68 DISCUSSION 



and still get a positive control. Our best materials run down to 2 or 



3 t^g. 



Brent: It would be interesting to know whether the production of 

 these antigens is specific to tumour cells or whether they also occur in 

 normal tissue cells. One way of doing this would be to examine peri- 

 toneal exudate cells of normal mice for the presence of the antigens 

 which you extract. 



Davies: We use tumour cells purely for convenience, because we can 

 get so much antigen ; we can get a kilogram dry weight of cells from 

 5000 mice, which makes a lot of things practicable which are otherwise 

 impracticable. I have no doubt that one could get the material from 

 other cells; it is present, for example, in "L" cells grown in tissue 

 culture. We have tried to get ascitic fluid in other ways, by some non- 

 specific irritation, but not with a great deal of success; though I would 

 not expect to find activity in such ascitic fluid. We have taken normal 

 mouse serum through the same procedure, and again one gets an in- 

 soluble fraction in the same position in the table, but this of course is 

 quite inactive. 



Woodruff: If it were considered safe to see if you could demonstrate 

 antigens in human beings (and I think there are circumstances in which 

 you could do this) it would be interesting to compare ascitic fluid in 

 people with disseminated tumours throughout the peritoneal cavity and 

 ascitic fluid in conditions like cirrhosis of the liver. 



Davies: I don't feel we've got far enough to embark on such a project ; 

 we would like to learn more about this mouse material first. However, 

 if one found the analogous human product, just by looking for the 

 same substance isolated in the same way, the specificity of this might 

 give a clue as to what the transplantation antigen is in man. 



Woodruff: Can you produce non-neoplastic ascites in mice by making 

 them cirrhotic, for example by giving them alcohol ? 



Davies: A number of papers describe the production of non-specific, 

 non-tumour-induced ascitic fluid in mice, but we have had much less 

 success than the authors of these papers have claimed — an occasional 

 mouse produces something, but we haven't had enough to put it 

 through the mill. 



G. Klein: Perhaps the best way to produce a comparable ascites of the 

 desired quality in large quantities would be to inoculate an ascites 



