34 y. KIKUCHI 



reddiyh coloured thin spougy crust of ];iva. This thin coating 

 examined under tlie microscope, is found to consist of a brown amor- 

 phous glass, sometimes with microcrystals of PIagiocla«!3, a few 

 Au"ite u'rains and Mno-netite. The interior of the crystal in fresh 

 spechnens is transparent, and has a hyaline lustre. The cleavage is 

 perfect parallel to V. and M. The hiHer is less perfect than P, and 

 the cleavage-face always exhibits twin-striœ. A slightl}^ pearly 

 lustre is observed on both of these cleavage-faces; on one portion of 

 the brachypinacoidal cleavage-plane, there has been observed in one 

 specimen, a play of colour somewhat like that of Labradorite. Grains 

 of Olivine are always found as enclosures, sometimes arranged in dis- 

 tinct zones. In some portions glass-enclosures are found often 

 arranged parallel to a crystallographic direction. »Some specimens 

 show the beginning of decom[)osition, becoming hazy in appearance. 

 Sino-le individuals are rather rare ; they are mostlv found as twins, 

 or several individuals are grouped together in a most irregular man- 

 ner, assumini!' a ulobular mass. 



C^rvstal-f jrm.^ — The thin coiilinn' of lava^ which always covered 

 tlje crystals, rendered the determination of the crystal-faces peculiarly 

 difhcult. The reflecting goniometer could Jiot be used, except in 

 the rleavaufe-faces. 1)V nieasurinu' tlie auQ-le between cleavag-e-faces 

 V : i)/, and by iinding the position of the face ij. Avhich is always 

 developed, and the position of tlie Pericline twin-lamella' on i)/, I was 

 able generally to iix the position of the crystal, so tliat the other faces 

 could be determined by approximate measurements with a contact 

 goniometer and by zonal relations. Ir was not easv to get a perfectly 

 clear cleavage-piece for goniometric and optical purjioses, ;is the twin- 

 stria^, Olivine-enclosures, and numerous fissures which make the 

 crystals very brittle, were always present. The cleavage-face parallel 

 to P usually gave a good reflection, but that parallel to J/, a dis- 



