in the Valley of Reposoir) by ‘glaciere’ M. Bourrit meant only a locality where ice is to be found, or a glacier district.  Dr. Boue, however, gives some references to the ’American Journal of Science’ which it is possible to make out by a careful search in the neighbourhood of the volume and page he mentions.  In vol. iv. (1822,—­Dr. Boue says 1821) there is an account by the editor[145] of a natural ice-house in the township of Meriden, Con., between Hartford and Newhaven, at an elevation of not more than 200 feet above the level of the sea.  The ice is found in a narrow defile, which is hemmed in by perpendicular sides of trap-rock, and displays a perfect chaos of fallen blocks of stone.  The defile is so narrow, that the sun’s rays only reach it for an hour in the course of the day; and even the trees and rocks, and beds of leaves, protect the ice from any very material damage.  Dr. Silliman visited this defile on the 23rd July, 1821,[146] with Dr. Isaac Hough, the keeper of a neighbouring inn, and found that the ice was only partially visible, in consequence of the large collection of leaves which lay on it:  they sent a boy down with a hatchet, and he brought up some large firm masses, one of which, weighing several pounds, they carried twenty miles to Newhaven, where it did not entirely disappear till the morning of the third day.  Seven miles from Newhaven, in the township of Branford, there is a similar collection of ice.  In both of these cases, the ice is mixed with a considerable quantity of leaves and dirt.

In the same volume (p. 331,—­Dr. Boue says p. 33), two accounts are given of a natural ice-house near the summit of a hill in the neighbourhood of Williamstown (Mass.).  In the next volume there is a further account of it by Professor Dewey, stating that since the trees in the neighbourhood had been cut, the snow and ice had disappeared each year about the first of August.

In vol. xlvi. (p. 331) an ice mountain in Wallingford, Rutland County (Vt.), is described, which is ordinarily known in the neighbourhood as the ice-bed.  An area of thirty or fifty acres of ground is covered with massive debris of grey quartz from the mountains which overhang it; and here—­especially in a deep ravine into which many of the falling blocks of stone have penetrated—­ice is found in large quantities.  It appears to be formed during the melting of the snow in February, March, and April, and vanishes in the course of the summer, in hot years as early as the last days of June.

These descriptions call to mind the Glaciere of Arc-sous-Cicon, in which many of the features of the American ice-caves are reproduced.  An American photograph is current in this country, in the form of a stereoscopic slide, representing an ice-cave in the White Mountains, New Hampshire; but it is only a winter cave, and in no way resembles any of the glacieres I have seen.  It is merely a collection of long and slender icicles, with beds of ice formed upon stones and trunks of trees on the ground; nothing more, in fact, than is to be seen in any tolerably severe winter in the neighbourhood of a cascade in a sheltered Scotch burn.