242 CARNEGIE INSTITUTION OF WASHINGTON. 



and hence of lower melting-points, they carried with them a certain 

 surplus of heat above that required to maintain their liquidity in the 

 new horizon, and this surplus was available for melting or fluxing 

 their way. They were at the same time, however, subject to loss of 

 heat by contact with surrounding rock of lower temperature. They 

 were thus probably at the same time taking up fusible material met 

 in their path and depositing old material as it became less adapted to 

 remain fluid under the new conditions, either because it had reached 

 the point of its saturation in the mixed rock solution that had been 

 developed or had cooled to its point of congelation. The liquid 

 thread was thus presumably taking on and giving up material con- 

 tinually as it worked its way outward, the process always being 

 selective and involving the retention of the more soluble or more 

 fusible portions and the rejection of the less soluble or more refrac- 

 tory portions. Since the included gases may be safely reckoned 

 with the former class, there was a selective accumulation of these, 

 and the ascending liquid became densely charged with them. To 

 this ascensive process those substances whose weight overbalanced 

 the differential pressure, such as metallic iron and possibly the 

 heaviest silicates, may be regarded as forming exceptions. 



Theory does not require that these threads should all succeed in 

 reaching the surface ; indeed, it does not require that an}' should in 

 the initial stages, before compression had developed a great excess of 

 heat in the central parts. The molten threads should simply rise 

 until their excess of heat, their working capital, was exhausted, 

 when they would return to the solid state and constitute tongue-like 

 intrusions. In doing this they would contribute heat to the tracts 

 which they invaded. This, in addition to conduction, was a mode 

 of conveying the intenser heat of the compressed central regions to 

 the higher horizons, where the original temperature was lower and 

 the fusing-points lower. The failure of the earlier threads to reach 

 the surface would thus be a means preparatory to the greater suc- 

 cess of later ones. The conditions for penetration would probably 

 be favorable up to the horizon where the temperature ceased to be 

 higher than the surface melting-point. Below this the retention of 

 the solid state was wholly due to pressure, the temperature being 

 above the surface melting-point. When the threads reached the 

 higher zone, in which the temperature was appreciably below the 

 surface fusing-point, the conditions were clearly adverse, and fur- 

 ther ascent was dependent on a sufficient excess of heat brought 

 from below to maintain the liquid state while this adverse tract was 



