Alfred Harker — Various Crystalline Rocks. 171 



angles up to about 31°. This agrees with disthene or cyanite, a 

 n)ineral described by Riess ^ as a common constituent of eclogites. 

 Among the other minerals included by the garnet may be mentioned 

 quartz and apparently omphacite, mostly in i-ound granules clustered 

 in the centre of the crystal; zircon, commonly in little well-formed 

 crystals, bounded by the pinacoids (100) and pyramid (111) ; and 

 locally, near the margin of the garnet, crowds of little prisms with 

 moderate double refraction and sensibly straight extinction, possibly 

 zoisite. Some larger rounded grains of quartz containing little 

 zircons are inclosed near the border of the garnets, and occasionally 

 a small flake of brown mica. Larger flakes of mica, with strong 

 pleochroism, occur in densely -packed bundles moulding the surface 

 of the garnets. 



The rest of the rock consists chiefly of a crystalline granular 

 aggregate of colourless omphacite and clear quartz, the latter giving 

 ' undulose ' extinction, and often crowded with the straight hair-like 

 rods usually referred to rutile. This aggregate, however, incloses 

 numerous idiomorphic, thougli rather rounded crystals of enstatite 

 with a longitudinal striation and strong cleavage. There are also 

 a few scattered rod-like crystals of reddish-brown colour, and not 

 more than a tenth of an inch in length. These belong to a uniaxal 

 mineral with exceedingly high double-refraction and slight, thougli 

 distinct, dichroism (E > 0). It is evidently rutile, a mineral not 

 admitted in Eiess' monograph, but recorded by Mohl ^ in some of 

 the Norwegian eclogites. The crystals are simple tetragonal prisms, 

 either (110) or (100). 



(iii.) Garnet- Ampliiholite from Sutherland. 



The locality is three miles south of Laxford Bay. The rock 

 shows abundant red gai-nets imbedded in a mass of greenish-black 

 hornblende. A curious feature is the well-marked series of parallel 

 cracks by which the garnets are traversed. These maintain a 

 constant direction throughout the specimen, but do not affect the 

 hornblende. They are evidently due to stress in the rock, and, 

 though better developed, are of the same kind as the cracks seen in 

 the garnets of various crystalline rocks, such as the Eddystone 

 gneiss and some of the Saxon granulites. These cracks are always 

 at right angles to the direction of ' stretching ' in the rock, and so 

 perpendicular to any foliation or schistosity that may be present. 



A slice [1254] shows the rock to be composed chiefly of garnet 

 and green hoi-nblende. The garnet retains its red colour, and shows 

 well the series of parallel cracks pi'eserving a common direction 

 throughout the slide. There is a less pronounced system of cracks 

 roughly perpendicular to the first. The crystals are of irregular 

 shape, but tolerably free from inclusions. They show no double- 

 refraction, nor does the rock as a whole present any trace of foliation 

 or other parallel structure. 



The hornblende is in rude columnar crystals without terminal 



1 Tsch. Mill. -u. petr. Mitth. 1878, vol. ii. pp. 165-172, 181-241. 

 - JNyt Mag. Naturvidensk. toI. xxiii. (1877), pp. 128-137. 



