DISCUSSION 23 



had been centrifuged (not at very high speeds, but so that it was free of 

 crude material not sohibilized by deoxycholate or Triton). We have 

 also tried Tween 80 and this did not solubilize nor did it greatly impair 

 sensitizing activity. Therefore with these three surface-active agents 

 inactivation and solubihzation went hand in hand. Although I fmd it 

 fundamentally objectionable to inject antigenic preparations containing 

 these surface-active substances, yet when we tried your method of 

 removing the Triton with acetone, reprecipitating the antigen, the 

 preparation was totally inactive in our system. 



HiUemann : I would like to raise two questions relative to your fmding 

 that soluble extracts would not sensitize by the intravenous route. 

 Will such extracts under any conditions lead to prolonged homograft 

 survival, as you supposed at one point in your presentation ? Secondly, 

 will such intravenously administered extracts induce humoral anti- 

 bodies and, if so, how does the titre compare with the antibodies 

 induced by, say, the intradermal route, which gives you effective 

 transplantation immunity as well ? 



Medawar: In answer to your first question: we have isolated obser- 

 vations of a prolongation of the Ufe of homografts by injecting these 

 materials intravenously; but I emphasize that these are just isolated 

 observations. 



As to whether these extracts provoke humoral antibodies when in- 

 jected intravenously: we have injected only four mice repeatedly with 

 these semi-soluble extracts intravenously. Three out of four of them 

 produced no haemagglutinins and no haemolysins and one of them 

 produced a moderate titre of both — so we don't know what to make 

 of it. 



Amos: Haemagglutination is a legitimate indicator in the mouse as 

 long as you are certain you are dealing with H-2, but one must bear in 

 mind that there are a number of haemagglutinating systems not 

 related to H-2 (we have already identified about seven of these) and 

 there are other cases where the antigen is not on the red cell (H-i and 

 H-3 are two good examples of this). If you now want a quantitative 

 estimate of the antibody response you must have some idea whether 

 there is a contamination with some of these other antibodies; this 

 should not be difficult to determine using back-cross animals as in- 

 dicators. Silvers' suggestion of using the co-isogenic lines is an excellent 



