58 BUNSEN ON THE FORMATION OF 



3 : 1'998, which may be regarded as identical ; and these numbers 

 certainly differ less than the results of the several analyses from 

 which these mean values are deduced. For this reason an almost 

 uniform composition, frequently corresponding more closely with 

 the calculation than when the normal pyroxenic substance is 

 applied to this purpose, is obtained when the normal palago- 

 nitic substance is made the basis of the calculation. But the cir- 

 cumstance that the composition of these palagonitic substances 

 deviates rather more from the mean results of analysis, explains 

 at the same time in the most simple manner the slight discre- 

 pancies observed in some few of the lavas occurring in the pala- 

 gonitic district of Iceland, the quantity of lime corresponding to 

 the palagonitic composition decreasing slightly in proportion to 

 the increase of magnesia. Even the almost perfect identity of 

 the constitution of the palagonitic tuffs with that of the normal 

 pyroxenic rocks, the absence of trachytic admixtures, and still 

 more their gradual transition into the almost anhydrous sub- 

 stance of the pyroxenic rocks, which may be observed on a large 

 scale, as well as in the separately-imbedded masses, render it in 

 the highest degree probable that the formation of tuff is most 

 intimately connected with the formation of the pyroxenic rocks. 

 Observations which Darwin made in the Cape Verde Islands, 

 and my own examination of specimens of the rocks to which they 

 refer, for which I am indebted to the kindness of that distin- 

 guished investigator, have chiefly contributed to give me a clue 

 to the remarkable processes concerned in the formation of pala- 

 gonitic tuff. 



In the neighbourhood of Porto Praja, a basaltic lava occurs, 

 which has been poured out over a recent deposit of limestone. 

 It may there be seen that the lava has, while in a liquid state, 

 acted upon the underlying limestone over which it flowed, and 

 become loaded with fragments of it. The product of this mutual 

 action is a breccia-like conglomerate, in which the altered lava 

 is mixed with a very pure mass of carbonate of lime. A closer 

 examination of this mixture, which even in its exterior has quite 

 the appearance of having been kneaded together in a pasty con- 

 dition, excludes every possibihty of supposing that the fragments 

 of limestone accompanying the lava have originated from sub- 

 sequent infiltration. The chemical change which has resulted 



