EXPERIMENTS on JVHINSTONE and LAFJ. 57 



no idea of the real charadler of the lava, which can only be feen 

 in the interior parts of the currents. In forming our colledion 

 we fcrupuloufly avoided this error, and chofe fuch fpecimens 

 only as were the moft compadl and free from the fcorified ap- 

 pearance of the furface. 



Wh£N thefe folid lavas are compared with our wliinftones, 

 the refemblance between the two clafles is not only flriking at 

 firft fight, but bears the clofeft examination. They both con- 

 fift of a ftony bafis, which frequently contains detached cryflals 

 of various fubftances, fuch as white felfpar and black hornblend. 

 The analogy between the two clafles feems to hold through all 

 their varieties ; and I am confident that there is not a lava of 

 Mount ^tna to which a counterpart may not be produced from 

 the whinfl;ones of Scotland. 



This refemblance in external characfler is accompanied with 

 an agreement no lefs complete in chemical properties. But be- 

 fore I mention the experiments which tend to prove this agree- 

 ment, it will be neceflary firfl to examine the opinion of two 

 very celebrated authors concerning lavas. M. Dolomieu and 

 Mr KiRWAN, though they differ widely in many i-efpe<fls, agree 

 in believing, that lavas have never been a(5led upon by a heat of 

 fufficient intenfity to produce complete fufion ; and endeavour, 

 each by an hypothefis peculiar to himfelf, to account for their 

 fluidity. The opinion of thefe gentlemen is of fuch importance 

 in the prefent queflion, and the arguments they h^ve ufed are 

 fo e^^traordinary, that I muft beg leave to quote their words at 

 full length. 



M. Dolomieu ftates his opinion in the following pafl~age, 

 (IJles Ponces, p. y.) : " II eft eflentiel de conftater, par beau- 

 " coup d'exemples et d'obfervations, quelques verites quej'ai 

 " annoncees il y a plufieurs annees, favoir, que le feu des vol- 



cans ne denature pas ordinairement les pierres qu'il a mifes 

 " en etat de fufion ; qu'il ne les altera pas au point de ne pou- 

 " voir les reconnoitre, de ne pas diftinguer quelle a pu etre la 



Part I. . H " bafe 



