348 Mr. William Phillips on the OxydofTm. 



slonal exceptions certainly exist, reliance cannot be placed but on 

 crystals so small, or rather so minute, as that it may reasonably be 

 doubted whether it be possible for the most skilful hand to obtain with 

 accuracy the admeasurement of the angles formed by the meeting of 

 their facets by means of the common goniometer. The larger crystals 

 are certainly best adapted to the use of this latter instrument, and 

 hence, as I conceive, must have arisen, at least in part, the differences 

 in the results obtained by it, and by the reflecting goniometer. 



The admeasurement of the angle formed by the meeting of the 

 planes 1 and 2 Fig. 27, PI. 16. is prominently noticed by Haiiy. 

 This angle is first given in his Traite as 135°, and secondly in his 

 Tableau as 133' 29' ; the value of almost if not of every other angle 

 in any degree connected with this, likewise differs very materially. 

 These circumstances induce the supposition that having assumed 

 the value to be first 135° and afterwards 133° 29', the rest were 

 arrived at by calculation in both instances, and if so, were, of course, 

 dependent on the truth and accuracy of this single determination. 

 It is not therefore surprising that they should be made to differ so 

 essentially in the two works. 



In attempting the admeasurement of the angle above noticed, 

 viz. that of 1 on 2 fig. 27. PI. 16. the reflecting goniometer I first 

 employed, being graduated only to 5 minutes, never satisfactorily 

 gave an incidence of 133, 30, or 133. 35, but generally approached 

 as nearly to the one as to the other. This caused the suspicion that the 

 true value lay somewhere between them, and induced the wish for a 

 goniometer more highly divided j and I have obtained one graduated 

 to half a minute, from Mr. Carey, whose ingenuity led him to add to 

 it some apparatus with a view to precision in its use. By this instru- 

 ment, I have repeatedly found the angle in question to be 133 o. 32'. 

 30". — being 1°. 27'. 30". less than the former determination of 



