4ngles of M 'mute Crystals. 307 



chemical agents. My attention, however, was much directed 

 to the beautifully minute crystallizations which took place m 

 that apparently limited space ; sometimes above a hundred ex- 

 quisitely symmetrical and perfect crystals lying together and 

 capable of distinct observation, in the space of the 5000dth of a 



square inch. . . 



A difficulty, however, arose as to the nature and composition 

 of these crystals. The substances brought together in this way 

 necessarily meet each other in unmeasured and, to us, indefinite 

 quantities. The whole quantity used was too small for weight ; 

 and, even if each substance could have been weighed, whetl 

 .pread with extreme tenuity over a comparatively large surface 

 of glass, it was impossible to ascertain the proportion, or the 

 proportional strength in which one substance would meet others 

 at any given point of the whole surface covered. When changes 

 took place, and new combinations and crystallizations appeared, 

 a degree of doubt therefore arose as to the precise nature of the 

 substance. At all events, the nature of the substance could 

 not be predicated so decidedly as in a case where the whole 

 measured quantity of two or three known agents, known also 

 in their relations, assumed in combination one marked and ta- 

 miliar character. And if this doubt would arise with respect to 

 minute crystallizations originating in the combination of known 

 substances, it would be still more the case in the attempt to ana- 

 lyze by this mode of microscopic investigation some small por- 

 tion of a substance whose composition was not known ; and m 

 which the varied results under the tentative process must be 

 crreatly, at least for a time, matter of conjecture. 

 " I have no doubt, after lengthened observation, that familiarity 

 with microscopic crystalline forms, and the accurate recording 

 and classifying of phenomena, would go far to remove much of 

 this uncertainly ; but besides this, it appeared to me that, as m 

 the case of many substances, the angles of crystallization are ac 

 curately known, the range of doubt might be considerably limit- 

 ed, if the angles of crystals so formed could be submitted to ac- 

 curate measurement. And, with this view, I adopted the fol- 

 lowing arrangement for the application of a micrometrical gonio- 

 meter to the microscope : • . j 



The eye-piece of my microscope is a tube about an mch and 



