36 Mineralogical Notices. 



Transactions for 1821, with onl}' one slight point of difference, 

 viz. that Mr. Kemp neutralizes his solution cold., while Sir J. 

 Herschel recommends the neutralization to be performed at a 

 boiling temperature, a precaution which, as will appear on a 

 perusal of the paper in question, is by no means to be neg- 

 lected when a mathematically rigorous separation of the oxides 

 is desired. 



Mr. Kemp also notices that where cheapness is an object, 

 carbonate of soda may be substituted for that of ammonia. 

 Sir J. Herschel has shown that any earthy or metallic carbo- 

 nate will equally answer the purpose, so that (for the prepara- 

 tion of the sulphate at least, for manufacturing purposes) 

 chalk would be that substance to which the manufacturer 

 would resort, as no harm can arise from adding it in any 

 excess, and the small quantity of sulphate of lime remaining 

 in solution would be readily separated by evaporation to dry- 

 ness and redissolution. An experiment is related in the paper 

 referred to, where a very minute per-centage of manganese 

 was insulated from an enormous excess of iron by simply fil- 

 tering the hot solution through chalk. 



1 am, Sir, &c., 



Annotatok. 



IV. Mineralogical Notices^. 



Analysis of the Greenovite, by M. Delesse. 



T^HIS mineral, found at Saint Marcel in Piedmont by 



-■■ M. Bertrand de Lorn, has been described by M. Du- 

 frenoy, who has given its crystallographic characters ; the ana- 

 lysis made by M. Cacarrie on a too small quantity presented 

 much uncertainty, which has induced me to submit it to fresh 

 investigation. 1 have in fact found that if the crystallization 

 gave an exact idea of that species, its composition had caused 

 it to be regarded, erroneously, as a titanate of manganese, 

 whereas it is a silico-titanate of lime, analogous to sphene, 

 as results from the following analyses : — 



I. Oxygen. 



Silica 29-80 15-48 



Oxide of titanium . .43-00 17-07 



Lime 23-60 6-33 



Protoxide of manganese 2-90 0*65 



Protoxide of iron, traces. 



99-30 100-50 



These elements are well enough represented by the formula 



(2 Si R3 + Tf R) + 3 Si ti2. 

 * From the CojnjHcs licndus, November 11, 1844. 



