397 



rate for BO (50 n«) is 805S, I. e. only 15% of tho animals escaped bcccaln g 

 tanourlng-bearlns during the remainder of xhe experiment. Ihls 15^ represents 

 the majdnium amount by which any increase of the dose level above 50 ag 

 no matter how great, can increase the steuTdardlsed response. It is no 

 surprise that under these circumstances an increase to 75 mg ca n on ly 

 produce a rise to 84Jb in t he starriardlsed rat e; this is about 25% of the 

 maxlmuin possible increase. Oius it is unrealistic to hope that when the 

 50 mg dose is producing a 7O-8056 response, the 75 mg dose will produce 

 an effect that is proportionally greater. To do so in BO, for example, 

 the standardised response at 75 mg would have to be 9<J6, i.e. virtually 

 every animal after about 28 weeks must become tu.-aour bearing; the 75 ng 

 dose must absorb 100^ of the available increase. And of course, this 

 means that a further increase in dose level from 75 mg to 100 mg could 

 produce no increase in the standardised rate at all; an equally anooalous 

 situation. 



Bie remedy for these high dose anocnalles In the standardised experiments 

 Is now obvious; the dose levels must be reduced. Ihe problem is to estimate 

 the magnitude of this reduction. If the dose levels are too low, the 

 tumour rates at the lowest doses will be so low as to be subject to great 

 uncertainty, and requiring many animals to produce reasonable confidence 

 levels. Hie consequences of the dose levels being too high are already 

 quite apparent. The ideal situation is one where we have a linear dose- 

 response relationship coupled with a reasonably high tumour yield even 

 at the lowest dose levels. 



OB 



e 

 c 



IS) 



I 



O 

 >-• 

 ^^ 



o 



