OLD RED SANDSTONE YOLCANIC ROCKS OF SHETLAND. 381 



form the main constituent. They are associated with a green fibrous mineral, 

 which is largely represented, and has apparently replaced the augite. In one 

 of the sections, the augite occupying the interspaces still occurs in forms which 

 are recognisable ; parts of the original crystals being quite fresh, while the 

 remainder has been converted into chlorite and a yellowish green mineral which 

 may be epidote. The unaltered portions show the characteristic play of colours 

 with polarised light. A little magnetite is diffused through the sections, but to 

 a much more limited extent than in the diabase lavas. It occupies quite a 

 subordinate position compared with the felspar and the green chlorite mineral. 

 We have already called attention to the fact that the boundary line between 

 the diabase and the granite is well defined. A section of the granite from the 

 junction contains triclinic felspars, clear quartz with fluid inclusions, and mica 

 partially converted into a green decomposition product. Numerous specks of 

 limonite occur, resulting from the hydration of the iron oxide. The rock 

 associated with this in the same microscopic section is a diabase similar to those 

 described. It contains no quartz, and the augite is faintly recognisable. 



The sections taken from the stream near Garder House, on the west bank 

 of Selie Voe, are coarsely crystalline like those from Skeld Hill. The plagio- 

 clase felspars are much kaolinised, and the pyroxenic mineral has undergone 

 intense alteration. 



Compared with these, the sections taken from the dykes in Northmavine 

 traversing the binary granite and felsite are exceedingly fine-grained. In one 

 instance, the crystals of plagioclase are very fresh, and the interspaces between 

 the crystals are filled with a bright green mineral, probably chlorite, there being 

 no fresh traces of the original augitic constituent. Magnetite is not abundant, 

 and small needles of apatite are disseminated through the mass. Another 

 example taken from Rooeness Voe exhibits much granular augite along with 

 the triclinic felspar, which has only undergone slight alteration. Magnetite is 

 also present in very small grains. 



The chemical analysis of a typical example of the Skeld Hill rock proves 

 that it is closely allied to the basic lavas ; the proportion of silica being 50*58 

 per cent. The percentage of magnesia is even larger than in the two examples 

 of the interbedded lavas, amounting to 8 "90 per cent. This feature plainly 

 indicates the great alteration which the pyroxenic mineral has undergone, and 

 the consequent development of magnesian silicates. It is worthy of note that 

 while the specific gravity of the bedded lavas is 27, that of the Skeld Hill rock 

 is 2 - 9 per cent. 



The microscopic characters of the great intrusive sheets in Northmavine 

 and Sandsting have many points in common. A section from the Heads of 

 Grocken shows that the rock is coarsely crystalline, consisting essentially of red 

 orthoclase and quartz. The felspar is much kaolinised ; but in one instance 



