450 Andersen — System Anorthite-Forsterite-Silica. 



called arrested (" abgescblossene ") crystallization are explained 

 as a result of cooling with equilibrium between the solids and 

 the liquid.* In giving the latter explanation Yogt refers to a 

 number of possible cases in certain types of ternary systems 

 with solid solutions, discussed by Schreinemakers. H. Brandf 

 has pointed out how resorption of crystals as well as recurrence 

 of crystallization are possible in ternary systems without solid 

 solution. Other examples of mixtures of this kind are 

 described by Shepherd, Rankin and Wright,:): in the prelim- 

 inary paper on the . system CaO-Al a O a -SiO a , in which the 

 resorption of CaO is discussed and attention is called to the 

 bearing of this phenomenon on certain petrologic problems. 

 H. E. Boeke§ has again emphasized the importance of apply- 

 ing these principles to the explanation of the corresponding 

 phenomena in igneous rocks. 



The first examples of petrologic systems (that is, systems 

 containing rock-forming mineral compounds) in which mag- 

 matic resorption and recurrent crystallization have been shown 

 to be a simple result of the process of solidification, were de- 

 scribed from this laboratory. In the final work on the system 

 CaO-Al 2 3 -Si0 2 Rankin and Wright || describe actual mineral 

 compounds (for instance, A1 2 3 ) showing magmatic resorption 

 on cooling. In the paper on the MgO-Si0 2 T[ system it was 

 shown that the forsterite crystals in the forsterite-silica part 

 of the system were always partly and sometimes completely 

 redissolved during the cooling of melts containing this mineral 

 as primarj 7 phase, and it was indicated that this might explain 

 the resorption of natural olivine so frequently observed in 

 rocks. N. L. Bowen*"* described the same phenomenon of mag- 

 matic resorption of forsterite in the system diopside-forsterite- 

 silica, and showed how in certain mixtures there must be 

 recurrent crystallization of forsterite. 



From the data contained in the present paper it is obvious 

 that in mixtures of anorthite, forsterite and silica, the crystals 

 of forsterite will be more or less resorbed whenever they occur, 

 and it may happen that forsterite appears in the earlier stages 

 of crystallization but is completely dissolved before the final 

 solidification. 



It has been pointed out in a previous chapter (p. 435) what 

 relations should obtain in a system without solid solution in 

 order that recurrent crystallization of one of the phases may 



* J. H. L. Vogt, Tsch. Min. petr. Mitt., xxvii, 141, 1908. 



f Neues Jahrb. Min., Beil. Bd., xxxii, 650, 1911. 



% Journ. Ind. Eng. Chem., vol. iii, No. 4, 1911. 



§Centralbl. Min., 1912, 266. 



|f G. A. Bankin and F. E. Wright, this Journal (4),. xxxix, 1, 1915. 

 T[N. L. Bowen and Olaf Andersen, loc. cit., p. 495. 

 ** This Journal, xxxviii, 207, 1914. 



