AT THE BOTTOM OF THE WATER 431 



temperature below 0° in a vessel or reservoir whose walls may be easily cliilled, 

 is always covered first on its surface by a tliin stratum cf ice, and tbat afterwards 

 the walls and bottom of the vessel also become covered Avith a stratum of ice. 

 II The thickness of this stratum is in direct ratio to the intensity of the cold, 

 and in inverse ratio to the conductibility of the walls ; the stratum of ire itself 

 operates as a wall, which is a bad conductor. III. Th-'-e remains, almost always, 

 at the centre a certain quantity of water which, surrounded on every part by ice, 

 with difficulty loses its latent heat, and does not freeze. Very often, also, a 

 bubble of air, which has been disengaged from the water during its congelation, 

 is comprised in the mass. But it sometimes happens, through intense and con- 

 tinued cold, that the upper ice cracks, and the caloric of the central mass escaping 

 by the fissures, the whole congeals, and the surface assumes a convex and pro- 

 tuberant form, from the expansion of the ice formed at the centre. IV. The 

 plashes or small puddles of water which occur in the roads or fields are promptly 

 covered with a thin crust of ice, which assumes the form of films crossing each 

 other at 30, 60, and 120 degrees, and leaving void spaces between them; the 

 water winch filled them is absorbed by the porosity and capillarity of the sur- 

 rounding earth, and, in turn, becomes congealed. V. It is quite otherwise with 

 water in large masses, for water at 4'^. 44, being specifically heavier than that 

 at 0"^, descends to the bottom, while the water of the surface continues to grow 

 colder, and finally is frozen. VI. In order, then, that ice may be formed at the 

 bottom of great masses of water, it is necessary, first, that the water should be 

 impressed with a movement sufficiently rapid to overcome the superpositicm by 

 strata according to the differences of specific gravity, and to bring to the bottom 

 the cold strata, so that the water may be chilled to 0° quite to the bottom and 

 the inferior walls, growing progressively colder, shall also be brought, to 0° ; 

 secondly, that there shall exist in the midst of the current an obstacle against 

 which the water impinges. 



In effect, whenever there occurs in a current a body forming an obstacle, the 

 collision with that obstacle augments the movement of rotation of the current, 

 and may even occasion a vortex or eddy in the water ; it must be recollected at 

 the same time that behind this obstacle there is a space where the water is in a 

 state of perfect repose — so much so that when the body which forms the obstacle 

 is of considerable volume, a deposit of sand and even pebbles takes place at that 

 point and forms a sort of delta. It is here that the ice of the bottom, the 

 grundvis, is formed, the adhesion of which gradually augments the volume of the 

 obstacle and the eifect produced, until the moment when, by virtue of its smaller 

 specific gravity, the ice is detached, and is borne to the surface of the water. 



At a session of the Society of Natural History of Strasbourg, May 3, 1864, 

 Professor Bertin presented a memoir on the polarization of light by ice, from 

 which it appears that the ice furmed at the bottom of water polarizes light in the 

 same manner as the ice formed at the surface. 



There is a valuable article on ice, by L. F. Koemtz, in the JEjncyrlopa:die of 

 Ersch and Gruber. The specific gravity of ice is there given as follows : ■ Ac- 

 cording to Kraft, 0.905 ; Irwin, 0.937 ; Scoresby, 0.9146, 0.9166, 0.9253 ; Royer 

 and Dumas, 0.950 ; Osonn, 0.9268. Thomson, in his Chemistry, gives 0.2900, 

 which is nearly the mean of the above numbers. 



As regards the form of the crystal of ice, Haiiy thought that it might be de- 

 duced from the octahedron. Brewster {Poggcndorff's Annalcn, t. vii, 509) re- 

 cognizes hexahedrons terminated by three planes. Hericart de Thury {Ann. de 

 Chhnie et Fhysique, xxi, 156) adopts the hexahedral prism. Clarke, ( Trans, of 

 the Philosophical Society of Cambridge, 1213,) in a seemingly very exact ar- 

 ticle, pronounces for rhombohedrons, with angles of 120 and 60 degrees. This 

 is confirmed by Marx and Brewster, {Ann. de Poggendorff, xxxii, 399,) and 

 seems corroborated by our own observation of lines crossed at 30, 00, and 120 

 degrees, as already noticed. See also Scoresby : Cristillisation de la ncige, in 

 the Ann. dc la Chimie et de Physique, t. xviii. 



