6 , JOURNAL OF GEOLOGY—SUPPLEMENT 



the bottom of the crucible. The analyses showed a considerably 

 larger percentage of silica in the upper portion, but it was also 

 noted that there were grains of undissolved silica floating in the 

 upper part. With siKca in process of solution near the top, it would 

 be a marvel indeed if the liquid in the upper part was not more 

 siliceous than that in the lower, when the extreme viscosity of the 

 liquid, and the consequent impossibility of diffusion and convection 

 of sufficient vigor to maintain uniformity, are considered. Moroze- 

 wicz' figures show also that there was a greater concentration in the 

 lower portion, as compared with the upper portion, of. iron oxide 

 and the alkaline-earth oxides and a lesser concentration of the 

 alkalis. The natural explanation of this is that some of the alkaline 

 feldspar collected in the upper portion in exactly the same manner 

 as the grains of quartz. After its solution, which for the alkaline 

 feldspar happened to be complete, apparently, diffusion and con- 

 vection were unable to establish uniformity of composition in the 

 liquid. It should be noted, moreover, that the percentage of iron 

 oxide and of lime was greater in both the lower and the upper parts 

 than in the original granite. We must, therefore, add to the com- 

 plex conditions pointed out above a contamination from the pot or 

 furnace, at least from some outside source. Clearly no significance 

 can be attached to the experiment in the matter of indicating" 

 the effect of gravity in producing composition differences in a 

 liquid. 



The rather rapid melting of a mixture of minerals such as a 

 granite cannot be a simple process. It begins with the Kquefaction 

 of those substances which flux most readily together, not necessarily 

 those of lowest melting-point. The substances first melted are 

 principally those rich in alkahne earths and iron. The liquid, as 

 gradually formed, tends to filter down through the porous material 

 and collect toward the bottom. Thus there results in the early 

 stages a relative concentration of alkalic and siliceous materials 

 toward the top. The more siliceous and alkalic substances, which 

 assume the liquid state with great reluctance, dissolve slowly, and 

 in time all becomes liquid. The melt will, however, be layered 

 as a result of the accumulation of the first-formed liquid at the 

 bottom and possibly the floating of crystal grains buoyed up by 



