168 



KNOWLEDGE. 



May, 1911. 



substance xsill interfere with each other's crystalH- 

 sation, so that we shall get a mass in which the 

 component crystals are very intimately intergrow n. 



Let us now turn from allo\s to the applications 

 of this theor5' to the crvstallisation of rocks. 



Few rocks have been svnthesised, but nian\- rock- 

 forming minerals have been crystallised from slags, 

 and Vogt has successfulh" tackled the determination 

 of some of their eutectics. Thus he has found that 

 68 parts of augite to 32 of oli\inc 

 form a eutectic, and so do 70 

 parts of plagioclase felspar toiO ot 

 olivine. .\nd this has been 

 proved to hold in the igneous '-; 

 rocks, for Harker has she\\n that t 

 in allivalite (a rock composed of 

 olivine and the lime-bearing 

 plagioclase vanorthite) the anor- 

 thite al\\a\"scrystallised first when 

 constituting more than 70 per 

 cent, of the rock, otherwise the 

 olivine crystallised first. 



In LS88, Teall suggested that 

 the curious intergrowth of quartz 

 and felspar known as micropeg- 

 matite was a eutectic, and \'ogt 

 shewed from a chemical analysis 

 that the ratios of quartz to 

 felspar in micropegmatite was 

 26 to 74. 



Hitherto we have been restrict- 

 ing ourselves to substances which 

 do not shew any crsstallographic 

 relationship. W'e must now 

 examine the melting-point curves 

 of isomorphous mixtures an< 

 touch upon the in\'estigations o 

 Bakhuis Koozeboom. He found 

 that the curves fell into five 

 groups, of which oid\' two neec 

 at present be examined. Taking 

 the enstatite-hypersthene series 

 of isomorphous metasilicates, 

 where crx'stals of ever\' possible 

 composition can be obtained, 

 he found that instead of one 

 he obtained two curves, the 

 freezing-point cur\e or liqiiidtis, 

 and the melting-point cur\e 

 or solidiis. Thus a liquid of 

 the conqiosition A gi\'es rise 

 to crystals of the composition 

 B, since crystals of the com- 

 position A would have melted 

 ature. This curve (shewn in Fig 

 his Type L Instances, however. 



FiCl'KI, 0. 1- 



XOlJZOtH 



at 



this temper- 

 re 5) constitutes 

 are frequent of 

 imperfect isomorphism in w hich one constituent. X. 

 will only hold a certain amount of the other, Y, just 

 as ether and water will only mix in certain 

 proportions. This imperfect isomorphism gives 

 rise to Type IV, shewn in Figure 6, where the upper 

 curve, the liqiiidus, is seen to have a eutectic point 



formed not of the substances Init of the mixed 

 crxstals A and 15. each of which contains both 

 constituents. This has a very interesting application 

 in the case of perthite, a peculiar intergrowth of 

 orthoclase and plagioclase. Orthoclase may contain 

 uj) to a certain percentage of albite, being then 

 known as soda orthoclase, whilst albite may likewise 

 contain a limited amount of orthoclase. This is 

 because ortlioclase being monoclinic and plagioclase 

 triclinic, perfect isomorphism is 

 impossible; nc\-ertheless the re- 

 3»6'- lationship is both chemicall_\" and 

 crvstallographically so close that 

 small amounts of one can form 

 mixed crvstals with the other. 



The importance of eutectics 

 in rock formation is ver}- great 

 and the conditions are often 

 very complicated, there being 

 frequently several successive 

 eutectics between different pairs 

 of minerals. A condensed 

 account of the crystallisation of 

 a granophyre ma\- give some idea 

 of the role of the eutectic. The 

 rock to be described consists of 

 f e r ro m ag n esi a n m e t as i 1 i c a t e s 

 which crvstallise out as augite, 

 alumino-alkali anhydrosilicates 

 wliich form felspar, and quartz. 

 The augite first separates out in 

 well-formed crystals ; then the 

 felspar builds large more or less 

 perfect crxstals termed pheno- 

 crx'sts, until the eutectic compos- 

 ition is reached, when the 

 fels()ar and quartz remaining 

 solidif\- as luicropegmatite, \\hose 

 felspar is in optical continuity 

 w ith the phenocrysts. 



Besides these micropegmatitic 

 or granophyric rocks, there are 

 manv whose ground-mass is too 

 fine to be resolved b\' the micro- 

 scope, but has been shewn chemi- 

 cally to be nearly eutectic in 

 composition. S[)herulites of 

 niiiuite radiating fibres are found 

 ni them, and these have been 

 shewn in some cases to consist 

 of micropegmatite. Hence this 

 grouiid-niass has been supposed 

 to represent the eutectic on a 

 scale, and is termed amphi-eutectic. 



Type I\', 



cr\'[itocrystalline 



that is, almost eutectic. 



It only remains to add that a classification of 

 rocks has been seriously proposed on the nature of 

 their eutectics. Its advocates regard the eutectic as 

 the dominant feature of the rock. It is being 

 \-igorously opposed by the school of petrologists who 

 place chemical composition first; but time alone can 

 decide between them. 



