On the Constituents of Various Minerals. 245 



ent earths, &c. I need hardly remark, that the analysis of the 

 magnesian minerals is the most difficult, and requires the long- 

 est practice. As soon as I find that the pupil has acquired suffi- 

 cient skill, he is left entirely to himself. All the precautions I 

 think necessary to take, is to give him two or three rather diffi- 

 cult minerals, which I have previously analysed myself. This 

 enables me to judge how far I can depend upon the accuracy 

 and sagacity of the pupil. It has sometimes happened, in these 

 cases, that the pupil has detected substances in the mineral 

 which I myself had overlooked. In such cases, I repeat the ana- 

 lysis again myself, and generally find that the analysis of the pu- 

 pil was more correct than the one that I had originally made. 

 The pupil is always exercised in these kinds of investigations for 

 a considerable time, and I do not give him new minerals (ne- 

 ver before investigated) to analyse, till I have had ample evi- 

 dence of his skill and accuracy. 



The minerals of which I mean to give the analysis in this pa- 

 per, were analysed in the laboratory of Glasgow College, almost 

 all of them since July last. I shall take them up without any 

 order, being guided, in some measure, by the comparative im- 

 portance of each. 



1. Sillimanite. 



This mineral was found at Petty Pog, in the township of Say- 

 brook, Connecticut. It was described and analysed by Mr 

 BOWER (Journal of the Academy of Sciences of Philadelphia, 

 p. 375). For the specimens which I was enabled to subject to 

 analysis, I was indebted to Mr NUTALL. 



It occurs in long four-sided prisms, generally bent, in a mica- 

 slate rock ; but the portion in which the Sillimanite is found is 

 quartz ; perhaps a vein. 



i i 2 



