Mechanism of Stromboli . 511 



out of the tube, which actually is emptied, at the end of each outburst, of 

 everything but gases and vapours, and these at a tension not greatly ex- 

 ceeding that of the atmosphere. We do not know the average section 

 of the tube, and therefore cannot calculate the volume of lava that would 

 be propelled thus out of the tube, if previously filled by each outburst ; 

 but it is manifestly so great that it would wholly change the character of 

 the -phenomena exhibited by the volcano, and must, during the last 2000 

 years, have produced a mass of ejected matter of enormous magnitude 

 instead of the insignificant amount of mixed lava and debris which alone 

 are to be seen. 



Liquid or semiliquid lava does, however, continually mate its way into 

 the bottom of the crater-funnel and amongst the fragments collected there, 

 which it more or less solders together, and along with which it is blown 

 out at the outburst. Some may ooze into the tube lower down, and may 

 more or less obstruct, but can never completely fill it. The walls of the 

 tube, and those of all the fissures or cavities below the level at which the 

 more or less fused lava reaches them, can scarcely have a lower tempera- 

 ture, and are most probably at a higher one than the lava itself. If the 

 tube of the volcano were the main, or only, ajutage through which the 

 liquid lava, as well as the steam to blow it out, were supplied if, in fact, 

 the tube were the main duct into the lowest depths of which both the 

 liquid and vaporous matters entered then, at irregular intervals, the tube, 

 and even the crater, must become filled, and the whole phenomena of 

 eruption would not differ in character from the highly irregular paroxysmal 

 efforts of any common volcano of like energy. The tube, then, here 

 plays a different part from what it usually does, and constitutes an addi- 

 tional element in its machinery, upon whose action in producing expulsion 

 the rhythmical recurrence of the outburst depends. 



We can now discern the very simple mechanism by which the actual 

 phenomena are produced, a description of which will be rendered more 

 intelligible by reference to the ideal Diagram No. 4, in which A is the 

 lower part of the funnel of the crater, filled more or less with the frag- 

 mentary mass which has fallen back into it from the preceding outbursts. 

 B is a lateral duct conveying more or less liquid lava into the bottom of 

 the crater. C is the tube leading to the bottom of the funnel from a 

 depth considerably below the sea-level, supposed to be, at the line L, at 

 about 400 feet below the upper lip of the tube. D is a duct communica- 

 ting with the sea, and enabling sea-water to find access to the interior of 

 the tube, and to rise therein, if otherwise unimpeded, to the sea-level. 

 E is either a lateral duct or a continuation of the tube itself, through 

 which steam at a high temperature and tension enters the tube at some 

 point much below the level of the sea. The lava- and steam-ducts, B 

 and E, may be supposed to come from the ancient great volcanic channels 

 still remaining under the more central parts of the island, and which 

 supplied its great ancient craters. The duct D may consist of many 



