRUSES 287 



The chief interest of these results is in regard to the nature of the X x X 2 

 fluids. From the results obtained with antisera, it appears that perhaps 80 % 

 of the virus particles have a mosaic surface of both antigens. This does not 

 necessarily imply that 80 % of the viable particles have such mosaics, but 

 the general correspondence of neutralization tests in ovo suggests that 

 such is probably the case. One can see no a priori reason why viable and 

 nonviable particles should not have the same type of antigenic surface 

 structure. 



In order to determine the genetic character of these mosaic particles, 

 Gotlieb and Hirst (1954) carried out titrations, going well beyond the 50 % 

 end point and analyzed the fluids so obtained for the presence of both serolo- 

 gical types. They found a clear excess of fluids from which both M and W 

 could be obtained in titrations of X fluids, as contrasted with mixtures of M 

 and W. They believed that this was direct evidence of the existence of 

 heterozygotes and, as will be discussed later, consider it likely that all 

 recombination occurs via a heterozygote condition. 



In our own experience (Lind and Burnet, 1957a), X fluids from limit 

 dilutions are excessively rare in the systems we have used, and we have also 

 found it much more difficult to obtain evidence of heterozygosis using the 

 same methods as Hirst and Gotleib. On one occasion, a typical X fluid was 

 obtained at an ID 20 level from the primarj^ harvest of a MEL/NWSE cross. 

 This fluid A was analyzed in much detail by repeated titrations at LD by 

 several methods. Direct titration showed LD fluids either WS — (a recom- 

 binant) or MEL in the ratio 17 to 7 (titers 10 8 * 5 and 10 8 - 1 , respectively). Only 

 2 of these 24 clones showed evidence of being derived from a heterozygous 

 particle. In addition, appropriate titration on the chorioallantois, with and 

 without antiserum, gave invasive WSE and M -j- (recombinant) strains at 

 titers 10 5 - 6 and 10 4 - 5 , respectively. 



Undoubtedly heterozygotes can occur — with important implications for 

 the general understanding of virus genetics — but everything suggests that 

 heterozygotes are less viable than homozygotes and that in Hirst and 

 Gotlieb's system a much higher proportion of heterozygotes are viable for a 

 few generations at least than in ours. 



This is equally evident in their studies of recombinants from WSN (similar 

 but not identical with NWS) and MEL which, using Hirst and Gotlieb's 

 (1955) nomenclature, will be called W+ and M— . Mixed cultures give 

 harvests which on test at limit dilution give rise to three predominant types, 

 the originals W + and M — and a W — recombinant. No directly neuropatho- 

 genic M -\- was obtained but a proportion of masked neurotropic M's could 

 be detected by back-crossing with W — . The point of greatest interest is that 

 if the clones serologically W were tested (1) for heterozygosity of the initiating 

 virus by subculture in anti-W serum and (2) for -f- or — neuropathogenicity, 



