in Fluids. " ^oi 



These descending currents meeting, in the region of 

 the constant temperature of 40^, with those cold currents 

 which ascended from the surface of the ice, it seems very 

 probable that the ascending currents, on the motion of 

 which the melting of ice depends, were checked by this 

 collision. 



By retarding the cooling of the hot water above by 

 wrapping up the jar in a warm covering, the velocity of 

 the descending currents was of course diminished ; but 

 when this was done, the results of the experiment shewed 

 that the melting of the ice was accelerated. 



When, the jar being naked, the cooling of the hot 

 water, and consequently the motions of the descending 

 currents, were rapid, no more than about 542 grains, or 

 at most 575 grains, were melted in 30 minutes ; but 

 when the jar was covered with a warm covering, 634 

 grains, and in one experiment (that No. i^S^ 747 grains, 

 were melted in the same time. 



As plunging the jar into a cold mixture of snow and 

 water could not fail to accelerate the cooling of the hot 

 water in the jar, and consequently to increase the ra- 

 pidity of the descending currents in it, ought not this to 

 embarrass, in an extraordinary degree, the ascending cur- 

 ^•ents of ice-cold water from the surface of the ice, and 

 to diminish the quantity of ice melted ? This is what 

 the following experiments, compared with the results of 

 those No. 39, No. 40, and No. 41, will shew. 



