Ch. XXIX.] AQUEOUS EROSION IN PALMA. 505 



longing to the upper formation in the Caldera, would leave behind them 

 few pebbles. Nearly all of these perishable deposits would be swept 

 down in the shape of mud into the Atlantic. Even the hard rounded 

 stones, since they were once angular and are now ground down into peb- 

 bles, must have lost more than half their original bulk, and bear witness 

 to large quantities of sedimentary matter consigned to the bed of the 

 ocean. We saw in the Caldera blocks of huge size thrown down by 

 cascades from the upper precipices during the melting of the snows, 

 a fortnight before our visit, and much destruction was likewise going on 

 in the lower set of rocks by the same agency. We also learnt that 

 a great flood rushed down the Barranco in the spring of 1854, shortly 

 before our arrival, damaging several houses and farms, and I have there- 

 fore no doubt that the erosive power even of rain and river water, aided 

 by earthquakes, might in the course of ages empty out a valley as 

 large as the Caldera, although probably not of the same shape. I am 

 disposed to attribute the circular range of cliffs surrounding the Caldera 

 to volcanic action, because they forcibly reminded me of the precipices 

 encircling three sides of the Val de Bove, on Etna ; and because they 

 agree so well with Junghuhn's description of the "old crater-walls" 

 of active volcanoes in Java, some of which equal or surpass in dimen- 

 sions even the Caldera of Palraa. The latter may have consisted 

 at first of a true crater, enlarged afterwards into a caldera by the 

 partial destruction of a great cone ; but if so, it has certainly been since 

 modified by denudation. Nor can any geologist now define how much 

 of the work has been accomplished by aqueous, and how much by vol- 

 canic agency. The phenomenon of a river cutting its channel through 

 a dense mass of ancient alluvium formed during oscillations in the level 

 of the land is not confined to volcanic countries, and I need not dwell 

 here on its interpretation, but refer to what was said in the 7th chap- 

 ter. (See p.. 84.) 



There remains, however, another question of high theoretical interest ; 

 namely, whether the denudation was marine or fluviatile. It was stated 

 that the materials of the great cone or assemblage of cones in the 

 north of Palma are of subaerial origin, as proved by the angularity of 

 the fragments of rock in the agglomerates ; but it may be asked, 

 whether, when the Caldera was formed long afterwards, it may not, like 

 the crater of St. Paul's (fig. 649, p. 509), have had a communication 

 with the sea, which may have entered by the great Barranco, and if, 

 after a period of partial submergence, the island may not then have risen 

 again to its original altitude. In such a case the retiring waters might 

 leave behind them a conglomerate, partly of river-pebbles, collected at 

 the points where the torrent successively entered the sea, and partly 

 of stones rounded by the waves. The torrent may have finally cut a 

 deep ravine in the gravel and associated lavas when the land was rising 

 again. Such oscillations of level, amounting to more than 2000 feet, 

 would not be deemed improbable by any geologists, provided thej r 

 enable us to explain more naturally than by any other causation, the 



