August 19, icSSo] 



NATURE 



569. 



faintest, and is only visible in a column of S feet. With 

 the exception of the 556 band, the other bands are so dis- 

 tinct that in a nearly saturated solution 4^ feet in depth, 

 they are clearly seen. This spectrum was so marked and 

 intense that at first we were sceptical with regard to its 

 belonging really to ammonia, thinking that possibly some 

 coal-tar product might be still clinging to the commercial 

 ammonia solution which in the first instance we used. To 

 satisfy ourselves that this was not the case, first we added 

 lime, and distilled the ammonia gas into pure water; this 

 acted inexactly the same way as the former solution ; then 

 we obtained from Messrs. Hopkins and Williams what they 

 guaranteed to be volcanic ammonia. A solution of this 

 gave exactly the same spectrum as the former solutions. 

 And lastly we prepared pure ammonia by Stas' method, 

 by the action of caustic potash and zinc — free from 



carbon — on potassic nitrite ; this also gave precisely the 

 same bands as the other ammonia solutions ; there could 

 therefore be no doubt that this spectrum belonged to the 

 ammonia solution. Thirty-six feet of ammonia gas did 

 not give us any indication of bands. Obviously this is 

 only a mere trace of gas as compared with the amount 

 held in the solutions before examined. To have as gas 

 the same amount as there was of solution in our 6-feet 

 tube, we should require a tube nearly a mile long. If 

 absolute alcohol instead of water be saturated with 

 ammonia, a spectrum (Fig. 3) still sharper than that 

 with water is obtained, but similar to it, excepting that 

 the band at 610 is wanting and the two bands at 650 and 

 630 now are of equal intensity, instead of the 650 band 

 being decidedly and constantly the darker of the two. 

 Ammonia giving so definite a spectrum it was evidently 



.u.iioc; 5, A.nyrAIculiol; 6, Ethyl AIv-Im 

 I, Aniline; 1 2, Toluidine ; 13, Turpentine. 



of much interest to ascertain what spectra would be 

 given by bodies of allied chemical constitution. Ethyl- 

 amine was the next substance we tried. Using a 33 

 per cent, solution, this gave a spectrum (Fig. 4) similar 

 in character to that of ammonia, but the dominant band, 

 as far as we could ascertain, has clearly moved towards 

 the red. It was now at 665 to 656. The next band is 

 also somewhat nearer the red than the corresponding 

 ammonia band. The position of the third band is very 

 nearly identical with that of the water-band, but instead 

 of being a wide band shading off on both sides, is now 

 narrow and perfectly sharp. It will be noticed that in the 

 alcoholic ammonia solution it is this band and the next 

 more refrangible one that are absent. 



For lack of material we have not yet examined the 

 spectra of other organic ammonias, but intend doing so. 



A solution of peroxide of hydrogen was also examined, 

 using the commercial 20-volume solution. The liquid 

 was not absolutely free from colour, and consequently 

 there was a very appreciable amount of absorption over 

 the whole spectrum. The water band was not visible, 

 and in fact no sharp band could be seen ; there was how- 

 ever a decidedly marked absorption commencing about 

 674, then the absorption is both dark and sharp ; it 

 extends, diminishing gradually, to 63S ; very probably this 

 absorption may prove to be a band, but the experiment 

 was not altogether satisfactory. 



We naturally returned to alcohol and other typical 

 organic liquids. Alcohol gives in the six-feet tube a very- 

 visible and fairly-defined band, more sharply defined 

 than the water band and nearer the red. It e.xtends from 

 632 to 624. The spectrum is given in Fig. 6. It will be 



