Pcle'^s Hair Under Banks. 



137 



It was very easj' to see what toppled down these fantastic cliffs, for the molten 

 mass within the lake was most active near the edges and under the banks which were 

 generally undermined horizontally to the extent of fifteen or twenty feet by the white- 

 hot, restless waves. From the under surface of these over-hanging shelves depended 

 long and flexible skeins of what seemed to be volcanic spun-glass, or Pele's hair, lapped 

 by the white waves, and seeming, in the glare in which they swung, to be hot to trans- 

 parency. These pendents were very numerous, often a foot in diameter and six to 

 ten feet long, fibrous as asbestos, and very flexible. Although they were one of the 

 most remarkable appearances at the southeast lake, it was nearly half an hour before 

 I had ■A.wy direct evidence of the process of their formation. Occasionally surface ex- 



/. 







rr^r=fsrr ^ 





From South East Lake— toward North. 



Fig. 8u. From southeast lakk, looking northward. 



plosions took place and the viscous fragments, thrown violently against the roof above, 

 spun out in falling back a glass thread, sometimes several from each lump, the frag- 

 ment being sometimes as large as a man's head. An attraction, probably electrical, 

 as the compass needle is stronglj- agitated in the vicinity of the currents from the 

 lakes, drew together these isolated threads until the hank was formed which floated 

 like seaweed in a falling tide. Although I watched several hours I did not see any of 

 these hanks fall into the lake beneath. 



The brittle nature of the banks which were formed by overflows and ejected 

 matter loosely cemented by subsequent overflows or spatters, would admit of any amount 

 of degradation, but how is the elevation to be explained? A prolonged stay at the 

 crater suggested the following explanation. The action in these fire-lakes or pools, 

 as has often been mentioned, is very irregular and intermittent, often apparently ceas- 

 ing on one side until the crust there is cool and hard; it then breaks out again from 



[515] 



