30 PROCEEDINGS OF THE NATIOMAIv MUSEUM. vol.60. 



sinking of early olivine crystals which went to form a layer of oliv- 

 ine diabase near the bottom of the sill." His discussion of the dif- 

 ferentiation of this sill is very instructive and may well in large part 

 be quoted, as follows: 



Inasmuch as the several tjpes of rocks described above occur as continuous 

 portions of a single intrusive sill, they must be regarded together as constituting 

 a unit. There is no evidence that they are products to any extent whatever of 

 separate intrusions, or even of successive pulsations of an extended period of 

 injection. Their present constitution and relations are best understood as the 

 results of differentiation, or separation of the constituents of the molten magma 

 after its intrusion and during the long period required for cooling and solidifi- 

 cation. 



The thickness of the sill or intrusive sheet varies considerably in its 100 miles 

 of outcrop in New York and New Jersey, but it is everywhere several hundred 

 feet thick, and in places, as along the Palisades above Weehawken, and in the 

 thicker parts of Rocky Hill and Sourland Mountain, it approximates 1,000 feet. 

 Under cover of a great blanket of overlying shales and sandstones, probably 

 many times its ov\'n thickness at the time of intrusion, though since partly re- 

 moved by erosion, this highly-heated molten magma cooled very slowly, and 

 probably remained in a liquid condition for a considerable period. The only 

 exceptions to this are the immediate contacts with the inclosing strata, which 

 must have been quickly chilled; on the other hand, the adjacent shales and sand- 

 stones themselves became highly heated, and subsequent cooling was probably 

 slow. The surrounding rocks are poor conductors of heat, and once a crust had 

 formed, and the strata at the contact were well heated, the inclosed liquid mass 

 became in a measure insulated. Under such conditions the outer crust of the 

 magma would slowly thicken until the whole mass became solid. 



Professor Iddings' conclusion that the process differentiation which gives 

 rise to variations in the character of different parts of such a magma "must be 

 of a chemico-physical nature; that is, a chemical process resulting from varying 

 physical conditions, especially temperatures,'' is doubtless true in most cases and 

 probabl}' to some extent in all, but in the present state of our knowledge, it 

 seems scarcely justifiable to exclude entirely the possibility of purely physical 

 processes acting alone. This applies particularly to the settling of heavier 

 crystals in the more basic magmas, which are highly fluid, and might well 

 remain so long enough for such a process to produce considerable effect. In 

 fact, the extent of such gravitation of the heavier minerals may be regarded as 

 a measure of the degree and duration of the liquidity after the beginning of 

 crystallization, and the absence of such effects only as evidence that the partic- 

 ular magma has become too viscous to permit effective differentiation from 

 this cause. 



Further, the time of crystallization of a particular mineral is held to have 

 some definite relation to its concentration in the solution, and this seems to 

 imply that the definite molecular group exists as the point of saturation is 

 approached, ready to crystallize when that point is reached. In acid magmas 

 the proportion of basic constituents is small, and saturation would occur only at 

 a correspondingly lower temperature than in those basaltic magmas which carry 

 basic substances in large amounts. Hence the crystallization of magnetite and 

 augite in rhyolite, for example, v/ould probably not take place before the whole 

 magma has cooled to a highly viscous condition, particularly as this condition 



"J. Volney Lewis. Petrography of the Newark igneous rocks of New Jersey, Ann. Rept. State Geol- 

 ogist N. J. for 1907, pp. 129-133. 



