296 



C. ANDERSON. 



are mostly smooth and brilliant, giving good reflections. 

 In the following and succeeding tables the best measure- 

 ments, suitably weighted, are used to fix the position of the 

 b pinacoid and in the accompanying text figures the orien- 

 tation of the various twin segments is indicated by the 

 position of the normal to b ; V c is the mean of the actual 

 goniometric readings, </> gives the angular distance from 

 the first meridian (position of segment I). 



Segment. 



v . 



Limits. 



Number of 

 Observations. 



4>o 



I 



II 



III 



IV 



117° 18 

 174 28 

 235 51 

 293 10 



117° 16 - 117° 22 

 174 17 - 174 32 

 235 44 - 235 56 

 293 10 - 293 11 



3 



5 

 3 

 2 



0° 



57 11 



118 33 



175 52 



Thus we have the following angular relations: — 

 lAlI-57° 11' (r-twin 57° 18'). IaIII = 61° 27(m-twin 62°46'). 

 Ill a IV = 57° 18'. IIaIV = 61°19 / . 



II a 111=61° 22. 



The divergence therefore from the position of an w-twin is 

 in the sense demanded by Goldschmidt's hypothesis. In 

 Text Fig. 2 the orientation is shown in stereographic pro- 

 jection, it being assumed that I and II and III and IV are 

 inclined to one another at the precise angle of twinning; 

 the position which the poles of III and IV would occupy if 

 these segments were twinned on m to I and II respectively 

 are indicated by the angular values enclosed in parentheses. 

 This projection shows clearly that the chief zones are 

 brought more nearly into parallelism than they would be 

 if the exact angle b \b were maintained ; for example r x 

 (r 2 ), a 3 are practically coincident as are a 4 and r a , r 3 and Oi, 

 a 2 and r 4 . Indeed one might describe III and IV as hetero- 

 twins to I and II in which the vertical axes and the zones 

 [c a] and [c r] are parallel. 



