264 Professor Sir J. J. Thomson [April 3, 



as the film is not transparent the measurement of the parabolas 

 offers great difficulties. I have got the best results with some plates 

 brought out by Dr. Schumann for studying the " Schumann rays." 

 which are light rays far in the ultra-violet, and so easily absorbed 

 that they are stopped by a layer of air only 1 mm. thick. The 

 instructions drawn up by Schumann for preparing these plates are 

 given in Baily's " Spectroscopy" (2nd Edition), and following these one 

 of my assistants has had considerable success in preparing these 

 plates. I may say that the more closely we followed Schumann's 

 instructions the better were the results ; all the variations we intro- 

 duced were failures. The films have their peculiarities ; they are no 

 good for some time after they have been made, they improve on 

 maturing, and may turn out good or bad. AVhen they are good they 

 make a wonderful diff'erence in tbe experiments, for an exposure of a 

 minute or two will give as good a picture as one of an hour with 

 some of the older plates. The saving of time is important not only 

 for itself, but because it is easier to keep the experimental conditions 

 constant for short exposures than for long ones, and thus to get 

 sharper lines. The plates after exposure show briUian^ colours due to 

 the reduced silver. A slide of a photograph of the positive rays through 

 a mixture of chloroform and chlorine was shown showing 21 parabolas 

 clearly defined on the plate ; these included parabolas due mercury 

 (200)5 CCI4 (152) CHCI3. mercury with two charges, Ctl.CU, C% 

 a line belong to a substance with a molecular weight 60 not identified 

 CH3CI, COo, a line corresponding to molecuhir weight 40, chlorine, 

 CO, chlorine with two charges, etc., lines corresponding to molecular 

 weights I'J and 17, oxygen, nitrogen, carbon. 



The choice of the developer is a matter of considerable importance ; 

 the plate just shov^n was developed with pyro. methol. A plate taken 

 under similar conditions, but developed with hydroquinone, will be 

 thrown on the screen, and you will see that the number of lines 

 is much less and the definition very infe^'ior to that developed with 

 pyro. methol. As examples of the results obtained by Schumann 

 plates I will show shdes showing the positive ray spectra of HCl, 

 SnCl^, and one of hexam taken by my son, Mr. G. P. Thomson, and 

 remarkable as showing on the negative side of the T^hotograph CH, 

 OH, C^, and Cg ; hitherto only atoms had been observed on the 

 negative side, tmd not molecules, whether of elements or compounds. 

 I will show later on another example of the occurrence of a compound 

 molecule PH3 on the negative side of the plate, indicating that the 

 molecule of this compound can carry a charge of negative electricity. 



I will now pass on to consider some applications of the method, 

 and will take first the question of the origin of the line corresponding 

 to the atomic weight 3 : this must be either a new element or a 

 polymer of hydrogen analogous to ozone. This line occurs when the 

 gas which is the sul)ject of the positive ray analysis is that which 

 comes off when metals and other substances are exposed to bombard- 



