Febedaey 24, 1899.] 



SCIENCE. 



275 



gave off generally also a trace of argon, as 

 was later shown by our diffusion experi- 

 ments. Naturally, it was impossible to be 

 certain that the few cubic centimeters of 

 gas which we collected from these miner- 

 als contained no new gas, but we failed to 

 detect the presence of any new lines with 

 the spectroscope. 



You will, undoubtedly, recall that, soon 

 after the discovery of helium, doubts were 

 expressed in many quarters as to whether 

 the gas was really uniform or a mixture. 

 In order to dispel these doubts and also to 

 search for the missing gas, Dr. Collie and I 

 carried out a long series of diffusion experi- 

 ments. Through these we reached the con- 

 clusion that it was, in fact, possible to sepa- 

 rate helium into two constituents, one of 

 which possessed a somewhat higher density 

 than the other. Later experiments, how- 

 ever, in conjunction with Dr. Travers, 

 showed that this conclusion was erroneous. 

 In this second series much larger quantities 

 of helium were at our disposal, and, to our 

 disappointment, we found that the heavier 

 fractions of our gas owed their greater 

 density to the presence of a trace of argon. 

 Here, again, we were unable to find any new 

 line in the spectrum, and thus far our search 

 was fruitless. 



We next directed our attention to mete- 

 orites and to mineral waters. Only one out 

 of seven meteorites examined by Dr. Trav_ 

 ers and myself showed the presence of 

 helium and with it a trace of argon ; the 

 others gave only hydrogen and hydrocar- 

 bons, which were also present in the gases 

 from the meteorite M'hich contained helium 

 and argon. Here, again, our search was in 

 vain. The mineral water from Bath has 

 been investigated by Lord Rayleigh ; in the 

 waters from Cantarets, in the Pyrenees, Dr. 

 Schlosing has found both argon and helium. 

 Dr. Travers and I examined these gases for 

 new lines, but, as before, none were found. 



Our patience was now well-nigh ex. 



hausted. There seemed, however, to be a 

 single ray of hope left, in an observation 

 which had been made by Dr. Collie and 

 myself. You will recall that the atomic 

 weight of argon was apparently too high ; 

 at all events, it would be more in harmony 

 with the periodic law if the density of argon 

 were 19 instead of 20, and hence its atomic 

 weight 38 instead of 40. Hence, after some 

 fruitless attempts to separate argon into 

 more than one constituent by means of so. 

 lution in water, we undertook a systematic 

 diffusion of argon. We did not, however, 

 carry this procedure very far, for, at that 

 time, we believed that helium was a more 

 probable source of the desired gas ; never- 

 theless, we found a slight difference in den- 

 sity between the gas which diffused first 

 and that which remained undiffused. We, 

 therefore, decided to prepare a large quan- 

 tity of argon, and, after liquefying it, to in- 

 vestigate carefully the different fractions on 

 distillation. 



Such an operation demands much time. 

 In the first place, the necessary apparatus 

 is not to be found in any ordinary chemical 

 laboratory ; the preparation cannot be car- 

 ried out in glass tubes in an ordinary fur- 

 nace, but requires iron tubes of large size 

 and an especial furnace ; in the second place, 

 the operation must be repeated several 

 times, for it is not convenient to work with 

 an excessively large quantity of magnesium. 

 As before, we removed the oxygen from the 

 air by means of copper at a red heat ; the 

 atmospheric nitrogen remaining was col- 

 lected in a large gasometer holding about 

 200 liters ; after drying over concentrated 

 sulfuric acid and phosphorus pentoxid, the 

 gas was passed through an iron tube of 

 5 cm. diameter, filled with magnesium fil- 

 ings ; the gas was then passed through a 

 second copper oxid tube to remove the hy- 

 drogen ; it then entered a gal vani zed-iron 

 gasometer, which was constructed like an 

 ordinary illuminating-gas gasometer, in or- 



