ARTICLES 217 



appears to be a diatomic carbon molecule, and going on with 

 successive additions of hydrogen atoms. This group contains 

 the very strong reference line 28 due to C2H4 or CO. The next 

 line is 32, the oxygen molecule, usually faint ; and finally we 

 have 44, due to CO2. The Ci group is well illustrated in Spec- 

 trum V, and the Cg group in Spectra II and III. In calcu- 

 lating masses, the " Oxygen " scale is used — that is to say, the 

 oxygen atom is taken as of mass 16 units. We have no evi- 

 dence whatever of the presence of isotopes in oxygen or carbon ; 

 indeed, from the extremely close whole-number relation of 

 their chemical atomic weights, none was expected. Hence 

 carbon may safely be taken as 12-00 to the accuracy of experi- 

 ment, giving us at once a direct scale of reference for masses 

 up to 44. 



Lines due to Multiply Charged Particles. — ^As has been pointed 

 out, what one really measures is not mass, but the ratio of mass 

 to charge. When the latter is the unit e itself (the funda- 

 mental unit of negative electricity : 477 10"^'' C.G.S.), the 

 lines give a measure of the masses themselves, and are called 

 lines of the first order. If, however, a particle carries a charge 

 2e, it will behave exactly as if it had half its normal mass ; with 

 a charge 3^, as if it had one-third its normal mass, and so on. 

 Lines formed in this way are called lines of the second, third, 

 and higher orders. Positive rays carrying multiple charges 

 have been fully investigated by Sir J. J. Thomson. The atoms 

 of most of the elements employed, with the exception of 

 hydrogen, seem capable of carrying two charges ; some, such as 

 argon, krypton, and xenon, appear with three or more ; while 

 mercury can carry no less than eight. First order lines are 

 always brighter, generally much brighter, than the corre- 

 sponding ones of the second order ; the second than the third, 

 and so on in decreasing strength. 



The mechanical details of the actual apparatus used to 

 obtain the following results have been given elsewhere {Phil. 

 Mag., vol. xxxix. May 1920, p. 611). It was somewhat diffi- 

 cult to construct and adjust, but even before the most advan- 

 tageous position of the plate had been fixed, it was realised 

 that the resolution was ample to settle the question of neon. 

 The results obtained with this and a few of the other more 

 important elements will now be considered in order. 



Neon {At. Wt. 20-20). — On introducing a little of this gas 

 into the mixture of CO2 and CH4 used for calibration, the new 

 lines expected were at once recognised. These were first order 

 lines at 20 and 22, and second order lines at 10 and 11. All 

 four of these were well placed for comparison, the first two 

 between the Ci and Cg groups, the second two between the 

 Ci group and the second order oxygen line 8. Several sets of 



