( 2o; ) 



iavas, have totally deftroyed thefe prifms; 

 but that it can have had that efFed on all, 

 and through fo extenfive a fpace, is utterly 

 improbable. Nor is it conceivable that 

 Felicuda, among the Eolian ifles, fhould 

 ftill preferve its prifms perfedi:, while the 

 reft of thofe iflands have entirely loft theirs, 

 notwithftanding they are all equally expofed 

 to the fhocks of the waves. 



I cannot here omit another remark. It 

 is certain that more than one of thefe 

 lllands w^ere not formed by one eruption, 

 but by fucceflive ejedions of lavas accu- 

 mulating on each other ; and in fome deep 

 fiflures, occafioned by the fea, this fuccefTive 

 formation is difcoverable by the eye, as we 

 may perceive five or fix different ftrata 

 of lava one above the other. The inter- 

 nal ftrata in very remote periods", having 

 flowed into the fea, as the external flowed 

 afterwards ; it is evident that, if the latter, 

 on touching the water, became prifma'ic, 

 the fame change muft have taken place, 

 alfo, in the former ; w^hich being defended 

 4 from 



