048 ANALYSIS. 



Defcrlption and are ctmpofed, and are light and delicate In the forms which 

 analyfis of a jj^g^ affumc, or they are grouped together according to a 

 Covijwall. variety of degrees of proximity and compadnefs. Sometimes 



they fill the v»'hoie cavify of a ftone, with little or no interrup- 

 tion; in other fpecimens they are feen partially fpreading over 

 the fides and pointed pyramids of quartz cryftals. 



In feme cafes thefe grouped (ufts adhere very pertinaceoufly 

 to the Hone which bears them; in others, they are eafily fe- 

 parable, in comparatively large pieces, from the quartz, the 

 imprelTed form of wliich the pieces thus feparated retain. The 

 furface of thefe, which was in immediate contact with the; 

 quartz, exhibits the feveral minute cryftals of which the mafs 

 confifls, matted together in various direftions, 



Thefe cryftalline afleftiblagcs are, in general, white; a 

 nearer infpeflion of tlie individual cryftals proves that they 

 are tranlparent. Sometimes they are ftained of a yellovvilh 

 hue by ochry water. 



The fize ot thefe cryftals varies confiderably in different 

 fpecimens. Sometimes (hey aflume the appearance of a white 

 powder raifed up in fmall heaps, upon the furface of the flone, 

 to which they adhere. In other fpecimens they refemble a 

 tender- down. And the larger tort varies, in relative fize, in 

 the proportion, perhaps, in which a human hair, horfe-hair, 

 9nd a hog's briftle, feverally differ from each other in magni- 

 tude. They feldom exceed a quarter of an inch in length. 

 The figure of thefe cryftals is not eafily afcertainable, on ac- 

 count of their minutenefs. By the help of a very powerful 

 microfcope, they appear to confitl of four-fided prifmS ; where 

 ibefe are broken off, the feflion exhibits a rhomboidal, ap- 

 proaching indeed to an elliptical figure, from the circumftance 

 of the angles of the prifm being worn away ; but that the 

 prifm itfelf is rhomboidal, cannot be inferred from hence, unlefs 

 we could be certified, that the fef^ion were at right angles 

 with the axis of it. 



Imbedded amongft thefe cryftals two fpecies of cryftalline 

 laminae are frequently difcoverable ; the one confifting of pa- 

 rallelopipedon plates with truncated angles, applied to each 

 other, of a green colour of various tints, from the emerald to 

 the apple-green : the other fpecies, confiding of an affemblage 

 of fquare plates, which vary in thicknefs. The angles of the 

 feveral fquare lamin??, which are applied to each othefi are 



not 



