By the Rev. A. C. Smith. 383 



the expansion of the air within the building will account for the 

 striking fact, that in the case of a somewhat lofty house with a 

 tiled roof at Cherhill, as well as in the case of several thatched cot- 

 tages both at Cherhill and Yatesbury, the sides of the roofs most 

 exposed to the full fury of the hurricane, and on which the storm 

 blew, were uninjured; while the opposite or eastern sides, which 

 were apparently protected from the wind, were carried off; where- 

 by it seems probable that the expansion of the air within the build- 

 ing forced off the roofs on the eastern sides, as the storm cloud 

 passed over, while those on the western sides were kept on by the 

 air pressing onwards towards the rarefied space in the cloud itself. 

 Perhaps the same principle may account for the apparent incon- 

 sistency, that in several cases low buildings attached to the eastern 

 sides of more lofty houses, and where it would seem impossible for 

 them to be affected by the storm, have been completely unroofed, 

 as is most strikingly shown at Quemerford Mill. And again, in 

 the same manner, on the N.E. side of Cherhill Church, where the 

 bank is very steep, and across the line of storm, and the low trees 

 and shrubs growing on it are apparently protected from the gale, 

 and are very little, if at all higher than the churchyard; where (in 

 addition to all this) there was a rick standing on the edge of the 

 bank and broadside to the storm, yet the trees on the bank were 

 crushed down as by an avalanche, which can only be accounted for 

 by the rarefaction of the air in the glen as the storm cloud passed, 

 and then by the rush of heavier air down into it. 



I come now to speak of the hail-stones which accompanied the 

 storm in large quantities, and which from their enormous size and 

 peculiar shapes were almost as extraordinary as the tornado itself: 

 moreover, their forms seem to have varied in different localities ; 

 thus Mr. Spenser of Bowood saw some more resombling flat pieces 

 of ico than hail : they were nearly half an inch in thickness, and 

 from two to three inches in diameter, star-shaped, with rays rang- 

 ing from four to seven in number, and tho rays of different sizes. 

 Others again woro wedge-shaped and about three inches in length, 

 and in some cases several of these were frozen togethor : these 

 hailstones fell clear of the rain-cloud; and Mr. Uowell suggests that 



