150 CUBES OF ICE FOE EXPERIMENT. [October, 



ways, upwards by evaporation, and below by the cus- 

 tomary process of freezing ; but here again we have a 

 very stubborn fact to contend with. By all the experi- 

 ments made, until I became sick of reported breakages of 

 thermometers, I could not detect a well-attested differ- 

 ence between the temperature at the bottom, in seven 

 fathoms, and that at the surface. If heat ascends, and 

 the cold will not descend through ice below 32, the sea 

 preserving 29 to 29 '5, why does it not freeze to the 

 bottom ? and why does it accumulate on ice freezing at 

 28, but only at 32 when frozen ? Our experiments 

 have not resolved this difficulty, every trial has been 

 marred by some mistake. 



Cubes of Ice. Experiments were now commenced on 

 cubes of sea-ice, and, as the thickness increased, they 

 were extended to the mean upper, mean centre, and 

 mean lower masses. Nor did it cease there. These 

 cubes, duly numbered, were exposed to the upper-deck 

 temperature during winter, and weighed at particular 

 seasons to determine their changes by exposure, evapo- 

 ration, etc. The water resulting from thawing these re- 

 spective divisions of the ice was bottled, as well as that 

 at the surface, for future examination. The atmospheric 

 air, in well-dried, stoppered bottles, and covered with 

 leather and bladder, was also obtained. 



On the 25th of October, a cubic structure was formed 

 with snow bricks, of twelve feet in length, ten wide, and 

 six high, the crevices being sealed with snow and water. 

 In this mass five Six's self-registering thermometers were 

 inserted, equidistant in height as well as lateral distance 

 from the air, as follows : 



