121 
IDDINGS. 
the formation of iron ore deposits in Norway and Sweden, 
a part of which he refers to u magmatic concentration ” in 
strongly basic eruptive rocks. The bearing of this principle 
on the question of the variation of igneous rocks is pointed 
out by Teall, to whom we are indebted for a review' of this 
important work from which the present notice is taken. 
After discussing the mineralogical character of the iron 
ores in question and that of the rocks with which they are 
associated, he concludes that they have resulted from a con¬ 
centration in the plutonic magmas for the following reasons: 
(1.) The different ores stand petrographically related to 
the surrounding rocks. 
(2.) The ore in many cases passes gradually into the sur¬ 
rounding rock. 
(3.) In no case is there any evidence of the introduction, 
of material by solutions. The segregations are character¬ 
ized by the minerals of the surrounding rocks, and by these 
alone. 
As to the process by which this concentration has taken 
place, he is of the opinion that it was by molecular diffusion 
without the actual separation of minerals. 
“The homogeneity of a solution may be destroyed by 
temperature differences and by gravity. The influence of 
the former has been experimentally established by Soret, 
and follows as a necessary consequence of van’t Hoff’s 
theorem that osmotic pressure in the case of dilute solutions 
obeys the laws of gaseous pressure. The influence of the 
latter has been deduced experimentally by Gouy and Cha¬ 
peron from the laws of thermo-dynamics. Where solutions 
become heavier by concentration the lower part will be 
more concentrated than the upper part. The difference is 
slight and can only be recognized with difficulty when a 
tube 100 meters high is used. The specific gravity of a 
molten magma will increase with an increase in the num¬ 
ber of molecules of magnetite, ilmenite, magnesia, iron- 
silicates, pyrite, etc. Hence the molecules will, according 
