34 PROCEEDINGS OF THE AMERICAN ACADEMY 



it follows with a very high degree of probability that 



Ag : Br : Cd = 108.00 : 80.00 : 112.31, 



within one ten-thousandth of the value of either quantity. It must 

 of course be regarded as absolutely proved that, in the material 

 analyzed, the proportion of bromine to the remainder of the mass, 

 assumed to be pure cadmium, is as 2 X 80.00 to 112.31, but it is always 

 conceivable that the material used contained some unknown impurity. 

 In the present case, however, such an assumption is highly improb- 

 able ; first, because such extreme care was taken in the preparation ; 

 secondly, because the products of the diiferent sublimations were so 

 constant in composition ; and, thirdly, because the presence of any of 

 the metals usually associated with cadmium would tend to lower and 

 not to raise the apparent atomic weight. If any impurity is present, it 

 wouM seem as if it must be some unknown element, and only on such 

 a bare chance as this can the evidence be invalidated which the 

 results of this investigation furnish against the hypothesis of Prout. 



