HYPOTHETICAL STAGES LEADING UP TO THE KNOWN ERAS. 121 



basic rocks in marine volcanoes. It is not to be lightly assumed, however, that 

 differentiation may not develop acidic rocks from even the prevailingly basic 

 magmas of the sub-oceanic segments. Volcanoes that arise from submerged 

 ridges and platforms that were once land, are obviously not to be regarded as 

 properly abysmal. 



Obstacles to minute differentiation. — While in most respects the conditions 

 postulated by the hypothesis were extremely favorable to the broader classes of 

 differentiation, there were certain respects in which complete differentiation 

 was hampered. One of the most notable of these was the apparent lack of a high 

 degree of fluidity, and of a protracted condition of fluidity even of the more 

 viscous kind. By hypothesis, the fluent material was developed by selective 

 fusion from a mixed mass presenting all grades of fusibility, and hence there is 

 little reason to suppose that any portion would become very highly fluent while 

 surrounded by matter on the border of fluidity, or already viscous and ready to 

 mingle with it. Moreover, the hypothesis postulates that as soon as any appreci- 

 able portion became mobile, it was forced toward the surface by the internal stress- 

 differences, and in its outward passage was subject to loss of heat, as well as 

 accessions and losses of material. Under all these conditions it seems most 

 consistent to suppose that little more than viscous mobility — not perfect fluidity 

 — was acquired. The matter is presumed to have been in slow movement upward 

 from the time it acquired mobility, and the rising threads or tongues of lava 

 were probably also subject to slow convectional currents, if at any time they 

 reached a sufficient degree of fluidity to permit this. The portion next the cooler 

 walls obviously tended to descend, while that in the central part tended upwards. 

 Up to a certain point, a slow movement may be regarded as favorable to differen- 

 tiation, by bringing together matter of the appropriate kinds. For example, 

 it is probable that the growth of crystals is promoted by very gentle movement, 

 just as it is probable that snow-flakes grow more rapidly by falling gently through 

 moist air than they would if they remained in contact with the same air. But 

 at the same time, motion must oppose the separation of the crystals from the 

 moving mass, and by keeping them in it, lead to their final inclusion in the solid- 

 ified mass. Between the high degree of viscosity and the appreciable amount of 

 motion postulated, it appears that the individual crystals were not usually permit- 

 ted to separate from the mobile mass by settling according to their specific gravities, 

 and hence arose the complex mixture of crystals which is found in most igneous 

 rocks, and which is so difficult to understand, if the lavas were perfectly liquid and 

 remained quiescent in reservoirs while slow crystallization took place, for the 

 heavier minerals often crystallized first. As the planetesimal hypothesis does not 

 recognize any great reservoirs of molten material, standing for ages in a quiescent 

 condition, the foregoing considerations relative to viscosity and to motion seem to 

 meet the requirements of this puzzling phenomenon, at least in some degree. 

 The specific gravities of the minerals common to igneous rocks vary rather widely, 

 so much that the petrologist separates them readily by a series of liquids of 

 graded specific gravities. The fact that they are not thus separated in nature 

 in the case of rocks that have been lavas may prove a very significant phenomenon 



