28 Mr. W. Crookes. On the [Jan. 13, 



Fig. 2. 



150 170 190 210 230 250 270 290 310 330 550 370 390 410 430 +50 



Spectrum of Spinel (Magnesium Aluminate). 



In the ' Comptes Rendus ' for December 6th last* appears a brief 

 note byM. de Boisbaudran, in which he announces, "to take date, that 

 alumina, calcined and submitted to the electrical discharge in a vacuum, 

 has not given him a trace of red fluorescence. This fluorescence, as well 

 as its special spectrum, shows itself brilliantly when the alumina contains 

 Y^oth and even T1 ^o^ of Cr 2 3 . With the 10 l 00 th part of Cr 2 3 we 



still obtain very visible rose colour From these observations 



the presence of chromium appears to be indispensable to the production of 

 the red fluorescence of alumina." 



This statement being opposed to all my experience, I immediately 

 instituted experiments with a view, if possible, to clear up the 

 mystery. I started with aluminium sulphate, which I knew to be 

 tolerably pure, and in which ordinary tests failed to detect chromium. 

 On ignition and testing in the usual manner in a radiant matter tube, 

 the alumina line was brightly visible in the spectrum of the emitted 

 light. Different portions of this aluminium sulphate were now 

 purified by various processes for the separation of chromium. All 

 gave as a result the absence of this impurity. The most trustworthy 

 process being that devised by W6hler,f I used it to purify the bulk. 

 The salt was dissolved in water, and excess of caustic potash added 

 till the precipitate first formed redissolved. Chlorine was now passed 

 through till no more precipitate fell down and the liquid retained a 

 strong odour of chlorine. The whole of the chromium would now be 

 in solution, whilst the alumina would be in the precipitate. The 

 alumina was filtered off, well washed, and a portion tested in the 

 radiant matter tube. It gave as good an alumina spectrum as did the 

 original sulphate, the crimson line being very prominent. 



The alumina thus purified was a second time dissolved in caustic 

 potash and submitted to the chlorine purification. Again in the 

 r adiant matter tube the alumina gave its characteristic crimson line 

 spectrum. 



* ' Comptes Rendus,' rol. 102, p. 1107. 



f ' Select Methods in Chemical Analysis,' 2nd edition, p. 124. 



