RESEARCHES IN THERMAL METAMORPHISM. 193 



■due to complete recrystallisation of the rock-mass : this is 

 what Beck terms "typical contact-structure". 



Numerous instances are known in which chiastolite is 

 produced in relatively large crystals in the early stages of 

 metamorphism, but gives place in the more altered rocks to 

 andalusite, usually in smaller and less perfect crystals. 

 Again, sillimanite or fibrolite, dimorphous with andalusite, 

 occurs abundantly in some very highly metamorphosed 

 slates, flags and grits. It is embedded as bundles of 

 parallel or sub-parallel needles in grains and patches of 

 quartz, to which it imparts a peculiar opacity and silky 

 lustre ("quartz sillimanitise " or Faserkiesel of some con- 

 tinental geologists), or it occurs in similar fashion in the 

 interior of flakes of muscovite or sometimes of biotite. It 

 has been described by Barrois in Brittany, and also occurs 

 in abundance in the highly metamorphosed rocks described 

 by Miss Gardiner (1) near New Galloway and by Barrow 

 in Forfarshire (2). In the latter area occurs still another 

 form of the pure silicate of alumina, viz., disthene or cyanite. 



Such facts as these raise an interesting question : Can 

 we to any extent specify the conditions which determine 

 what particular aluminous silicate shall be formed in a given 

 case ? The minerals produced during metamorphism must, 

 of course, depend in the first place upon the nature of the 

 rocks affected ; but in the group of minerals considered 

 some (chiastolite, andalusite, cyanite, sillimanite) have identi- 

 cally the same chemical composition, and some others do 

 not differ very greatly from one another in this respect. 

 We are therefore led to infer that physical conditions, and 

 especially the temperature, attained during the metamor- 

 phism, must be a considerable factor in determining the 

 formation of one or other of the allied minerals. Barrow's 

 work in Forfarshire has an important bearing on this 

 question. There can be little doubt that the striking features 

 of that area are due, in the main, to thermal metamorphism 

 on something approaching a regional scale, but the pheno- 

 mena of extreme metamorphism are localised in connec- 

 tion with certain intrusions of gneiss and pegmatite. In 

 the altered rocks, which have undoubtedly once been 



