Vol. 59.] IN GLASSY IGNEOUS EOCKS. 437 



oxide) of which one acts as a kind of matrix to the other. 1 At 

 first, the crystal-growth in a spherulite is simply radial ; bnt, alter a 

 time, the interstices between the growing 'stems' become sufficiently 

 wide to allow them to throw off side-branches. These in some 

 cases contrive to interlock and form a kind of mat of rectilinear 

 branches, 2 but more commonly they interfere, with a result some- 

 what resembling that observed when trees in a wood are ' drawn ' 

 by being planted too closely. 



But, as the history of the ordinary spherulite has been traced by 

 Mr. Parkinson, I need say no more than that my conclusions accord 

 with his, and that I regard the ordinary ' graphic ' or ' pegmatitic ' 

 structure, whether on a minute or a large scale, as the result of a 

 struggle for independent crystallization between two minerals 

 (commonly felspar and quartz), one of Avhich has gained a very slight 

 advantage over the other in freezing. The crystals thus formed are 

 skeleton-crystals, the intervening parts being occupied by a more or 

 less continuous definite mineral, instead of by magma or aggregates 

 such as iron-oxides ; but we sometimes find that the one mineral, 

 either in the outer part of a spherulite or a pegmatite, assumes (b) 

 the curvilinear or a root-like growth. This is also a result of 

 obstruction, but of a slightly-different kind ; and its history I think 

 can be inferred from a remarkable and suggestive experiment in 

 the formation of colloid silica, described several years ago by 

 Messrs. J. I' Anson & E. A. Pankhurst. 3 A certain amount of an 

 alkaline carbonate was mixed with a strong solution of an alkaline 

 silicate, and then some strong sulphuric acid was slowly discharged 

 from a pipette at the bottom of the vessel containing the liquid. 

 Bubbles of carbonic-acid gas formed immediately and rose upward, 

 carrying with them some of the other acid. This on its part decom- 

 posed the alkaline silicate, causing precipitation of the silica, so that 

 in a few minutes a tube of it was formed, reaching from the bottom 

 to the surface of the solution. Its walls at first were very thin, but 

 as the acid percolated through them the process of decomposition and 

 deposition was maintained, and it continued so long as the agent 

 was supplied, thus forming a hollow ' stalactite.' These stalactites, 

 to quote the authors' words, 



' do not grow up by any means in constantly straight regular forms, but 

 assume irregular and branched ones, more like those of coral than anything 

 else, according to the direction in which the bubbles of gas or the acid escape 

 from the end, or from points of least resistance in the sides, of the tube.' 4 



I can see no other explanation of this wavy structure than a slightly- 

 variable opposition to the passage of the disturbing agent ; and thus 

 regard the coralloid of root-like structure of a mineral in a rock as 



1 Though the two minerals are often hardly to be distinguished under the 

 microscope, the published analyses of spherulites show that free quartz must be 

 present as well as felspar. 



2 So far as my observations go, this is more usual in artificial glasses. 



3 Min. Mag. vol. v (1884) p. 34. 



4 Op. cit. p. 36. 



