396 



The key to understaivllng the reasons for s-ich things as the 

 negative dose-respcmse relationship In the obsei-ved response rates 

 In B6 Is to be found In Table 3 and Graph 3. where Is shc*fn plotted the 

 number of animals dying before the appearance of the first tunour in 

 each of the experiments. Now two distinct patterns of events can be 

 seen. Experiments B2, BJ and B/i show little change in the number of 

 deaths prior to first tumour as the dose level Increases; on the other 

 hand in experiments BO, B6, and B7, the deaths Increase sharply with 

 dose level. Comparison of Graphs 2 and 3 shows that It Is the experiDents 



m 



With this sharp increase of early deaths with dose rate (BO, B6, B?) 

 which exhibit the most anomalous behaviour. 



This high initial mortality Is (usually) due to the toxicity of 

 the codensate, and It has the effect of reducing the number of animals 

 at risk of producing a tuoxjur; If the increase of death rate with dose 

 level Is high enough, the number of animals producing tumours will 

 decrease with dose level, rather than increase. This is whet happened 

 in B6. Age standardisation is aimed at reaoving effects such as these, 

 and generally speaking it does so quite well; the anomalous effects 

 that remain in the standardised data arc not primariljr due to toxicity* 

 effects. 



Cigarette smoke condensate does not produce instant tunours. The 

 earliest tumours produced at Battelle have occurred about the 20th week 

 of an experiment, and usually the time is about 28 to 32 weeks. Now consider 

 BO. In the age standardised experiment 95$^^ of the animals survive to 

 about the 28th week, so that even if every animal then became tumour bearing, 

 th» standardised response rate would be 931^« I n fact, the age stajxiardised 



BW-\i(2-01105 



