SCIENCE. 



105 



The carbon tetrachloride was prepared by Dr. Hodgkin- 

 son, who very kindly supplied me with sufficient for my ex- 

 periments. 



On passing the spark without the j\.r in this tube, the 

 spectrum observed consists of those sets of flutings which, 

 according to Messrs. Liveing and Dewar, are due to hydro- 

 carbon, and the set of flutings which is reversed in the sun, 

 and ascribed by Messrs. Liveing and Dewar to cyanogen, 

 also appears in a photograph of the violet end of the spec- 

 trum, Fig. 2. On connecting a Leyden jar with the coil and 

 then passing the spark the flutings almost entirely vanish 

 and the line spectra of chlorine and carbon take the place of 

 the flutings without either a line of hydrogen or a line of 

 nitrogen being visible. 



As a long experience has taught me that these tubes often 

 leak slightly at the platinums after they are detached from 

 the pump, so that the evidence of such a piece justificatij is 

 only good for a short time, I took the occasion afforded by 



principal double line in the green being seen. The hydrogen 

 line H«(C) was faintly visible when I first observed the 

 spectrum, but it got gradually weaker and finally disappear 

 ed altogether. When this line was no longer visible the con 

 denser was taken out of circuit again, and the same carbon 

 bands were seen as before. These bands, therefore, show 

 themselves with great brilliancy when a strong and powerful 

 spark does not reveal the presence either of hydrogen or 

 nitrogen. (Signed) Arthur Schuster." 



"March 21, 1880." 



This result, which entirely endorses the work of Attfield 

 and Watts, has been controlled by many other experiments. 

 I have also repeated Morren's experiment and confirm it 

 and I have also found that the undoubted spectrum of cyan- 

 ogen is visible neither in the electric arc nor in the surround- 

 ing flame. 



Hence then in the case of carbon, as in the prior cases 

 of hydrogen, nitrogen and the like, those who hold that 



Fig. 



a visit of Dr. Schuster to my laboratory while the experi- 

 ments wese being made to get my observations confirmed. 

 He has been good enough to write me the following letter 

 and to allow me to give it here : — 



''March 21. 



"My Dear Lockyer. — The following is an account of 

 the experiment which I saw performed in your laboratory on 

 Monday, March 15 : 



" A tube containing carbon-tetrachloride was attached to 

 the Sprengel pump. As exhaustion proceeded the air was 

 gradually displaced by the vapor of the tetrachloride. The 

 electrodes were a few millimetres apart. If the spark was 



the flutings a;e due to impurities must, it would seem, 

 abandon their position ; for the flutings are undoubtedly 

 produced by carbon vapor. Nor is this all ; the sugges- 

 tion that the various difficulties which have always been 

 acknowledged to attend observations of this substance may 

 in all probability be due to the fact that the sets of carbon 

 flutings represent different molecular groupings of carbon, 

 in addition to that or those which give us the line spectrum, 

 and that the tension of the current used now brings one set 

 of flutings into prominence and now another, seems also 

 justified by the facts. This suggests the view that a body 

 may have a fluted spectrum of compound origin as well as 



SPARK WITHOUT 

 JAR IN CXIa . 



SPARK IN 



BUNSEN 



FLAME 



tARBON 

 FLUTINGS 



Fig. 2. 



taken without a condenser in the vapour the well-known 

 carbon bands first observed by Swan in the spectrum of 

 a candle were seen with great brilliancy ; I also saw the 

 blue band which you said was identical in position with 

 one of the blue bands seen in the flame of cyanogen or 

 in the spectrum of the electric arc. When the condenser 

 and air-break were introduced this spectrum gave way to a 

 line spectrum in which I could recognize the lines of chlo- 

 rine. The lines of nitrogen were absent, not a trace of the 



a line spectrum. 



This conclusion is greatly strengthened by the preliminary 

 discussion of a considerable number of photographs of the 

 spectra of various carbon compounds. 



A general comparison of the photographs first enables 

 us to isolate the lines in the blue and ultra-violet portions 

 of the spectrum (wave lengths 4300-3800) of the substance 

 associated with the carbon in each case. 



In this manner the lines seen in the photographs of the 



