PRODUCTION OF VOLCANIC DYKES. 495 



number of these opened cones show no trace of ever having con- 

 tained lava. The true nature of the mechanism by which the) 7 have 

 been opened may, I believe, be inferred from the preceding pages. 

 The opening has been effected not by bursting* but by the under- 

 mining of one part of the base of the crater-wall, either by a small 

 outburst of lava beneath or, in the case of those opened cones which 

 present no trace of lava, by the erosive leakage of rainwater filling 

 to some depth the cavity of the crater. Although these hollow dust 

 heaps are porous, the bottoms are more or less stanched to rain by a 

 sort of natural puddling with very fine material washed down from 

 their inner sides, in the same way as certain portions of the Atrio 

 del Cavallo have been made watertight and contain ponds of water 

 in the winter time. 



Q. J. G. S. Ko.123. 2 m 



