28 ROYAL SOCIETY OF CANADA 



jVIr. Keefer, in the discussion following Mr. CI. H. Henshaw's paper 

 on frazil, published in the Transactions of the Canadian Society of Civil 

 Engineers, for March, 1887, after describing an observation made by him- 

 self on the condition of the river opposite Montreal, before there was any 

 floating ice, during a very cold day, saj's : "In this condition of the river 

 the water, no doubt, at the deepest point, is loaded with ice spicules to 

 the bottom, densely and uniformly distributed throughout the whole 

 mass, and would supply the raw material for the formation of anchor ice 

 at the bottom, whenever the latter was prepared to receive it." He goes 

 -on to state that " ice will attach itself to ice or to other frozen bodies, but not 

 to the unfrozen bed of a river." The '' frozen bed of a river " must be taken 

 to mean that there is a layer of ice already formed. It becomes, then, a 

 matter to investigate how this first layer appears. 



With the water at the freezing-point, the bottom cannot possibly be- 

 come cooled below O" C, except by radiation. There is indeed always a 

 tendency for it to become warmed above the freezing-point, on account of 

 the slow conduction of heat from beneath. How, then, can the bottom be- 

 come cooled sufficiently to cause an abstraction of the latent heat in the 

 water in immediate contact, necessary to cause ice to form ? The bottom 

 at 0° C. cannot radiate heat into the water at 0°. 



Fvom the experiments by Prof. Tyndall on radiant heat, it was shown 

 that heat rays may be passed to a moderate extent directly through water 

 or clear ice. In this case there is no reason why the bottom of a river 

 should not radiate heat straight through the water and atmosphere into 

 space. From the cooling of the bottom by this means the first layers of 

 ground ice would grow with considerable rapidity. The growth would, of 

 course, be continued by radiation from the surface of the ice itself, as well as, 

 during the extreme weather, by the sticking of fine surface formed ice 

 carried down by currents. Eadiation from the bottom would be some- 

 what retarded as the ice increased in thickness. "When the ice became so 

 thick as to practically check radiation, on account of its granular texture, 

 heat conducted through from beneath would serve to slowly bring up 

 the anchor ice, as well as to prevent any further formation of ice. 



From the results of the observations on soil temperatures,^ carried 

 out by Professor Callendar in conjunction with Professor McLcod, in the 

 McDonald Physics Building, with delicate electrical resistance thermom- 

 eters placed at different depths in thegi'ound, it has been shown that the 

 conduction of heat is exceedingly small. It may be easily calculated, from 

 their work, that in one hour there would be enough heat transmitted to 

 the bed of a river sufficient only to melt a layer of ice not more than -05 

 mm. in thickness. It will take very little radiation, then, on a clear night, 

 to overcome this minute source of heat. 



1 Transactions of the Royal Society of Canada. Vol. I., Sec. III., p. 63, 1895, and 

 Vol. II., Sec. III., p. 109, 1896. 



