346 F. A. Perret — Circulatory System in the 



highly charged with minute gas vesicles which gradually rise 

 and escape during the passage across the lake, the flow being 

 then represented in the diagram by short lines. The crast 

 grows in thickness and, on arrival at the eastern end, the sur- 

 face layers — cooled, condensed and gas-free — enter the spatter- 

 grotto and sink vertically as a continuous heavy sheet to form 

 and maintain the powerful syphon effect which is the main- 

 spring of the circulatory system. 



Arrived at the bottom and turning westward the current 

 passes over the mouth of the conduit from which are rising 

 the vesicles of juvenile gas. The large, intermittent, fountain- 

 forming bubbles rise rapidly and buoyantly to the surface, 

 suffering but little deviation by the horizontally moving strata 

 — of which the one corrects the other — but the innumer- 

 able minute vesicles can rise but slowly and these are en- 

 trapped by the west-flowing lava, which is thereby re-heated 

 and re-vivified. In this condition it has acquired ascensional 

 power and — as it must, in any case, rise to balance the descent 

 at the east — this power is thus added to the other forces in 

 maintaining the circulation. But it is important to note how 

 comparatively feeble is this ascensional power of the gas-vesic- 

 ulated lava which, instead of rising directly, is deflected to 

 the extreme end of the lake by the powerful effect of the 

 eastern downflow. 



The critical reader may here move the objection that, for a 

 paper wherein it is sought to confirm each statement by actual 

 observation, confident mention is now being made of a west- 

 ward return current in the lower part of the lake whose exist- 

 ence, however strongly indicated, is, nevertheless, deduced. 

 And we are, therefore, fortunate in being able to record a 

 circumstance which shows the reality of this undercurrent 

 almost as conclusively as direct vision would have done. At 

 the time of the sinking of the island of 1911, as described in 

 the preceding paper of this series, a quantity of gas was 

 evolved which formed and inflated a floating cylinder of black 

 lava, below which the sunken mass of the island hung down- 

 ward in the lower part of the lake, remaining attached by one 

 end to its floating buoy. The sinking occurred at the east end 

 of the lake and the cylinder at once began to make its way 

 westward against the surface current, it being towed by the 

 mass of the island suspended in the west-flowing undercurrent. 

 Nor is this all, for, on approaching the west bank, the sunken 

 mass was again brought to the surface, a point of the original 

 island being seen and photographed emergent from the lake. 

 This re-flotation was, no doubt, caused by the rising, gas- 

 charged lava current of the west end and, no sooner was it 

 accomplished, when the whole system proceeded eastward 



