32 CLARKE— STATISTICAL METHOD [April i8 



and I have indicated some of the desiderata. From chemical evi- 

 dence we can draw deductions relative to the volumes and masses 

 of the sediments, and also gain something towards the measure- 

 ment of geological time. Even if the conclusions to be reached by 

 these methods are not final, they are at least helpful. In order to 

 be convincing, the chemical evidence must follow lines convergent 

 with other testimony, and when it does so we may be satisfied that 

 we are approaching the truth. 



