ON THE FORMATION OF CRYSTALLINE ROCKS. 287 



one as in the other case, tlie water appears to favor the elimination of 

 substances which would remain mixed, and to permit the crystalHza- 

 tion of silicates at a temperature very much lower than their point of 

 fusion.* It is again through the influence of this kind of mother 

 water that the same silicates crystallize in a succession which is often 

 opposed to their relative order of fusibility. We know, for instance, 

 that amphigene, a silicate of alumina and potash, which is infusible, 

 has been developed in the lavas of Italy in crystals which are often 

 very large, and in which are imbedded numerous crystals of pyrox- 

 ene, a substance of known fusibility. These apparent anomalies are 

 shown in a still more striking manner in granite, which differs from 

 ' all the products of dry fusion with which we are acquainted, and dif- 

 ferent conjectures have been offered to explain them. It can be ex- 

 plained nearly in the same manner ; only in the formation of granite 

 the action of water, according to the observations of Elie de Beau- 

 mont, appears to have been aided still more than in the lavas by some 

 auxiliaries, such as the chlorides and the fluorides. In the felds- 

 pathic porphyries containing quartz, water alone would have sufficed 

 to produce the bipyramidal crj^stals which characterize this rock. 

 This is again another capital phenomenon which has nothing analo- 

 gous in the products of the dry way. 



The remarkable association of anhydrous and hydrous silicates which 

 basalt, phonolite, and other rocks present, is not surprising after 

 the experiments that I have jast described. For during the same 

 operation, and in the same tube, I have obtained crystals of pyroxene 

 disseminated in the midst of a zeolite ; that is to say, the two con- 

 stitutive elements of basalt simultaneously. A still greater difficulty 

 presented itself when, on the one hand, the soft state or even fluidity 

 of some eruptive rocks was considered; and on the other, their low 

 primitive heat, which is well established by different circumstances. 

 This difficulty is also done away when we consider what has taken 

 place in these same experiments. The glass tubes which were per- 

 fectly regular were found, after the operation, bent, deformed, and 

 covered with ampullar in such a manner as to prove that they had un- 

 dergone an unmistakable softening. Furthermore, sometimes the tube 

 had alro.ost disappeared ; it was transformed into a kind of mud, pre- 

 senting probably great analogies, as well in consistence as in com- 

 position, with the original state of some eruptive rocks. A very re- 

 markable phenomenon, to which I shall have occasion to revert, is 

 here produced ; although the glass in the transformation loses a part 

 of its elements, it augments considerably in volume ; this augmenta- 

 tion is more than a third of the primitive volume. When we see the 

 important part which water plays in the phenomena just passed in 

 review, should we not also be led, with still greater reason, to attri- 

 bute to it the most important part in metamorphic action, especially 



^' I will here recall the opinion of Dolomieu on the crystallization of lavas : "I repeat, 

 perhaps for the hundredth time, the compact lavas are not vitrifications, and their fluidity 

 at their exit from volcanoi^s, which continues a much longer time than their cooling should 

 permit, is the very singular effect of a cause which is not yet determine 1." — [Journal des 

 Mines, No. 37, p. 402,^1797.) 



