xvni THE CLASSIFICATION OF THE ELEMENTS 497 



of octaves. In this arrangement hydrogen was i and gold was 

 49, but in 1878 a more open system was used, in which gold 

 was 86. Moseley in 1913-1914 showed that integral atomic 

 numbers could be determined experimentally by studying the 

 high-frequency spectra emitted by different elements when used 

 as targets in an X-ray bulb. In this system, if aluminium be 

 taken as 13, then gold is 79, and there are only four vacant 

 places between these two elements after filling in all elements 

 of known atomic weight. The total number of missing ele- 

 ments between hydrogen and uranium appears to be nine, 

 namely, numbers 43, 61, 72, 75, 84, 85, 87, 89, and 91. The 

 atomic number for nickel was found to be one unit greater 

 than for cobalt, in spite of its smaller atomic weight, thus 

 justifying MendeleefTs view that nickel should be placed 

 between cobalt and copper ; in the same way the atomic 

 numbers for chlorine and for potassium were found to be 

 separated by two units, leaving a place between them for 

 argon, the atomic weight of which is greater than that of 

 potassium. Atomic numbers may also be deduced from the 

 absorption of X-rays by solid elements and the scattering of 

 X-rays by gases. These atomic numbers satisfy all the require- 

 ments of Prout's hypothesis, but the irregularities shown by the 

 atomic weights are still unexplained. 



X K 



