308 THE CANADIAN NATURALIST. rVol. vil. 





Chemical and Bloiiqnpe Characters. — Before tlie blowpipe 

 colours the flame intensely yellow, becomes opaque, and often 

 exfoliates or swells up into cauliflower-like forms. After ignition 

 in the forceps, or in the closed tube, gives a strong alkaline 

 reaction. Fragments which have not been ignited, when placed 

 upon a piece of moistened turmeric paper, shew no alkaline 

 reaction ; but if the finely pulverised mineral is treated with 

 waterj the water is rendered slightly alkaline. In the closed 

 tube gives ofi" water and carbonic acid. With nitrate of cobalt 

 gives a fine blue colour (alumina.) With hydrochloric or nitric 

 acid dissolves in the cold completely, with evolution of carbonic 

 acid ; and this even when the mineral is in fragments and the 

 acid exceedingly dilute. Addition of ammonia to the solution 

 gives a copious precipitate of alumina. Acetic acid decomposes 

 it, but does not appear to dissolve it completely ; the solution, 

 however, gives an abundant precipitate of alumina with ammonia. 

 Through the kindness of Dr. Dawson I have been enabled to 

 obtain sufficient material for two analyses. The first was made 

 some months ago, but so strange did the results appear, that I 

 was unwilling to publish them before making a second analysis, 

 in order to ascertain whether the mineral was at all constant in 

 composition. 

 , The first analysis gave me the following results : 



I. 



Carbonic acid 29.88 



Alumina* 32.84 



Lime 5.95 



Magnesia traces. 



Soda 20.20 



Potash 0.38 



Water 11.91 



Silica 0.40 



101.56 



The carbonic acid was determined with an ordinary single 

 flask apparatus, and the water with a small chloride of calcium 

 tube. Together they equal 41.79 per cent. Direct ignition of 

 a separate portion of the mineral in a covered crucible gave a 

 loss of 41.16 per cent. 



For the second analysis, the material was ■ obtained from at 

 ^east twenty difi"erent specimens, and considering that the amount 

 employed was small, the results are sufficiently close to those of 



* With traces of peroxide of iron. 



