568 PROCEEDINGS OF THE PHILADELPHIA MEETING 



probable cross-section of the conduit or system of conduits of eruption was about 

 1,500 feet in diameter. Such a vast mass of rigid lava certainly would offer greater 

 resistance to ascending eruptive forces than would the surrounding fragmental beds, 

 and it would not be lifted bodily into the air. A new vent would more probably 

 be opened beside the old plug or the rising heat would melt its way through the 

 plug. 



A violent eruption like that of Mont Pele on May 8, 1902, would certainly have 

 cleared the conduit of the volcano or at any rate have given a practically free 

 course to the rising eruptive material. The clearing of the conduit had been going 

 on gradually, but with increasing violence, for some weeks before the catastrophic 

 explosion occurred which overwhelmed Saint Pierre. 



If the ancient conduit were occupied with congealed lava prior to 1902, the re- 

 vived volcanic agents found their way through the plug or ejected it in fragments 

 into the air before the massive dome and spine began to form. Such action might 

 or might not leave a tubular ring of solid lava around the new conduit. If such 

 a tube were left between the conduit and the main tuff beds of the mountain, it 

 hardly seems possible that the eruptive forces could get under it or between the 

 tube and the matrix in such manner as to lift the shell bodily into the air. If no 

 tube of solid lava were left, there would be no shell of ancient lava to be elevated 

 as even the remnants of a plug. The positive arguments, however, in favor of the 

 third theory are more convincing than the negative arguments which have been 

 advanced against the second hypothesis. 



The third theory is that the dome and spine consisted of lava which solidified 

 and became rigid as it was extruded or else in the upper part of the conduit, at 

 least to such an extent that no flow of lava resulted. This is the theory which 

 the author has advocated* and which accords with that advanced by Professor 

 Lacroix,f who first recognized the fact that the new cone of eruption was mainly 

 massive and not fragmental in character. More recently the theory has been 

 elaborately and ably defended by Israel C. Russell. J 



The lava of Pele is andesitic in type and therefore rather difficult to fuse. The 

 result is a viscous stream at the ordinary temperature and under the usual condi- 

 tions of eruptions. In some of the preceding eruptions of Pele streams of molten 

 lava have issued from the vents and flowed short distances ; in most, however, there 

 has been sufficient excess of water vapor present to produce entirely explosive erup- 

 tions. The presence of expanding gas and vapor lowers the temperature of the 

 containing fluid, the effect increasing with increase in the proportionate amount 

 of gas and vapor present. The sudden diminution of pressure at the vent of the 

 volcano produced corresponding expansion of the occluded gases and rapid lower- 

 ing of temperature, with resulting greater or less solidification. § At Pele, after 

 the outburst of May 8, 1902, the balance between forces was so nicely adjusted 

 that the lava welling up out of the conduit congealed as it rose and formed the 

 famous dome and spine instead of flowing down the mountain in a stream. 



The lava near the walls- of the conduit would naturally be cooler and therefore 

 more viscous than that near the center of the rising column. The result would be 

 the freer rise of gases and vapor through the center than through the sides of the 



*C. R. IX Congr. Geol. int. de Vienne, 1903, pp. 724, 734. 



-fComptes Rendus, 27 Oct., 1902. 



J Science, n. s., vol. xxi, p. 924. 



gSee A. C. Lane, in Science, n. s., vol 18, p. 760, and G. K. Gilbert, in Science, n. s., vol. 19, p. 927. 



