22 Johnson and Warren— Geology of Rhode Island. 



was wholly a part of the act of crystallization. Whether the 

 olivine preceded the ore in time of crystallization or whether 

 the two crystallized at the same time, we have to explain how 

 a material in the act of orientating its molecules can forcibly 

 expel another and move it through a molecularly great dis- 

 tance. If the expulsion is a part of the act of crystallization 

 then it must have begun about a great number of irregularly 

 situated centers and proceeded gradually outward, ihe growth 

 of the one mineral (in the present case the olivine) forcing the 

 other toward the surrounding centers until finally the expelled 

 material from the many centers would necessarily occupy an. 

 interstitial position with reference to the first mineral. If the 

 interstitial material were still liquid it would then solidify 

 when its own freezing point was reached. The position and 

 character of the ore is quite in keeping with such a procedure. 

 Another supposition which is possible is that the minerals 

 become immiscible before their freezing point is reached 

 and being then in a relatively mobile condition they would 

 segregate readily, and subsequently solidify independently, 

 their textural relations to each other being in such case deter- 

 mined by their individual freezing points, their relative veloc- 

 ity and power of crystallization, and their relative volumes. 

 The inclusion of ore grains, as well as that of the minute 

 opaque particles, a part of which may reasonably be considered 

 as primary crystallizations, within the olivine crystals and 

 along their boundaries, is quite in keeping with either of the 

 above suppositions. The inclusions are merely a portion of the 

 ore which was prevented from segregating with the rest by 

 the decreasing mobility of the crystallizing olivine. 



This last supposition involves the idea of liquid immiscibil- 

 ity in silicate melts. This has been proposed by some to explain 

 rock differentiation,* but there appears to be no convincing 

 evidence, certainly not experimental, f of immiscibility in molten 

 silicate mixtures. Nevertheless, the mechanical difficulties in- 

 volved in the segregation of the various minerals into the posi- 

 tions occupied by them in the rhodose (as well as of the minerals 

 in many other rocks) seem to be considerably less if liquid 

 immiscibility can obtain, than would be the case if the separa- 

 tion took place during cn'stallization, at which time one would 

 expect, in most instances, great resistance to freedom of motion 

 and the minerals might be expected to crystallize as a fine- 

 grained, intimately mingled aggregate such as is the case with 

 eutectic mixtures. The idea of liquid immiscibility seems at 

 least worthy of careful consideration. 



*See H. Backstroni, Jour, of Geol., vol. i, 1893, p. 778. 

 ■j-See Vogt, Die Silicatschnielzlosungen, vol. i, p. 100, vol. ii, p. 228, and 

 A. L. Day and E. S. Shepard, Econ. Geol., vol. i, 1905, p. 286. 



