Lewinsoii-Lessiiig' s Classification of Rocks. 367 
render it viscous and difficultly fusible. According to Vogt a 
certain percentage of Al^O-i in the magma obstructs crystalli- 
zation. Lewison-Lessing says that the peridotic magma 
\vith seven to ten per cent, of AlgOg crystallizes with a great 
quantity of glass, while magmas without alumina have no glass 
on consolidation. 
Resume. Dififerentiation takes place, according to Lewin- 
son-Lessing, in the following manner: The igneous magma 
mixes with a second magma, or assimilates rocks which it 
encounters in its passage. In this way it becomes rich or sat- 
uiated with one or more component patts. The hisfher the 
temperature the greater the liquidity and the more abnormal 
the composition may become. Distinct differentiation tliei 
takes place only by the lowering of the temperature. When 
the temperature approaches that of solidification as the mag- 
ma becomes viscous, chemical afifinity begins to play a domi- 
nant role. In this magma the bases share the SiO^ between 
them, according to their respective af^nities. At this moment 
the minerals are already formed in the state of silicates which 
will crystallize or will form in glass, on eruption, according to 
the conditions of crystallization. This last moment before sol- 
idification appears so important to the author that he proposes 
to distinguish three kinds of dififerentiation, viz: (i) Magmatic 
differentiation, or static, taking place in the depths of the 
earth (dans la profondeur). (2) Dififerentiation by cooling, 
during the ascent to the surface (a little before solidification). 
(3^ Crystalline dififerentiation. 
In the first the principal factors are, specific gravity, tem- 
perature, pressure; while the second takes place as liquation 
under the influence of assimilated inclusions, and during the 
third phase the principal role is played by the chemical affini- 
ties. 
Lewinson-Lessing rejects in a few lines, the principle of 
Soret as applied to magmas, and presents some considerations 
UDon the application of the rule of Gibbs of phases in the phe- 
nomena of dififerentiation, reaching the conclusion that the 
minerals formed cannot be heterogeneous and indefinite but 
that thev are determined a i)riori; and rests on the "Grenge- 
lialt" in saturated solutions, etc. These considerations, occu- 
pying several pages, are rather su])erficiallv treated, and are 
omitted here. 
