298 



C. ANDERSON. 



V is a hypothetical segment twinned to I on m; to assume 

 its existence is permissible as the position is a possible and 

 probable one. We then have the following relations: — 



I A II = 57' 

 III a IV = 57 



III a II = 59 



IV A V = 62 

 III A V = 60 



15'} 



19 



26 ) 



06 



35 



r-twin 57° 18' 

 ra-twin 62° 46'. 



Here again the two r-twins I and II, III and IV conform 

 very closely to the theoretical angle, while III to II and 

 III and IV to the assumed segment V (twinned to I on m) 

 show the required divergence. 



Group V. — In this the segments are unequally developed, 

 I and II being much larger than III and IV. 



Segment. 



v„ 



Limit. 



Number of 

 Observations. 



</>o 



I 



II 



III 



IV 



164° 21 

 221 39 



286 58 

 281 30 



164° 19 - 164° 23 

 221 33 - 221 42 



286 56 - 287 

 281 29 - 281 30 



3 



7 

 2 



2 



/ 







57 18 



122 37 



117 9 



In this group, therefore, I is twinned to II and to III on 

 r (angles 57° 18' and 57° 23', calc. 57° 18'), and I is twinned 

 to III on m at an angle of 62° 51' (calc. 62° 46'), the varia- 

 tion from the calculated angles being within the limits of 

 observational error. 



Group VI (Plate LII, fig. 2, Text Fig. 4).— This consists 

 of thirteen segments twinned on r and extended in the 

 direction of the vertical axis to form a columnar aggregate 

 measuring about 8 by 2 cm.; the edges formed by the r 

 faces project horizontally, radiating from the central part 

 of the column. The group may be described as a polyet of 

 arrow-head twins on r elongated parallel to the vertical 

 axis and united by the points of the arrows. Reflections 

 are moderately good, but for some of the segments the 

 measurements were rather meagre, and a high degree of 



