IS6 POPULAR SCIENCE MONTHLY. 



wonderful closeness to exact multiples. This is shown by the following 

 table, taken from the report of the Committee on Atomic Weights of 

 the American Chemical Society for 1899: 



Arsenic 75.0 Lead 206.92 Phosphorus 31.0 



Boron 11.0 Lithium 7-03 Rhodium 103.0 



Bromin 79.95 Manganese 55.0 Silver 107.92 



Carbon 12.0 Mercury 200.0 Sodium 23.05 



Cerium 139.0 Nitrogen 14.04 Sulfur.... 32.07 



Cobalt 59.00 Osmium 191.0 Tin 119.0 



Gallium 70.0 Oxygen 16.00 Yttrium 89.0 



Iron 56.0 Palladium 107.0 



In addition to these at least nine others have atomic weights differ- 

 ing not more than 0.1 from whole numbers. By the law of probabilities 

 this close approach to whole numbers cannot be the result of chance, 

 but no satisfactory explanation has as yet been offered. 



Prout's hypothesis was not unique in concerning itself with an 

 effort to show a unity of matter. Very early there was noticed a con- 

 nection between the atomic weights and the properties of certain 

 groups of elements. Attention was first called to this by Professor 

 Dobereiner, of Jena, and an account was given of it in print in 1816, 

 just after the first enunciation of Prout's hypothesis. Dobereiner 

 noticed that the equivalent weight of strontium was 50, while the 

 values then accepted for calcium and barium were respectively 27.5 and 

 72.5. Fifty is the mean of 27.5 and 72.5, and the properties of stron- 

 tium may be looked upon as being an average of those of calcium and 

 barium. It was soon clear, however, that strontium was as much en- 

 titled to recognition as an element as calcium or barium. Hence it 

 appeared that there was a numerical relation between the weights of 

 the atoms of these three elements, barium, strontium and calcium, 

 which corresponded to both the chemical and the physical properties 

 of the elements. Several other similar groups of three were discovered 

 by Dobereiner, and this, which was really the earliest germ of the 

 Periodic Law, became known as Dobereiner's law of triads. It is of 

 especial interest, as having enabled its author to predict the atomic 

 weight of bromin, which was later confirmed by experimental investiga- 

 tion. In this respect it anticipated the Periodic Law, and may be said 

 to represent a phase of this later and greater generalization. 



Little attention was attracted by the speculations of Dobereiner, 

 and a quarter of a century or so later the subject was taken up anew 

 by the great French chemist, Dumas. He developed to some extent 

 the law of triads, though he made little actual advance beyond the 

 point attained by Dobereiner. Dumas's work was, however, widely 

 noticed, and proved very stimulating to the chemists of his day. It 



