242 Mr. G. P. Thomson on the 



on which they showed in a darkened room as bright para- 

 bolic arcs, one for each A^alue of elm. The method of 

 experiment was so to adjust the conditions in the discharge 

 tube that one of the two parabolas, corresponding to the 

 atoms and molecules respectively, was as bright as possible 

 in comparison with the other — then to take a photograph of 

 the spectrum of the rays in the observation-chamber. 



In order to secure the purity of the hydrogen, a slow 

 steam of gas was passed into the discharge tube through a 

 capillary during the whole of the exposure, and exhausted 

 at the other end by a Gaede pump ; to prevent mercury 

 vapour from the pump diffusing back into the apparatus, a 

 mercury trap was fitted, as shown in fig. 1, and kept cool 

 with liquid air. It was found necessary after any alteration 

 which involved admitting air into the apparatus, to pass a 

 stream of gas in this manner, with the discharge passing, 

 for some hours before the impurities in the tube were 

 driven off — a result shown by the change in colour of the 

 positive rays from a greyish hue to a deep pink. The 

 hydrogen used was obtained by the electrolysis of barium 

 hydroxide and kept in a glass container dried with phosphorus 

 pentoxide. 



The discharge was produced bv means of an induction 

 coil, and the voltage, as measured by the equivalent spark- 

 gap, was in the neighbourhood of 15,000 volts. The 

 exposures used were 2f-5-| hrs. ; the plates Imperial Flash- 

 light. These are not sensitive beyond the blue, but it was 

 found that panchromatic plates would have required too 

 long an exposure. 



3. Results. 



It was found that the nature of the positive rays was very 

 sensitive to slight changes in the form of the discharge 

 tube, and in the position of the cathodes in the neck of the 

 tube. In addition, thern was a well-marked pressure-varia- 

 tion, the proportion of atomic to molecular lays being 

 markedly greater at the higher pressures. 



Atoms. 



The highest proportion of atoms was obtained with a tube 

 of the dimensions shown in fig. 2. With this tube the 

 molecular parabola was quite invisible at the higher pres- 

 sures (dark space less than half across the tube). A 

 photograph taken under these conditions is shown in PL IV. 

 No. 1. It will be seen that the only lines visible are two of the 



