^Ol. 57.] EOCKS FROM DTJPTOX TIKE. 35 



In polarized light, the few clearly recognizable crystals and portions 

 of crystals of felspar are, as already stated, apparently to be referred to 

 oligoclase and andesine; but, with regard to the more highly 

 altered fragments which chiefly constitute the rock, still less can be 

 affirmed with any confidence, since a definite boundary indicative 

 of an idiomorphic crystal, other than a ragged and approximately 

 straight line, can rarely be discovered to form part of the boundary of 

 one of these fragments or lapilli. When, however, such a boundary 

 does occur, it is often found to lie at 0° to approximately 2i^ 

 with the direction of maximum extinction, so that it seems 

 probable that these fragments may in many cases be referred 

 to orthoclase. Some of them, moreover, show indications of 

 twinning on what looks like the Carlsbad type. 



The most remarkable feature about these fragments is that they 

 appear to consist of a meshwork of small, colourless rods, lying 

 apparently in any direction and intersecting at any angle. That 

 these rods are not individual crystals is proved by the fact that, 

 in each separate fragment, all the rods undergo simultaneous 

 extinction between crossed nicols. It is evident, then, that they 

 all belong to one and the same crystal. Each of these fragments, 

 therefore, represents not only the breaking-up of a crystal into 

 fragments, but the partial erosion of that fragment both superficially 

 and internally (PL I, figs. 4 &■ 5). 



There is one, and, so far as I can ascertain, only one, fragment m 

 this section that afi'ords clear proof of the foregoing statement. It 

 is a rudely triangular fragment of unaltered felspar, which is seen to 

 pass into a mesh of reticulating rods (PI. I, fig. 7). Both the fresh 

 felspar and portions of its adherent mesh extinguish simultaneously, 

 thus proving that the meshwork of rods and the unaltered fragment of 

 felspar are parts of the same fragment, the whole having originally 

 formed part of one crystal. The little rodlike bodies which constitute jr 



the mesh, partly fringing and continuous with the fragment o£ felspar, 

 have apparently the same refraction and other characters as the 

 felspar- fragment from which they proceed. Why the felspar has been 

 removed from the spaces lying between these small intersecting rods 

 is a matter which I leave others to decide. The removal of the 

 felspar could scarcely have taken place along planes of more ready 

 solubility, because the interbacular spaces are for the most part 

 small triangular or polygonal areas ; yet that some kind of selective j| 



solution has caused this peculiar honeycombing of the felspar- P 



frao-ments seems an unavoidable conclusion. The matrix or ground- 

 mass in which these fragments are embedded appears dark during 

 all stages of rotation between crossed nicols ; and, when tested with 

 a Klein's plate, scarcely any perceptible difference in the uniformity 

 of the tint is to be observed, except that here and there a few minute 

 birefringent specks may be discerned. When, however, _ this prac- 

 tically isotropic matter is examined in ordinary transmitted light, 

 it is seen to be crowded with globulites and little rods like those 

 which constitute the altered felspar-fragments, except that, as a 

 rule, they show no double refraction. Without the aid of polarized 

 liaht it is, therefore, difficult to distinguish the altered fragments 

 from the matrix in which they lie ; but in those parts of the 



D 2 



