92 F. LOEWINSON-LESSING 



solved salts in the lower parts. Direct experiments are wanting; 

 the oceans do not seem to support this assertion. Should this 

 phenomenon really exist, it would assume long periods of rest in 

 the magma, undisturbed by convectional and other currents. Let 

 us not further insist upon this hypothesis. 



Let us now examine the remaining appHcation of gravitative 

 adjustment in the liquid magma, namely the gravitational distribu- 

 tion of the minerals during crystallization. Here again we may 

 assume two cases: (i) the distribution, according to their densities 

 of those minerals which are the definite components of the growing 

 igneous rocks, i.e., the components corresponding to- a full equi- 

 librium; (2) the sinking or floating of the first products of crystaUi- 

 zation which may issue from the dissociation of the final products. 

 These minerals correspond to an unstable and temporary equi- 

 Ubrium. If they are not removed by sinking or floating, they are 

 redissolved and replaced by those definite components from whose 

 dissociation they were derived. This is Bowen's view. It is 

 necessary to analyze each of these two cases separately. 



I . Examples of such a gravitational distribution of phenocrysts 

 in lavas, diabases, and nephelite syenites are weU known, for 

 instance, two facies of the Vesuvius lavas, one rich in leucite, the 

 other in augite; the more pyroxenic and the more feldspathic facies 

 of the Olonetz (Petrosawodsk) diabase studied by the author ;'' 

 and the nepheHte-sodahte- and nephelite-eudaUte-syenites of Ussing. 

 It is only natural to attribute to this process the melanocratic 

 facies of certain laccoliths or intrusive sheets, the granophyric upper 

 layer of certain gabbors, the variohtic structure of certain diabase 

 masses, and so forth. Gravitational differentiation understood in 

 this sense certainly reaUy exists. More than that, I think that 

 Schweig is right in assuming that the early crystallized minerals, 

 sinking into the deeper parts of the magma where the temperature 

 may be higher, can be dissolved. In this way certain components 

 of the magma may be transported by the aid of a temporary solid 

 phase into the deeper horizons, and finally there may result two 

 different layers and the formation of a batholith or a laccolith con- 



' F. Loewinson-Lessing, "The Diabase Formation of Olonetz," Trav. d. I. Soc. d 

 Natur. St. Petersh., 1888. 



