March 12, 1885] 



NATURE 



447 



igous to those « ' 

 mcentrated aqueous solutions. Cooling 

 nation of the different minerals onl; 

 tin solubility of the constituents of the minerals in 

 -magma. The point at which a mineral forms from 

 ii.]-, then, no more connection with its fusibility 

 than the point at which graphite forms in molten iron has with 

 ibility. 



■ of very great importance is this: the differ- 

 entiation of crystals in an originally homogeneous magm i mu ;l 

 impanied by a variation in the composition of 

 1 :s, then, of great interest to determine the 



general order of the formation of crystals in igneous magmas. 

 ' we have a most valuable and suggestive paper 



nlitled " Ueber das Wesen der kbmigen und 

 ischen St> uctur bei Massengestein 

 1882, ii. p. I). Before proceeding to give an account of the 

 portion of this paper which bears more particularly on the sub- 

 ject we are now discussing, it may be well to call attention to 

 the methods available for the purpo e of determining the order 

 of the f irmation of the minerals in a rock. There are two. In the 

 first place we may observe the phenomena of inclusions, and in 

 the second place we may observe the extent to which crystalline 

 form his been developed. If one mineral is seen to be included 

 in another, then we may safely infer, subject to certain precau- 

 tions, that the included mineral is the earlier of the two, and if 

 one mineral shows a more perfect form than another witli which 

 iated, then we may infer — again subject to certain pre- 

 cautions — that the mineral with the more perfect form is the 

 earlier. 



Now in the paper above referred to, Rosenbusch divides the 

 constituents of igneous rocks into four groups : — 



(1) The ores and accessory constituents (magnetite, hematite, 

 •lmenite, apatite, zircon, spinel, sphene). 



(21 The ferro-magnesian silicates (biotite, amphibole, pyro- 

 xene, olivine). 



(3) The felspathic constituents (felspar, nepheline, leucite, 

 melilite, sodalite, haiiyn). 



(4) Free quartz. 



He then calls attention to the contrast which is presented by 

 the granites and syenites on the one hand, and the diabases on 

 the other. In the former the following law is adhered to with a 

 very great amount of persistence : — The order of formation is 

 that of increasing basicity : the ores and accessory constituents 

 are first formed, and the quartz is the final product of consolida- 

 tion. In the diabases and gabbros there is apparently an excep- 

 tion to this law of increasing basicity, the augite consolidating 

 after the felspar. Rosenbusch proposes to divide the granular 

 holo-crystalline rocks into two classes: (i) those in which the 

 minerals of the 2nd group in the above classification consolidate 

 he 3rd, and (2) those in which the reverse rela- 

 tion hold-. He then calls attention to cases illustrative of the 

 law of increasing basicity which are furnished by the order of 

 separation in the individual groups. Thus in the ferro-magnesian 

 group, olivine is older than biotite, amphibole and pyroxene ; 

 ite is older than the bisilicates. In the felspathic group 

 triclinic felspars are older than monoclinic [there are exceptions 

 to this rule, as, for instance, in the porphyroid of Maim 111 the 

 Ardennes, where orthoclase crystals are seen to be surrounded 

 by a narrow zone of oligoclase], and the basic triclinic felspars 

 are older than those which contain a large percentage of silica. 



The views of Rosenbusch are based on the assumption that 

 the order of formation of crystals in igneous magmas is deter- 

 min d solely by chemical conditions. That these conditions are 

 the more potent seems quite clear, but there are fa 

 appear to show that physical conditions are not altogether without 

 influence on the result. 



The law of incre ising 1 1 1 deity may be accepted without hesita- 

 pressing in a general way the truth as , 

 order of separation of the different constituents of igneous 

 rocks. 



Now a very interesting conclusion follows as a natural con- 

 sequence of this law. The effect of progres, 

 in a magma must be to increase the percenl 

 decrease the amount of lime, iron, and magnesia, 1 



1 mount of alkalies, and to increase the potash relatively 



it which remain- liquid. It is alwa)s 



satisfactory to find independent evidence confirmator) ol anj 



conclusion at which one may have arrive I. Now I think we 



nmatory evidence of this kind in the present case. It 



will be admitted on all hands that the crystals in porphyritic 

 rocks, such as hypersthene-andesite, have been formed in a 

 magma the composition of which is represented by the bulk 

 analysis of the rock. If, then, we compare the bulk analysis 

 with the analysis of the ground-mass deprived of its crystals, 

 we ought to find confirmation of the above conclusion. 



Dr. Petersen has isolated and analysed the ground-masses of 

 two of the Cheviot porphyritic rocks, and by comparing these 

 with the bulk analyses of the rock the truth of the conclusion 

 is most strikingly illustrated. The effect of progressive crystal- 

 lisation in the andesitic magma has led unquestionably to an 

 increase in the amount of silica, a decrease in the amounts of 

 lime, iron, and magnesia, an increase in the amount of alkalies 

 generally, and an increase in the potash relatively to the soda. 

 In the rock itself soda is in excess of potash ; in the ground- 

 mass p itash is in excess of soda. 



There is yet another piece of independent confirmatory evi- 

 dence. Every geologist is familiar with the phenomenon of con- 

 temporaneous veins. The general view held with regard to 

 them is that they represent portions of material which remained 

 fluid after consolidation had progressed to a considerable extent. 

 If this view be correct, then they should hold the same chemi- 

 cal relation to the main mass of the rock as the ground-mass of 

 the Cheviot andesite does to the main mass of the andesite. 

 Mr. Waller has recently analysed certain contemporaneous veins 

 which occur in the bronzite-diabase of PenmEenmawr. He finds 

 that they contain about 7 per cent, more silica than the normal 

 rock, less lime and magnesia, more alkalies, and more potash 

 than soda, although in the normal rock soda is in excess. Con- 

 temporaneous veins in the Rowley rag dolerite have also been 

 investigated by Mr. Waller, with the same result as far as in- 

 crease in silica and total alkalies is concerned. The relation of 

 potash to soda has not yet been determined. 



I believe it is admitted to be a general rule that contempora- 

 neous vins contain more silica than the rock with which they 

 are associated. It will be seen that there is abundant evidence 

 of an independent character to confirm the general truth of the 

 conclusion which follows from a consideration of the facts 

 brought forward by Rosenbusch. 



I should not have treated this subject at such length did it 

 not appear to have an important bearing on the origin and 

 sequence of volcanic rocks. I can best explain this by referring 

 to the Cheviot district, with which I am slightly acquainted. 



Andesitic lavas and tuffs cover large tracts of this district. 

 These are unquestionably the products of surface volcanic action. 

 In the central portion of the volcanic area there is a mass of 

 augitic granite. A consideration of the mineralogical compo- 

 sition of this granite shows that it cannot belong to the acid 

 group of rocks, and this conclusion is confirmed by an examina- 

 tion of the chemical composition of allied rocks from the Vosges. 

 So far as we can judge in the absence of analyses there appears 

 to be a very close connection between the composition of the 

 plutonic and that of the volcanic rocks of the Cheviot district, 

 and we therefore seem justified in concluding that the plutonic 

 masses differ in character from the andesitic lavas merely in con- 

 sequence of differences in the conditions of consolidation. The 

 plutonic rocks represent the consolidation of the andesitic 

 magma beneath the surface, and therefore under great pressure ; 

 the lavas and tufts represent the consolidation of the same 

 magma at the surface. 



I now come to the point to which I wish to direct special 

 attention. The andesitic lavas and tuffs are traversed by quartz- 

 felsite dykes in such a way as to show that a magma of rhyolitic 

 composition must have been erupted by the Cheviot volcanoes 

 subsequently to the period characterised by the eruption of the 

 andesitic magma. Contemporaneous veins similar in character 

 to the quartz-felsite dykes also occur in the plutonic rocks. 

 Again, an analysis of one of the quartz-felsite dykes by Mr. 

 Waller agrees almost exactly with the analyses of the ground- 

 mass of the hypersthene-andesite by Dr. Petersen. 



Putting all these facts together, we conclude that the eruption 

 of an andesitic magma was followed, in the history of the Cheviot 

 volcanoes, by that of a rhyolitic magna in consequence of pro- 

 gressive crystallisation in the deep-seated plutonic source. The 

 rhyolitic magma is, so to speak, the mother liquor out of which 

 varipu basic minerals have crystallised. Suppose a half-con- 

 solidated plutonic mass, originally of andesitic composition, to 

 dijected to a powerful crush such as that which un- 

 questionably arises in the earth's crust under certain circum- 



o; The mother liquor will be squeezed out of the mass, 



