J. D. Dana — JErwptions of Kilauea and Jit. Loa. 101 



fused rock seldom exuding ; (5) the blocks of the upper part 

 of the front, as the stream creeps on, keep tumbling down the 

 high slope, owing to retardation at bottom from friction, and 

 thus a rolling action in the front part. 



The aa-field, owing to its crevices and shaded recesses, 

 retains moisture, and decomposition at surface early com- 

 mences, which favors germination of seeds ; and often the 

 stream, as I am informed by Mr. Baker, becomes forest- 

 covered when the pahoehoe alongside remains bare. 



One of the best published descriptions of an aa-flow is that 

 of Judge Hitchcock (p. 31) which says : "Along the whole line 

 of the advance, the stream, twelve to thirty-five feet in height, 

 was one crash of rolling, sliding, tumbling, red-hot rock, no 

 liquid rock being in sight ; with no explosions, but a tre- 

 mendous roaring like ten thousand blast furnaces all at work 

 at once.'' Mr. Baker writes (letter of February, 18S8) : " I have 

 stood by a wholly molten stream of lava which miles below was 

 cooling into aa." 



Under the restrictions of such facts the aa cannot be ex- 

 plained by referring it to simply a partial cooling of a stream 

 and then a breaking up of the crust on a new accession of 

 flowing lava — a common explanation ; for there is no evidence 

 of a crust from surface -cooling analogous to that of pahoehoe. 

 It is not dependent on the mineral constitution of the lava ; 

 for one and the same stream may take either condition ; and 

 adjoining fields near Punaluu are at opposite extremes as to 

 the amount of chrysolite. 



The first conclusion we may draw, in view of the facts, and 

 especially the abrupt transitions from aa to pahoehoe and the 

 reverse in the flowing stream, and the independence of kind 

 of lava, is that the difference must be connected with some 

 condition in the region flowed over ; and, the second, that 

 where the transition from one kind of stream to the other 

 occurs, the conditions must be such as will allow of extreme 

 liquidity in one part (the pahoehoe), and occasion imperfect 

 liquidity or a pasty state in the other (the aa). 



It follows also from the size and rough character of the 

 blocks of lava, thirdly, that in an aa stream the lava must 

 have been subjected to some deeply-acting cooling agency to 

 have made a crust thick enough for blocks 10 to 20 feet and 

 more in dimensions — far thicker than the crust over the tunnels 

 in a pahoehoe stream. Fourthly, that the cooling was not from 

 above downward, as in the pahoehoe, for there are no remains 

 of a crust in the true aa field, but largely from below upward ; 

 and thence comes the absence of a crust and of the usual amount 

 of vesiculation. 



These four conclusions appear to lead directly to a fifth : 



