20 THE NEW KNOWLEDGE. 



weights were curiously approximate to whole numbers, they 

 were with positive certainty not whole numbers. The 

 atomic weight of oxygen, for example, is certainly ^nm 16, 

 but seems to be established as 15.879. 



The decided approximation, however, referred to above, 

 was certainly curious, and it soon became most conspic- 

 uously so; for it turned out that by taking the atomic 

 weight of oxygen as 16 and calculating the weights of the 

 other atoms to that basis, a most remarkably large number 

 of them became actually whole numbers, and another large 

 number very approximately whole numbers. The table of 

 the atoms .arranged ; above is arranged on the basis of 

 oxygen =16, which for certain practical considerations is 

 the favorite arrangement among chemists; and this ex- 

 plains why the atomic weight of hydrogen =1.008 instead 

 of 1, as we should expect. yJi+Y] 



If you care to look at this table, you will find that no 

 less than 21 (marked in capital letters) out of the number 

 are whole numbers to the second decimal place; that 17 (in 

 italics) are remarkably approximately so; and that the re- 

 mainder drift along anywhere. Now, that 21 out of the 78 

 are actually whole numbers, cannot be accounted for as 

 chance on the basis of the law of probability, nor can the 

 wonderful approximation of the other 28. 



Whatever the explanation of this curious approximation 

 to symmetry, there is reason enough at any rate to suspect 

 strongly that there is some hidden relation here. There is 

 enough to whet our curiosity and to lead us on. 



TRIADS OF DOBEREINER. 



Another curious and interesting discovery soon came to 

 light. It became apparent that out of the list of elements 

 little groups could be chosen here and there with strongly 



