THE ARGON FAMILY OF GASES 655 



convincingly that a substance is an element, and not merely 

 an undecomposed compound. In neither case is any single 

 proof absolutely satisfying. In both cases a review must be 

 made of all the available evidence. I shall use the word mole- 

 cule generally to indicate a single particle, whatever its com- 

 plexity, whether mon-atomic or not, which is capable of existing 

 an appreciable time, long enough to be studied, and the word atom 

 only for the smallest particle that, as in a chemical change, has 

 an existence so momentary that it cannot be studied alone. 

 Thus the hydrogen atom, H, cannot yet, under any known 

 circumstances, properly be termed a molecule, though no 

 doubt extension of our means of high temperature research 

 would result in its becoming experimentally known as is 

 the case for the single iodine or bromine atoms which at 

 high temperatures exist as molecules. But the hydrogen ion 

 (H + ), the a-particle (He ++ ), the various positive-ray ions studied 

 by Sir J. J. Thomson recently, are molecules, in the strict sense 

 of the definition, and whether they are also single atoms or not, 

 need not be prejudged by the name by which they are called. 

 They exist, unchanged in mass, for periods long enough to 

 enable their mass to be determined. I shall treat it as beyond 

 dispute, that from Avogadro's Law, the molecular weight of a 

 gas in terms of hydrogen is identical with that of its density in 

 terms of the same unit, and shall follow the usual convention that 

 the unit of gaseous density is H = 1, and the unit of molecular 

 weight H 2 = 2. The molecular weight of helium is thus 4, and 

 that of the radium emanation, from diffusion, effusion, and direct 

 density determinations by the micro-balance, is 222 (± say 

 10 per cent, at most). The molecular weight of the a-particle 

 is experimentally found by comparing the value of the ratio of 

 its mass to its charge, m/e, as determined by electrostatic and 

 electromagnetic deviation, with that of the hydrogen ion, and 

 by comparing the charge on the single a-particle (which it is 

 possible to determine experimentally since the number of 

 a-particles can be counted and their total charge measured) with 

 that on the hydrogen ion. As Perrin has shown the value for 

 the single atomic charge carried by the hydrogen ion — or 

 what is the same thing, experimentally, the determination of 

 Avogadro's constant, the number of molecules in a cubic 

 centimetre of gas at N.T.P. — can be determined by at least 

 nine independent methods, with results in agreement far closer 



