284 ANNUAr. RRPORT SMITHSONIAN INSTITUTION:, 1913. 



placement type and vacuum tubes especially arranged to meet the 

 conditions which we had found. These tubes Avere of the same ca- 

 pacity as the individual tubes in the previous experiment (one-half 

 liter) , but were provided with a long stem, on the remote end of which 

 was blown a thin glass bulb. The plan was to attach these tubes to a 

 pole of convenient length and to thrust the end carrying the thin 

 bulb into the dome, where the heat might be expected to explode the 

 thin glass immediately, permitting the tube to fill with the gases, and 

 as quickly to seal it again by melting down the broken end. The 

 tubes were dried in contact with phosphorus anhydride and the de- 

 gree of exhaustion checked by electrical discharge tests from a small 

 static machine. When a number of these tubes had been prepared 

 and everything was ready for a second attempt, the top of the lava 

 dome had fallen in and the liquid lava in the basin had gone 

 down to such an extent that it offered no further opportunity to 

 collect gases under conditions which should assure original gas with- 

 out contamination from the air or otherwise. In fact no other op- 

 portunity offered until December 4.^ 



THE SECOND ATTEMPT TO COLLECT GASES. 



On December 4, with the lava surface 360 feet below the rim, and 

 therefore even less conveniently accessible than on the previous occa- 

 sion, a similar dome formed directly on the border of the lava lake, 

 and the second attempt was made to collect a quantity of gas — this 

 time in the vacuum tubes. In order that there might be no possible 

 doubt about the excess gas pressure within the lava dome, the 

 descent into the crater was made at night, when the pale blue flame 

 of the escaping gases could be plainly seen emerging from the crack 

 in the dome. The manner of collecting the gases was exactly that 

 which was planned and described above, and six tubes were filled 

 with gas under these conditions. 



On descending into the crater to collect gases from the December 

 dome, it was found that in addition to a long slit or crack across the 

 top, from which the gases were discharging constantly, there was a 

 second opening near the base Avhich was not noticed before the descent, 

 but which gave access to air at the base of the dome and thus behaved 

 like an air blast in a furnace. The gases were therefore partly 

 burned within the dome instead of outside, and the tubes, which were 

 filled at the upper opening, were accordingly found to contain chiefly 

 burned gases — that is, the free hydrogen had become water, the free 

 sulphur had burned to SOo, the CO appeared as CO,, etc. 



Although the identity and something of the relation of the gases 

 discharged from the basin of Halemaumau can be established from a 



1 In tho meantime one of the authors (Day) was obliged to return to Washington, 

 leaving the other (Shepherd) to finish the task alone. 



