VOLCANIC NECKS OK HoKNSHV AND MWDAS. 503 



Vogt, 1 who considers the felspar crystals to have been 

 originally supersaturated in these dark coloured oxides, 

 which have since crystallised out from the solid felspar. 

 Still it must be noted tliat it is only in the more altered of 

 the Dundas rocks, where, from other evidence, there is 

 most sign of the percolation of solutions that these small 

 secondary plates or cavities have formed, for occasionally 

 they appear to be cavities not filled by oxides. 



A second alteration, more plainly in accord with Judd's 

 views is not confined to felspars only, but is common to the 

 rock as a whole. Bands of very minute irregularly shaped 

 cavities either filled with liquid or with oxides, pass in a 

 serpentinous manner through the rock from mineral to 

 mineral (see fig. 4). When these cavities are filled with 





Fig. 3 — Band of cavities filled with liquid in felspar, highly magnified. 

 black oxides they recall Harker's typical figure of a den- 

 dritic separation of material with which the host (there 

 olivine) 2 had been supersaturated. They are in Dundas, 

 however, generally in droplets, not in such com plex dendritic 

 forms. Felspar, pyroxene, olivine, and even spinel occas- 

 ionally may contain these cavities. 



A third common mode of alteration is the deposition, in 

 cleavage cracks and other partings, and in intergranular 

 spaces, of a dull green mineral of rather high birefringence 



' Quart. Journ. Geol. Soc. 1909; see also Harker, Natural History of 

 Igneous Rocks, p. 257. 



2 Natural History of Igneous Eocks, p. 258. 



