298 ANISrUAL REFOFvT SMITHSONIAN INSTITUTION, 192 8 



Second. The asthenosphere is taken as occupying a deep zone, 

 possibly several hundred miles in depth, the upper part of which 

 has the composition of a diorite, and is decidedly more acidic than 

 the lower part, which probably has the composition of a gabbro or 

 basalt. The dioritic zone has appreciably lower density than the 

 basaltic one, quite apart from the effects of pressure. Most of the 

 igneous activity within the continents is thought to originate in the 

 dioritic zone, and this was notably true of the extensive fusion which 

 gave rise to the Archean granites. A dioritic magma has between 

 55 and 65 per cent of silica, and is therefore intermediate in compo- 

 sition between a granite (65 to 75 per cent silica) and a gabbro 

 (45 to 55 per cent silica). This intermediate magma, when it ap- 

 pears at the surface in molten form, gives rise to dark lavas known as 

 andesites, many of which approach basalts in appearance. During 

 the existence of the molten reservoir in the outer crust a separation 

 of the fluid rock tends to take place, possibly by an early crystalliza- 

 tion of the basic minerals, splitting the magma into basalts on the 

 one hand and granodiorites or granites on the other. The latter 

 accumulate especially in large bodies in the crust, and these are 

 exposed in after ages by erosion and are revealed as the cores of 

 mountain ranges. The processes of differentiation will therefore 

 produce all kinds of rocks from whatever depth in the crust they 

 come, but the relative volumes of the different kinds will be notably 

 different according to the original composition. Such a difference 

 in proportions is thought to distinguish the lavas of continents from 

 those of oceanic islands, the latter being more dominanth' basaltic. 

 The difference pointed out is probably to be ascribed to the original 

 depth from which the magma rises. It is true, however, that ex- 

 trusions and intrusions of dominantly basic nature are known within 

 the continents. When these rise to great volume they are thought 

 to have risen from a deeper source ; when occurring over broad areas 

 they set up strains which work toward continental fragmentation in 

 obedience to the laws governing isostasy. 



Third. The periodic cause of the generation of magmas is be- 

 lieved to be radioactivity. The radioactive elements are known to 

 be present in appreciable quantities in granites and basalts, so that 

 if these rocks exist below a depth of about 40 miles,^^ heat is gener- 

 ated by them faster than it can be conducted away, unless some 

 unknown inhibiting action is there present. If so generated, the heat 

 must slowly increase and express itself in the liquefaction of rock. 

 As the liquid rock increases in quantity it becomes able to work its 

 way to higher levels. Most of it will slowly cool there, but some 

 will melt its way to the surface. This convection effect will tend to 



1' This figure Is greatly reduced by recent estimates. See Editorial rrcface, with foot- 

 notes. 



