346 VOLCANIC ERUPTIONS ON MARTINIQUE AND ST. VINCENT. 



The breaking of telegraph cables a few miles offshore from St. 

 Pierre at the time the city was destroyed and during subsequent erup- 

 tions of Mont Pelee seems best accounted for on the supposition that 

 landslides occurred on the steep sul)merged slopes, similar to those 

 observed at the margin of the adjacent land. 



Electrical d I H2}lay^. — The graphic accounts that have been published 

 of the recent eruption gi^'e a better idea of the magniiicence of the 

 electrical phenomena accompanjdng volcanic explosions than was pre- 

 viousl}" attainable. These observations show that an interesting and 

 difficidt problem here awaits solution. The most striking phase of 

 what is assumed to have been an electrical displa}^ during a primary 

 eruption of Mont Pelee on the evening of May 26 is thus described ])y 

 George Kennan: 



"The feature of the eruption that made the deepest impression upon 

 me was the stellar lightning. The uprush of black smoke, the glow 

 over the crater, and the shower of incandescent stones and cinders 

 were all phenomena that had been observed and described before, but 

 the short, thin streaks of lightning, followed by star-like explosions 

 in the volcanic mantle — not only above the crater, but miles away from 

 it — were entirely new. The distinctive characteristics of this light- 

 ning were the shortness of the streak, the comparatively great size and 

 brilliancy of the spark, or light burst, at the end of the streak, and the 

 single l)ooming report that followed. Sometimes three or four great 

 sparks, connected l)y fiery streaks, would flash together in this way, and 

 at other times the stars would burst so far back in the cloud that the 

 streaks were invisible and there was only a circular irradiation of 

 the vapor. If there was any storm lightning of the ordinary kind 

 in the earlier stages of the eruption, it was so much less noticeable 

 than the stellar lightning that it escaped my observation, and I am quite 

 sure that there was no rolling, reverberating thunder at all until near 

 the close of the display, when reddish lightning bolts began to dart 

 down on the volcano from the developing stoi-ni cloud over the crater. 

 Before that time all, ov nearly all, of the electric discharges had ended 

 in stellar ligiit bursts, and all of the thunder had been made up of 

 separate and distinct reports, like the thunder of a heaxv and rapid 

 caimonade."" 



Supplementing the above account we are fortunate in having an 

 instructive picture of a night eruption of Mont Pelee, drawn b}^ 

 George Varian (PI. V), in which the appearance of the so-called 

 stellar lightning is well shown. 



In reading Kennan's vivid description with the aid of Variants 

 equally faithful sketch one can scarcely avoid making the tentative 

 suggestion that the streaks of light and brilliant explosions, appar- 

 entl}^ resembling the trails and occasional bursting caused by meteoric 

 bodies entering the earth's atmosphere, may have been due to intensely 

 heated solid projectiles blown out of the volcano and entering the 

 oxvgen-charged air. 



