482 R. MALLET ON THE MECHANISM OE 



they may have been originally deposited. Any snch inference as to 

 a particular case might be grossly in error, and would probably be 

 so in the case of curved or twisted though more or less continuous 

 dykes. The error may be greatly less, or perhaps no error may exist, 

 where the inference is drawn from nearly flat dykes, whose planes 

 present neither considerable twist, curvature, nor fracture. But as 

 we never can be certain that any dyke may not have had its position, 

 especially its orientation, more or less changed or shifted subsequently 

 to its production by injection, and as we also never can be certain 

 that any two dykes, whether distant or adjacent, and of which we can 

 only see some one section of a part of the dyke, which may be chiefly 

 a vertical section as in Somma, or may be more commonly horizontal 

 sections as well as vertical ones, such as are found surrounding the 

 Val del Bove, so we never can be certain that any two, though 

 now apparently converging, must have emanated from a common 

 volcanic axis or crater. The evidence as to such common origin 

 can never be more than probable, though the probability may cer- 

 tainly be strengthened in some ratio to the number of dyke-planes, 

 or planes of orientation, which we may observe (after sufficiently 

 exact geodetic operation) all to appear to converge to a single axis. 

 The divergences from even approximation to any such single axis, 

 found by me in the dykes of Somma, at one time raised doubts in 

 my mind as to whether they were injected dykes at all, or might 

 not be ancient lava-beds which had flowed down the slope of 

 the mountain at various epochs, and had, by some unexplained 

 causes, been broken up, more or less turned on edge in planes 

 in a line with, or transverse to, the slope, or even occasionally 

 inverted altogether. This view I was subsequently compelled to 

 abandon, though some facts observed during the careful lithological 

 survey which I made of these beds seemed for a time to support it ; 

 and I should add that one class of facts observed and about to be 

 referred to, which seemed to support this hypothesis strongly, are 

 to me still without an explanation and remain for the more com- 

 plete investigations of others. I may remark here that the descent 

 of large surfaces of lava-beds, along with the sand and detritus 

 which, from the causes already referred to, and aided by rains, are 

 in many instances observed to have descended more or less upon 

 volcanic flanks, and the turning up of such beds more or less 

 on edge, either in a direction transverse or approaching to their 

 line of descent, are by no means uncommon phenomena in volcanic 

 countries. If the large plate of lava thus descend upon the detrital 

 slope of an uninterrupted volcanic cone, the lower or forward end 

 of the huge slab, like a flat coulter, tends constantly to increase 

 the volume of a ridge of detritus heaped up before its foremost 

 edge; and a time may arrive when, from this cause and the 

 slowly digging deeper into the detrital bed of the advancing edge, 

 the whole huge plate is brought to rest. If sufficiently large, 

 the detrital matter of the bed is interrupted in its descent by the 

 upper or higher edge of the plate, while it is washed out more or less 

 from the lower edge by the percolation of rain, which makes its exit 



