Church — On Native Phosphates of Aluminium. 555 



§ VIII. Calaite. — 'Eow we have in calaite a mineral adapted for 

 the trial of this question. There can be no donbt that the essential 

 part of pure turquoise or calaite may be represented by the empirical 

 formula — 



2AI0O3, P2O5, SKO. 



But we have to ask whether the constitution of this species may not 

 be expressed by some more precise arrangement, such as 



AI2O3, PA, 2H2O + AI2O3, SH.O, 



namely, one molecule of a hydrated normal aluminium phosphate with 

 one molecule of the normal aluminium hydrate ? ISTow we have seen 

 that the latter compound as occurring in nature loses no water at 

 100°, while the nearest approach to the former compound (hamely 

 variscite) loses all its water at this temperature. Calaite, then, if 

 containing such a phosphate, should part with fibs, of its water 

 below 100° C. As it retains all, even when heated beyond this point, 

 it will be more in accordance with the results of experiment if we 

 express the molecule of calaite somewhat after the following fashion : — 



2A1,03 ( P.O.. ___ Al, ( 2P0, 



H2O (4H,0, "" H2(8HO, 



which may be abbreviated into 



The analyses upon which this view is founded are three, one pub- 

 lished in 1864,^ and the others lately made for the purposes of this 

 report.* The relation of the cupric, ferrous and manganous phos- 

 phates to the main constituent of calaite having been fully discussed 

 elsewhere, a few only of the later analytical results need be cited here. 

 In the case of a very pale and pure specimen of calaite from Mshapur, 

 Persia, the following figures were obtained : — 



•584 gram lost "018 gram H2O in vacuo over H2SO4 - 3-08% 

 •584 „ ,, -118 „ „ on ignition =20-85%. 



Nothing was lost at 100° in the air-oven: but of the 20*85% water 

 driven off on ignition the last 4% was retained with greater tenacity. 

 Similar results were yielded by a Thibetan turquoise. 



§ IX, Two other aluminium phosphates have been described, appa- 

 rently belouging to the calaite or turquoise group, and differing from 

 calaite in the proportion of water only. One of these is the peganite 

 analyzed by Hermann, a Striegis mineral ; the other is fischerite from 

 Nischne Tagil, also analyzed by Hermann. I find that three distinct 

 minerals from Striegis go under the name of peganite. One of these is 



3 Chemical JSfeivs, x. p. 290. 



^ Zellner's analysis of a Silesian calaite confirms this yiqw. 



