462 Messrs. J. H. and G. Gladstone on Collyrite, and a 



in the first column. The second column gives the theoretical 

 proportions calculated from the formula SiO 2 , 2Al 2 3 + 9HO. 



It is perfectly clear that the silicic acid and the alumina are 

 in this ratio, but the amount of water is rather low for 9 equi- 

 valents. 



I. II. 



Silicic acid 14*74 14*14 



Alumina 48*25 48*02 



Water 37*01 37*84 



10000 100*00 



The mineral agrees, both in" physical characters and in chemical 

 composition, with that which has been described under the name 

 of Collyrite, and to which the formula SiO 2 , 2 A1 2 3 + 10 HO 

 has been attributed; but our specimen appears to have been 

 purer, and to have given more accurate numbers for the disilicate 

 of alumiua than those analysed by previous observers. 



But no two portions analysed gave exactly the same composi- 

 tion. Some had a larger amount of carbonate of lime : thus a 

 piece which was considerably harder, and broke with a conchoidal 

 fracture, was found to contain between 5 and 6 per cent, of lime- 

 salt. Some had a much smaller amount of silica : thus a piece 

 which very easily fell to powder, and had the specific gravity of 

 1*99, gave about the following proportions : — 



Silicic acid 3*4 



Alumina 60*5 



Carbonate of lime . . . 0*6 



Water and loss .... 35*5 



100*0 



But the most remarkable specimens were from another part 

 of the quarry. They had the same physical characters as those 

 already described ; but quantitative analysis showed that they 

 contained more carbonic acid than was necessary to saturate the 

 lime, and that there was no other base present except the 

 alumina. Now as bicarbonate of lime is soluble in water, and 

 carbonate of alumina is unknown as a mineral species, and has 

 seldom, if ever, been procured even in the laboratory, it seemed 

 desirable not to depend on one analysis. A quantity of one 

 specimen was therefore pounded up and several determinations 

 were made of each constituent, and that by different processes, 

 the mean results of which are given in the first column of the 

 subjoined Table. Another specimen was reduced to powder, 

 and exposed over sulphuric acid in vacuo for a whole month, by 



