410 H. S. WASHINGTON — IGNEOUS COMPLEX OF MAGNET COVE 



pure condition, as foyaite, and gradually becoming more basic (richer 

 in the .solute) as the freezing process crept toward the center. 



Assuming that a certain portion of the magma acts as solvent and 

 another as solute, as it seems we must, such a process is a natural one. 

 It has been advocated by Becker in his paper on Fractional Crystalliza- 

 tion,* but in a somewhat different form. According to him the " pro- 

 cess depends essentially upon convection currents," and the crystalliza- 

 tion proceeds from the least to the most fusible constituents. 



On these two points I cannot agree with him. While convection cur- 

 rents would undoubtedly be set up to some extent, and have been ap- 

 parently in the outer syenitic zones at Magnet Cove, yet they do not seem 

 to me to be essential to the process. It would go on by a simple crys- 

 tallization of the solvent, thus collecting along the rough borders in 

 accordance with the well known tendency of crystallizing bodies to grow 

 about sharp nuclei, the solute molecules being mechanically pushed aside 

 toward the center. t 



That it is not the least fusible substance which crystallizes first is 

 shown by the general truth of Rosenbusch's law of the order of crystal- 

 lization and the numerous dikes and laccoliths with basic and more 

 fusible borders and acid centers. This is not a question of relative fusi- 

 bility, but of solvent and solute, degree of dilution and relative solubilit} 7- . 



Different types of laccoliths. — While the conditions are complex, and sev- 

 eral possible processes may be involved, yet there is reason for thinking 

 that the composition of the magma as a whole has most to do with the 

 order of arrangement of the ciystallized differentiates. If we examine 

 the three classes of laccoliths already mentioned, it is seen that, while in 

 one (Henry Mountain type) the magma is undifferentiated, in the second 

 (Square Butte type) it is differentiated and with a basic border, and in 

 the third (Magnet Cove type) it is differentiated, but with the border 

 acid ; yet the conditions of cooling and solidification could not, as far as 

 we can tell, have been very different in the different cases. 



I have already pointed out that the three types seem to be distinguish- 

 able by the chemical characters of their respective magmas. 



The first type of laccolith is granitic or dacitic — that is, very rich in 

 Si0 2 , and with relatively small amounts of all other oxides, either alka- 

 lies or lime preponderating among these. It may also be noted that the 

 structure of these is always porphyriticj The second is either mon- 

 zonitic or dioritic, referring to the general character of the magma as a 



*<;. F. Becker: Am. Jour. Sci., vol. iv, 1897, p. 257. 



f Examples of this on a small scale are to be seen about the phenocrysts of certain dike rocks. 

 Cf. Pirsson : Am. Jour. Sci., vol. ii, 1896, p. 191 ; and Washington : Jour, of Geology, vol. vii, 1899, 

 p. 113. 



JCf. Pirsson : Twentieth Ann. Rept. U. S. Gcol. Survey, L9Q0, pt, iii, p. 561, 



