328 Prof. W. D. Harkins on the 



oxygen basis from a whole number which is more than two 

 or three hundredths of a unit greater than the experimental 

 error, may be considered as a basis for the prediction 

 of isotopes, particularly among the light atoms, provided 

 (a) that the whole number from which the deviation occurs 

 is that specified by the Harkins- Wilson equation 



W = 2(M+/) = 2M + n, 



where the probable value of / (or n) is determined from the 

 adjacent elements and from the general stability hyperbola 

 represented later in fig. 5 (PI. XII.); and (b) that other 

 complication factors are absent. 



2. No isotope for which the value of N/P is less than 0'5 

 should be predicted. 



3. The probability ol the existence of a detectable 

 quantity of an isotope is small if the value of N/P for the 

 species is abnormally high for its particular value of M. 

 Thus the atomic weight of lithium, 6*94, indicates the 

 existence of an isotope of atomic weight equal to either 6 or 

 5 } but the value 5 is very improbable since its N/P value is 

 less than 0'5. On the other hand this does not exclude the 

 existence of an isotope of atomic weight 8, as predicted by 

 Rutherford. However, the value of N/P for this latter 

 isotope would be 0625, which is extremely high. This does 

 not make its existence impossible, but makes it probable 

 that if it exists at all, it does so in only very small 

 quantities. 



4. In the region of abundant isotopes for elements of even 

 atomic number, quite definite predictions may be made in 

 many cases for elements of odd atomic number, since the 

 latter are not so numerous. In fact, if the chemical atomic 

 weights in this region were known with considerable pre- 

 cision, practically all of the more abundant isotopes of this 

 class could be predicted with few mistakes. If the precise 

 chemical atomic weight of the element of odd atomic 

 number is equal to 4^ + 3, where q is a whole number, then 

 the element should consist of only this one pure species. 

 This is illustrated by arsenic (at. wt. = 74*96), and iodine 

 (126*92), both of which meet this condition within the limits 

 of experimental error. Each of these should consist essen- 

 tially of one pure species, while bromine (at. wt. = 79*92) 

 should, on the basis of the principles given, have atomic 

 weights of 79 and 81, which is just what has been found by 

 Aston. Just what isotopes of odd atomic number and even 

 atomic weight will be found cannot be predicted, but they 

 wall be few in number and of low abundance. 



.5. In the case of elements of even atomic number in this 



