58 



still more conclusively on this point.* He poured water into iron boilers 

 which were insulated from the influence of soil temperature by being 

 elevated on trestles, and they were at the same time fully exposed on 

 all sides to the action of a freezing temperature. The inside of one 

 boiler was smooth, while another was interiorly coated with a few chips 

 of iron and wood. Ice was formed in both boilers along the sides and 

 bottoms, as well as on the surface, while the middle was still occupied 

 with unfrozen liquid. From a comparison of both vessels, it seemed 

 that the inequalities on the interior of the second greatly favoured the 

 formation of rough crystalline bunches of ice. The residual unfrozen 

 liquid suggests an explanation of the difficulty to which I have alluded 

 with reference to the exclusive freezing of a river at surface and on its 

 bed. This phenomenon is indeed only a particular instance of a 

 general thermological law — namely, that all substances in passing from 

 the liquid to the solid state evolve a certain amount of latent heat. It 

 is thus, after various metals, sulphur, and other substances commence to 

 crystallize from a state of fusion, we find, on breaking the crust of solid 

 matter first formed, that a residuum of liquid enclosed in a solidified 

 matrix may be decanted off. "With regard to water, this process has 

 been very clearly described by Professor Curtis, of Queen's College, 

 Galway ; j and, he refers, moreover, to the low conductivity of water for 

 heat as an agency for confining the communication of the latent heat 

 of congelation to the adjacent particles. If, therefore, from the pre- 

 valence of conditions favourable to freezing both at surface and along 

 the bed of the still parts of such a river as the Dodder, ice should be 

 formed in these positions, its growth will in itself interpose obstacles 

 to the freezing of the middle waters. 



The explanation here given of the formation of ground ice is, in 

 substance, the same as that propounded several years ago by the late 

 M. Arago ;j but I venture to believe that there are some peculiar fea- 

 tures in the phenomena which I have described, which may further 

 elucidate the whole question. It cannot be maintained, as has been 

 done, according to Arago, § by one of our countrymen, that freezing at 

 the bottom of cold still and clear water arises from the greater facility 

 presented by still water as compared to moving water for the trans- 

 mission of radiant heat from the underlying bed. In a discussion of 

 the physical properties of water with reference to terrestrial climate at 

 different geological epochs, published in 1859,|| I alluded to the manner 



* "Memoires de la Societe des Sciences Naturelles de Strasbourg," tome vi. 



f " On the Freezing of Water at Temperatures lower than 32° F. :" " Philosophical 

 Magazine" for December, 1866. 



\ Arago, " GEuvres," vol. viii. § Loc. cit., p. 176. 



|| " Atlantis," vol. ii., p. 208, January, 1859. Some of my conclusions regarding 

 climate having been lately reproduced as new, I may be excused for briefly stating the 

 properties of water to which I appealed when attempting to establish these conclusions : — 

 1, its great capacity for heat ; 2, its mobility; 3, the influence of evaporation and con- 

 densation ; 4, the impermeability of water to obscure heat. The first three are distinctly 

 adduced in section 2 of my essay (p. 210): while the 4th, now noticed in the text, is 



