238 Rejjorts and Proceedings — Geological Society of London. 



show the obtuse lobate character, previously suspected by Huxley 

 to obtain in this genus. Large basal scales lie on each side of each 

 pectoral fin. 



The ventral fins are abdominal. The right, which is best pre- 

 served, exhibits the arrangement of the scales which is described, 

 and which gives a clue to the disposition of the underlying bones or 

 cartilages. This must have closely resembled that in some Elasmo- 

 branchs. The same type of fin may be traced, though with important 

 modifications, in Pohjpterns, Pohjodon, and Acipenser, whilst in other 

 recent Ganoids and in Teleostei the arrangement is widely different. 



Between the ventral fins are three large scales, one median and 

 two lateral. On the left side of the median scales lies what appears 

 to be the anus. A similar arrangement seems to occur in Fterichthys. 

 This region is rarely exposed in fossils. 



The anal fin has also its pair of large basal scales. The caudal 

 fin cannot be well made out. There are indications of the under- 

 lying skeleton, but nothing can be distinctly made out. 



All the features of the present fossil confirm the opinion long ago 

 expressed by Pander and Huxley as to the near afiinity of Megalicli- 

 thijs to Osteolepis and Diplopteriis. 



3, " Studies on some Japanese Eocks." By Dr. Bundjiro Koto. 

 Communicated by Frank Kutley, Esq., F.Gr.S. 



The author has studied a series of Japanese rocks from the collec- 

 tions of the Tokio University and the Geological Survey of Jaj)an. 

 The microscopical investigation was carried on at the Miueralogical 

 Institute of Leipsic, under the direction of Prof. Zirkel, and the 

 chemical analyses were made in the laboratory of Prof. Knop. 



The most abundant rocks are the pyroxene-andesites, which are 

 not of a glassy texture, but for the most part holocrystalline. The 

 most abundant mineral in these rocks is a plagioclase felspar with 

 twinned and zonal structure, which is proved by its extinction -angles 

 and by the chemical analysis of its isolated fragments, to be labradorite. 

 Sanidine is present in small quantities. 



The augites of these rocks present many peculiarities ; they are all 

 decidedly pleochroic ; and they exhibit the oblique extinction in 

 basal sections, first pointed out by Mr. Whitman Cross, and which is 

 characteristic of triclinic and not of monoclinic crj'stals. A careful 

 examination of the question has led the author to conclude that the 

 mineral, which has lately been regarded as a rhombic pyroxene 

 (probably hypersthene), is really only ordinary augite cut parallel to 

 the optic axis. He does not regard the property of pleochroism as 

 distinctive of hypersthene, while the absence of a brachypinacoidal 

 cleavage and the presence of 10 jDer cent, of lime in the mineral 

 forbids our referring it to that species. 



The other abundant minerals in these augite-andesites are mag- 

 netite, which is always present, and quartz, which occurs in some of 

 them, both as a primary and a secondary constituent. Hornblende 

 is very rare in these rocks, and when present the peripheral portions 

 of the crystals are seen to be converted into augite, probably by the 

 action of the caustic magma upon them. Enstatite is rare in these 



