given by Carbon and some of its Compounds. 121 



is possible to secure that condition. The expense o£ this 

 investigation was defrayed by a Government grant from 

 the Royal Society; and the experiments were (partly) m;ide 

 at the Davy-Faraday Research Laboratory. I am indebted 

 for specially purified materials to Sir Edward Thorpe, for 

 heptane from Pinus Sabiniana, and to Prof. Sydney Young 

 for normal hexane, normal octane, carbon tetrachloride, and 

 chlorobenzene. 



In order to avoid oxygen contamination, as far as possible, 

 a discharge-vessel made of fluor-spar was employed. This 

 consisted of a rectangular block of transparent colourless 

 fluor-spar, with a cylindrical cavity, the bottom of which 

 was closed by a copper electrode ground into the spar. The 

 upper electrode consisted of a platinum wire. The fluor-spar 

 vessel was contained in a large glass test-tube, the mouth of 

 which was closed by an indiarubber stopper through which 

 the upper electrode passed, and which was also connected to 

 an air-pump. A platinum wire sealed into the bottom of the 

 glass tube made contact with the lower copper electrode. 



The liquid under examination was distilled over phosphoric 

 anhydride into the fluor-spar vessel, and the tube was kept 

 exhausted. The upper electrode was heated red-hot within 

 the liquid by causing the two electrodes to touch for a short 

 time, and the spectrum was observed with a minute spark 

 within the liquid which presently boiled. On account of the 

 rapid separation of carbon, only a few seconds were available 

 for observation, and consequently it was impossible to photo- 

 graph the spectrum. 



The results obtained were exactly the same as those reported 

 in the Philosophical Magazine for 1906*. No difference was 

 observed as the result of employing these liquids of ex'ra 

 purity and of the extra precautions taken to exclude traces 

 of oxygen ; nor does the intentional introduction of traces of 

 oxygen affect the result. 



Heptane gave a brilliant " Swan " spectrum in which the 

 groups 7(5635 &c), 8(5165 Ac), e (4737 &c.) and/(4315 &c.) 

 were identified. 



Hexane and octane gave the same result, as did also 

 carbon tetrachloride and chlorobenzene ; but with these last 

 the observations are more difficult and the blackening more 

 rapid. 



The first observations of the spark within liquids seem to 

 be those of Daniel")". He experimented with ether, alcohol, 



* Watts and Wilkinson, Phil. Mag. xii. p. 581 (1906). 

 t Daniel, C. £. lvii. p. 98 (1863). 



