86 PROCEEDINGS OF THE AMERICAN ACADEMY 



Professor Cooke made the followins communication : — 



PonoP =^ 122° 15' 



PonP = 129 56 



RonR = 85 47 



a = 1.374 



"The above drawing is an accurate representation of a crystal of 

 rhombic arsenic which was picked out from a quantity of sublimed 

 arsenic. The length of the lateral axes is about one third of an inch, 

 and the crystal almost perfect. The angle of P on P was meas- 

 ured by a reflective goniometer, which reads to minutes. The faces 

 were very perfect, but somewhat dull, owing to a slight tarnish. This 

 made it difficult to measure with the usual signals, but the difficulty 

 was easily overcome by making the measurements in a darkened room, 

 and using for the upper signal a horizontal slit cut in a piece of turned 

 iron about four inches in length and one fourth of an inch in diameter, 

 which was fastened to the upper part of a window-sash, and the rest 

 of the window covered with black cloth. The lower signal was a 

 black line drawn on a white card. In this way very perfect and suf- 

 ficiently bright images of reflection were obtained, as will appear from 

 the fact that the greatest difference between twelve measurements did 

 not exceed two minutes. The angle given (P on P) is the mean of 

 the twelve. The other angles were obtained by calculation. 



" The angle R on R, as given by G. Rose, is 85° 4', by Breifhaupt 

 85° 26', by Miller 85° 41'. My own measurements gave 85° 47'. 

 This difference, too great to be referred to any errors of observation, 

 is probably occasioned by apparent variations in the angle produced 

 by stride. Of ten or twelve crystals which I have examined, the one 

 described above was the only one which had perfect planes. On 

 some larger crystals, having the same form as the one described, the 

 angle R on R measured approximately 85° 20' and 80° 31'. The 

 variation from the normal angle was evidently caused by striae, which 

 spread out the image of the signal into a broad band formed appar- 

 ently by several images overlapping each other. (In measuring, the 

 brightest portion of this band was selected as the starting-point.) 



