134 DR. A. L. BU TO IT OX THE [vol. IxXY, 



scapolite in Mesozoic limestones in the Pyrenees is interesting, the 

 invading rock being lherzolite. The invariable contiguity of this 

 reaction-zone to the igneous junction argues for only a limited 

 range of movement for the silica, alumina, alkalies, and chlorine in 

 this process of magmatic ' excretion.' 



The presence of a shell of phlogopite postulates the transference 

 from the magma of silicon tetra fluoride, potash, and in many cases 

 alumina, the marbles being normally low in these two bases ; an 

 excess of the latter will obviously be used in the production of 

 spinel. 



The greater volatility of the fluorine-bearing emanations must 

 tend to make this zone travel farther outwards, leaving between 

 it and the granite a layer of silicates devoid of fluorides, while 

 calcitization in the region beyond implies some inward movement 

 of magnesia also. With a low content of alumina and potash, or 

 in default of their outward transference, forsterite would normally 

 form ; but, in the presence of hydrofluoric acid, one of the chon- 

 drodite minerals instead, for all that would be required would be the 

 introduction of (MgP. OH) into the forsterite molecule Mg.,Si0 4 

 in certain proportions to produce one or other of that fluorine- 

 bearing group. The antipathetic relationship of chondrodite to 

 forsterite normally, but the occasional envelopment of the latter 

 by the former (as shown in PL IX, fig. 8), strongly supports this 

 view. 



In the presence of alumina, but absence of potash, spinel Avould 

 form, from which it may be inferred that, other things being 

 equal, mainly the presence or absence of potash determines whether 

 phlogopite should crystallize, or else a mixture of chondrodite 

 and spinel. The possible existence of authigenic felspars in the 

 sediment could, moreover, be invoked to explain the occurrence 

 of mica along the planes of original stratification in the mass 

 far away from contacts, the fluorine-bearing vapours preferentially 

 following the bedding-planes ; some of the flakes might even be 

 regenerated clastic micas. 



The endomorphic changes are much less marked, and result 

 primarily from the absorption on a very limited scale of the 

 dolomitic limestone by the invading magma. 



Magnesia, having a greater affinity than lime for silica and 

 alumina, is more easily digested ; hence pyroxene and calcite 

 separate from the dyke, whereas, owing to desilication, the magma 

 becomes richer in alumina, potash, and soda, the first two seem- 

 ingly tending to leave the dyke, while the last determines the 

 development of perthite and soda-pyroxene. In brief, this is 

 Dr. R. A. Daly's hypothesis of the derivation of the alkaline 

 igneous rocks from carbonate syntectics. 1 In the small-scale 

 phenomena observed in the Marble Delta this alkaline facies is 

 strictly confined to the active margin, and does not obtain in 

 the more central part of the intrusion, where the abnormally- 



1 Proc. Nat. Acad. Sei. Washington, vol. iii (1917) i>. 659. 



