J. D. Dana — Explosive Eruptions. 107 



a shelf. The j were manifestly local, and had their center near 

 the surface — an effect, not a cause; and they thus prove that 

 the immediate cause of the eruption was local. The facts as 

 to the Krakatoa earthquakes are similar. The deafening roar 

 in each was made chiefly by the violent projectile actiou, the 

 incessant detonations and the thunderings of a storm. 



Such eruptions are of a wholly different cast from the ordin- 

 ary outbreaks and discharges of Hawaii. The projectile agent 

 must gain access to the conduit lavas to produce such extraor- 

 dinary violence. 



The eruption of Tarawera Mountain was probably brought 

 about by the opening of a fissure that let subterranean waters 

 into the reservoir of lavas ; for Lake Rotornahana, situated on 

 the line of fracture and only three or four miles distant, lost 

 its waters, and probably in the process of supplying water for 

 the projectile work. The volcanic mountain had been long 

 extinct ; but the widely distributed geysers and boiling springs 

 were testimony to the existence of liquid lavas just below 

 the reach of descending atmospheric waters. The geyser lakes 

 of Rotorua and other localities became hotter during the night 

 of the eruption and continued so afterward. Under such con- 

 ditions an old volcanic mountain, perhaps hollow from former 

 discharges, might be burst open again. Had the ingressing 

 waters passed into the lava-reservoir at a great depth below 

 the surface, the generated vapors would necessarily have added 

 outflows of lava to the projectile discharge. 



The volcano of Krakatoa, was probably started into action 

 by a similar incursion, but of marine waters. 



In both cases there were enormous chasms and crater-like 

 depressions made, with a loss of the old foundations, and of 

 the rocks that occupied the depressions. But the facts, while 

 they include the projection of large stones over the vicinity, 

 show positively that the stones were few compared with what 

 would be needed to fill the great cavities left in the region. 

 The explosive eruption blew to great heights fragments of the 

 liquid lavas in the shape of scoria and sand or ashes, but did 

 not blow off the solid rocks of the mountain. The disappear- 

 ance of these and the making of the cavities are explained by 

 the engulf men t or down-plunge of material to fill the space left 

 empty by the projectile discharges. 



An explosive eruption is then simply one in which the pro- 

 jectile action, instead of ceasing at the time of eruption, be- 

 comes enormously increased ; in which the erupting agent, 

 instead of being roused to action outside of the lava-conduit, 

 gains access to its interior, and hence the terrific boiler-like 

 explosion. 



For further explanation I repeat that the ordinary activity 



