214 



Dr. F. Horton on the 



It may be mentioned that some slight trouble was ex- 

 perienced in adjusting the positions of the solenoids C^ and 

 C 2 . In the new thermostat, where the solenoids have been 

 placed co-axially with the screws B x and B 2 , no difficulty 

 has arisen. 



Another slight difficulty that should be stated is that to 

 get the best conditions of control it is necessary that the air 

 in the room should be rapidly stirred by means of a fan ; this, 

 however, causes air currents to impinge upon the sensitive 

 strip, thus making it vibrate. Wire gauze round, and a dash- 

 pot at the end of*, the strip (which is hung vertically), 

 prevent the trouble. Neither the wire gauze nor the dash- 

 pot is shown on the drawing. 



In the course of the experiments with the interferometer 

 it was found that, between the observations, while the ob- 

 server was out of the room, it was advantageous to replace 

 him by a heated dummy, thus keeping the conditions as 

 constant as possible. 



The following results give the temperatures of the air inside 

 the interferometer case. During the hour from 10.5 to 

 11.5 A.M. the observer replaced the dummy, and was seated 

 within 1 foot of the case : — 



Time. Temp. 



9.0 a.m. 1915 C. 



9.31 19-15+ 



10.5 19-16- 



10.20 19-16+ 



Time. Temp. 



10.20 a.m. 19°16 + 



10.35 19-16- 



10.50 19-16 



11.5 19-16 



Time. Temp. 



11.20 a.m. 19°16+ 



11.35 19-16+ 



11.50 19-16+ 



12.15 1917 



XX. On the Origin of Spectra. By Fkakk Horton, D.Sc, 

 M.jL, Fellow of St. John's College, Cambridge^. 



IN a paper recently communicated to the Royal Society t, 

 the author has given an account of some experiments 

 with mercury vacuum-tubes which show that mercury is 

 capable of giving several distinct line spectra when rendered 

 luminous by an electrical discharge. The particular spectrum 

 appearing in any given case depends upon the energy of the 

 discharge in relation to the mass of vapour through which it 

 passes. The spectrum most easily produced consists of five 

 bright lines, but as the energy of the discharge is increased 



* Possibly improved results could be obtained if the contact points 

 were placed under oil. 



f Communicated by the Author. 

 X Roy. Soc. Proc. 1911. 



