From these and other examples which have been observed, it is difficult to escape the conclusion that the differences between the basic margin and the acid centre are due to some process of segregation or differentiation while the mass was still in a liquid condition, and its constituents could pass from one part of the boss to another. According to Professor Brögger, it may be stated as a general law that differentiation sets in during consolidation, and is determined by, and dependent on, the laws of crystallization in a magma, in so far as the compounds which, on given conditions, would first crystallize out, diffuse themselves towards the cooling margin so as to produce in the contact-stratum a peculiar chemical composition in the still liquid material before crystallization takes place.[41] [Footnote 41: This general conclusion is stated by Professor Brögger from his investigation of the rocks of Gran, _Quart. Journ. Geol. Soc._ l. (1894), p. 36.] If during the process of differentiation, and before consolidation, injections of the magma occur, they may be expected to differ in character according to the portion of the magma from which they are derived. Professor Brögger believes that among the basic eruptive rocks of Gran in the Christiania district, one and the same magma has in the bosses solidified as olivine-gabbro-diabases, and in the dykes as camptonites, bostonites, pyroxenites, hornblendites, and more acid augite-diorites.[42] [Footnote 42: _Quart. Journ. Geol. Soc._ l. (1894), p. 35.] Various opinions have been propounded as to the cause or causes of this so-called differentiation, but none of them are entirely satisfactory. We must await the results of further exploration in the field and of continued research in the laboratory. What appears to have taken place within a subterranean molten magma which has been propelled into the earth's crust as a boss or laccolite, with or without a connected system of dykes, may possibly be made to throw some light on the remarkable changes in the characters of lavas successively erupted from the same vent during the continuance of a volcanic cycle. Whether or not any such process of differentiation can be proved to take place within a subterranean volcanic reservoir, the sequence of erupted lavas bears a curious resemblance to the order in which the constituents of some large bosses succeed each other from margin to centre. The earliest lavas may be of an intermediate or even basic character, but they generally tend to become more acid. Nevertheless alternations of basic and acid lavas which have been noted in various districts would seem to show that if there be a process of differentiation in the magma-basins, it is not regular and continuous, but liable to interruption and renewal. The return to basic eruptions, which so often marks the close of a volcanic cycle, is likewise not easily explicable on the supposition of continuous differentiation.