110 Dr. J. H. Vincent on a Numerical 



Deviations from the Formula. 



The elements deviating most from the formula, judging 

 from the differences, between the computed and experimental 

 values, are : — 



w-W. 



Iodine + 3*9 



Barium + 3*2 



Cobalt -2*73 



Praseodidymium — 2*4 



Zirconium —2*37 



Calcium -2-30 



Selenium +2*19 



Molybdenum —2*1 



Neodidymium — 2*1 



Iodine is thus a very marked exception to a rule which 

 holds fairly correctly for 57 other elements. This may be 

 explained by assuming either (1) that the real atomic weight 

 of iodine is less than that accepted, or (2) that the order in 

 the table is too low ; and thus the suspicion of error in the 

 atomic weight falls upon its next highest neighbour tellurium. 

 The atomic weight of tellurium is a matter of some un- 

 certainty owing to two causes. One of these is that successive 

 attempts to determine its atomic weight do not yield con- 

 cordant results, and the other is that unless the atomic weight 

 of tellurium is below that of iodine these two elements con- 

 stitute a serious exception to the periodic law. If we assume 

 that its atomic weight is less than that of iodine, then iodine 

 would have 



n = 52. 



Its computed value would be .124*8, which differs from the 

 determined value by 1*09, a quantity little greater than the 

 mean deviation. The most probable atomic weight for 

 tellurium on this view is 122. 



The fact that the greatest variation of the rule leads to the 

 same conclusion as the periodic law may be taken as tending 

 to confirm the legitimacy of the former. No general relation 

 has been found to connect the deviations, but a term could 

 be inserted in the formula whose magnitude depended on 

 whether n were odd or even, which would make the errors 

 less. 



It may be noted that iodine, molybdenum, and selenium 

 are exceptions to Dulong and Petit's law. 



