﻿290 ME. J. V. ELSDEX 0^' THE C^^}' I908. 



pyroxenes owing to some peculiarity of surface-tension, leading to 

 capillary repulsion instead of attraction. AYe may, in such a 

 case, imagine, in the early stages of consolidation of the magma, 

 a scattered dust of the minerals of first separation, in this case 

 apatite, and a kind of selective inclusion of these in the subsequently- 

 formed crystals. There are reasons for believing that this selective 

 action does in fact take place in certain cases. An objection to 

 this view, however, is that in the rocks of other localities the 

 pyroxenes often do contain apatite-crystals in fair abundance/ In 

 my opinion, the greater preponderance of apatite in the more acid 

 rocks in the St. David's-Head area seems to point to the conclusion 

 that the bulk of the phosphoric acid was originally contained in 

 the acid portion of the magma. 



Y. Eelaiioxs or the Eocx-Ttpis. 



T^hatever view we may take of the difierentiation-process which 

 has resulted in the various types of rock described in the foregoing 

 pages, it is clear that these suggest a typical rock-series, 

 primarily derived from a common magma. It will be convenient 

 here to illustrate graphically the chemical relations of the different 

 types. Eor this purpose we may use the method of Prof. Iddings,* 

 plotting the silica-percentages along an abscissa, and the other 

 oxides along ordinates erected at the successive points thus obtained. 

 For the sake of space a part only of the complete diagram is shown 

 (fig. 2, p. 291). It is evident, from an examination of this figure, 

 that the curves thus obtained deviate in some cases from even an 

 approximate linearity. ^Ye have, of course, to remember that many 

 things may occur during the consolidation of a magma to interfere 

 with a linear result.^ Thus, owing to convection-cnrrents, magmatie 

 flow, and other causes, crystals do not necessarily remain in the 

 positions in which they consolidated. This must be particularly 

 true in the early stages of crystallization. It is only by making an 

 extensive series of analyses, and by averaging the results, that these 

 factors can be even partly eliminated. Analyses of isolated samples 

 of rock can only be expected to give a rough approximation to the 

 composition of the magmas from which they originated. It is 

 probably for this reason that we find the lime-curve following a 

 somewhat erratic course instead of conforming, as might have been 

 expected, to those of magnesia and iron. Xeither the curve for 

 titanium nor that for potash suggests diff'erentiation of the Carrock- 

 Fell type, and in view of the fact that the soda-lime felspars 

 were almost invariably among the first minerals to separate, it is 

 difficult to explain, upon the theory of differentiation in situ, the 

 large proportion of soda in the aplite-veins. The whole of the facts, 



1 See J. D. Falconer, Trans. Roy. Soc. Edin. vol. xlv (1906) p. 141. 



- 'The Origin of Igneous Eocks ' Bull. Phil. Soc. Washington, vol. xii 

 (1892-94) pp. 89-214. 



' Upon this point see A. Harker, ' On Tgneous-Eock Series <fc Mixed 

 Igneous Eocks ' Journ, Geol. Chicago. toI. viii (1900) p. 389. 



