June i, 191 i] 



NATURE 



467 



cent, the atomic weight of a gaseous substance, without 

 requiring more than i/ioo milligram of the substance. 

 Another very important advantage of this method is that 

 it is not dependent upon the purity of the material ; if 

 the material is impure, the impurities merely appear as 

 additional lines in the spectrum, and do not affect the 

 parabola due to the substance under examination, and 

 therefore produce no error in the determination of the 

 atomic weight. The method would seem to be peculiarly 

 suitable for the determination of the atomic weights, not 



Fig. I. 



merely of the emanation from radio-active substances, but 

 also those of the products into which they disintegrate. 



The rays, too, are registered within less than a millionth 

 of a second after their formation, so that when chemical 

 combination or decomposition is occurring in the tube, 

 the method may disclose the existence of intermediate 

 forms which have only a transient existence, as well as of 

 the final product, and may thus enable us to gain a clearer 

 insight into the process of chemical combination. 



I will now show a few slides prepared from the photo- 

 graphs we have taken of the positive-ray spectra. The 

 first (Fig. i) is that of nitrogen prepared from air ; the 

 measurements of the photograph showed that the atomic 

 weights of the carrier producing these curves were as 

 follows : — 



Positive Negative 



I H+ I H_ 



1*99 H.,+ 1 1 20 C_ 



6-8o N++ 15-2 0_ 



11*40 C+ 



ir95 N+ 

 28-1 N2+ 

 39 A.g+ 

 100 Hg++ 

 198 Hg+ 

 The symbol H + denotes that the carrier is an atom of 

 hydrogen with one charge; Hj+ that it is a molecule of 

 hydrogen with one charge; N^.+ that it is an atom of 

 nitrogen with two charges ; and so on. 



I 



Fir.. 2. 



With nitrogen from NH^NO, the lines were as follows 

 (the magnetic force was so large that some of the lines 

 corresponding to the lighter particles were thrown of! the 

 plate) : — 



61 C+_f. 442 CO0+ 



7-02 N++ 65-5 Hg+++? 



i2-o8 C4. 100 Hg++ 



14-01 N+ 204 Hg4. 



27 '9 N.J+ 

 NO. 2170. VOL. 861 



The next slide (Fig. 2) is the positive-ray spectrum for 

 CO, and again the magnetic field is so great that the 

 lighter carriers do not appear. 



From the measurement of the lines we find that the 

 atomic weight of the carrier is 



Positive Negative 



6-OOC++ 12 C_ 



6-95 N++ 16 0+ 



7950++ 



1202 C+ 



13-9 N + 



1595 0+ 



28 05 CO+ 



43 CO.,+ 



69-5 Hg+++(?) very faint 

 100 Hk++ 

 202 Hg+ 



for CO, is represented in Fig. 3 ; the 



43-9 CO2+ 



62 5 Hg+++(?) very faint 

 996 Hg++ 

 200 -o Hg+ 



The spectrum 

 atomic weights are : — 



5-98 C++ 

 8 00 0++ 



I2'00 C+ 



i6-oo 0+ 

 2802 CO+ 



The spectrum of CHj, of which a small region with five 

 lines close together is shown in Fig. 4. This is interest- 

 ing, because the measurement of these lines shows that 

 their atomic weights are 12, 13, 14, and 15, 16, and thus 

 that we have here C, CH, CH^, CH„ CH^. If I am not 

 mistaken, this is the first occasion when the atoms CH, 

 CH,, CH, have been observed in a free state. 



The spectrum of the analogous compound chloroform, 

 CHCI3, is represented in Fig. 5. The atomic weights 



represented in this are : — 



1 Hh. 



I -Si?) 



2 H2+ 



3 (?) 



6 C++ 

 8 0++ 

 1 1 -9 C+ 

 137 N+ 

 16 0+ 



18-50++ 

 277 CO+ 

 36 C1+ 

 46-5 CCI+ 

 63 (?) faint 

 81 CCI, 



102 ng++ 



201 Hg+ 



The carriers with atomic weights 1-5 and 3 have not 

 been identified. They are of frequent occurrence. I have 

 here two slides, one of SiH« (Fig. 6) and the other of the 

 residual gas in the tube, in which they are well marked, 

 though at their best they are only faint lines. In Fig. 7 

 we have the positive ray spectrum of air, taken under 

 conditions which produce very narrow lines, which can be 

 accurately measured. 



Let us now consider some of the results brought to 

 light by these photographs. In the first place, they show 

 that a gas through which an electric discharge is passing 

 is a much more complex thing than a collection of mole- 

 cules all equal to each other. Even an elementary gas 

 becomes in these circumstances a mixture of a great many 

 different substances. Thus, to take oxygen as an example, 

 the photographs show that when a current of electricity 

 passes through it, we may have present simultaneously 

 o-xygen in the following states : — 



