49 



optically negative witli, as it seems, weak double refraction, 

 while parisite has strong, positive double refraction. 



If the parisite from Muso is in reality holohedral hexagonal, 

 then cordylite agrees with it also in this respect in contra- 

 distinction to the parisite from Narsarsuk, which is decidedly 

 trigonal. It is not, however, proved that the mineral from 

 Muso and cordylite may not be trigonal, as what is regarded 

 as hexagonal bipyramids of the first order might be such pyra- 

 mids of the second order. Such is, for instance, the case 

 with the mineral s pang о lite, the rhombohedral nature of 

 which could be established only by etching ^). 



Cordylite has been found only at the locality No. 2 on 

 Narsarsuk. The mineral occurred here very sparingly partly 

 implanted on loose specimens and partly on the pegmatite in situ. 

 On the specimens it is mostly associated with parisite and nep- 

 tunite, and also with the ancylite described below. The cordylite 

 crystals are often implanted on ægirine crystals, generally in 

 small depressions on the latter. Sometimes the mineral is met 

 with on neptunite crystals of type II and also on tabular 

 crvstals of lepidolite. The mode of occurrence shows that the 

 mineral belongs to a comparatively late generation, as both the 

 neptunite and the parisite, which are among the most recent 

 formations, are nevertheless older than the cordylite. 



11. Ancylite. 



This name is derived from the Greek word а^;!ГУ>^ос, curved, 

 and has been chosen in allusion to the fact that the crystals 

 of the mineral always have their planes strongly curved. 



This new mineral has only been met with in crystals, but 

 these are small in size and not very distinctly developed. There 

 are, strictly speaking, two different kinds of them, though a 



M Amer. Journ. of Science, 1890, Vol.39, p. 370. 



XXIV. 



