io6 



SCIENCE. 



spectra of CCL, C,„H S , CN, CHI 3 , CS,, C0 3 , CO, &c, 

 have been mapped, and both the common and special lines 

 and flutings thus determined. 



The phenomena seen with more or less constancy are a 

 blue line, with a wave-length of 4266 ; a set of blue flut- 

 ings, extending from 4215 to 4151 ; and another set of ultra- 

 violet flutings, which extend from 38S5 to 3843 (all approx- 

 imate numbers). 



In a photograph of the spectrum of the electric arc 



the spectrum which contains the blue line alone and that 

 which contains the blue fluting alone (Fig. 4). In compar- 

 ing the spectra of carbon under different conditions, I find 

 this to be true. The blue line never appears in conjunction 

 with the blue flutings, unless the ultra-violet flutings are also 

 present. In other words, the highest and the lowest hypo- 

 thetical temperature spectra are never visible together with- 

 out the spectrum of the intermediate hypothetical temper- 

 ature. 



Highest temperature. 

 Intermediate. 

 Lowest temperature. 



Fig. 3 — Action of three different temperatures on a hypothetical substance, assuming three stage of complete dissociation. 



(with a weak battery) between carbon poles in an atmos- 

 phere of chlorine, the blue flutings alone are visible, whilst, 

 when the spark is similarly photographed, the ultra-violet 

 flutings and the blue line (4266) are also visible, whilst 

 the blue flutings become fainter. 



From this we may assume, in accordance with the 

 working hypothesis of a series of different temperature 

 furnace?, as set forth in the paper of December, 1878 (see 



But this is not all. By placing the spectra of the sub- 

 stances at different heat-levels, so to speak, I was enabled 

 to construct a map, which not only indicates the mere pres- 

 ence or absence of , the lines and their relative intensities, 

 but shows a perfect gradation between the spectrum which 

 contains the line alone and that which contains the blue 

 flutings alone (Fig. 5). I would point out that there is noth- 

 ing theoretical in this map. All the horizons depicted are 



Furnace A 



Inter- ( 1 

 mediate. \ 2 



Furnace B 



Inter- j 1 

 mediate. ( 2 



Furnace C 



■BIIIHHHHMI«Hi 

 ■KIIHHHHKMM 



■iMIIII'ill'tlillllliilllllllillg— 







Highest temperature. 



Lowest temperature. 



F'G. 4 — Spectra of the hypothetical substance, in intermediate furnaces, assuming that the vapours are not completely dissociated. 



Fig. 3), that the different flutings and the line correspond 

 to different temperature spectra, the blue flutings to the 

 lowesi and the blue line to the highest temperature, whilst 

 the ultra-violet flutings occupy an intermediate position. 

 According to this working hypothesis there should be 



copied from photographs of carbon under the conditions in- 

 dicated, and theory has merely enabled me to arrange them 

 in order. 



This map I submit, therefore, bears out the hypothesis of 

 differences of temperature indicated above, for it is seen 



Highest temperature. 



Lowest temperature. 



HE^S Hm£^I 



^■nHHMBHHHBl 



IH 



■ mil 





■inn mm 



1 



■ mil 





1 mil niii 





HUB nilllHI 



1 





Spark in Ci H B and CHli 

 Spark CO 



" CO a 



" CN. 



" CS. 

 Arc in CI (B). 

 " CI (C). 



FlC. 5 — The photographed spectra of some carbon compounds. 



a scries of horizons forming a perfect gradation between I flutings appear first and grow in intensity. As these in- 

 that, while the blue line gradually thins out, the ultra-violet 1 crease the blue flutings become visible, and further, as the 



