at its Maximum Density. 263 



the top or brim of the cylindrical vessel in such a manner 

 that the centre of the circular hole coincided with the 

 axis of the cylindrical vessel, when the short cylindrical 

 projection belonging to the ball was introduced into that 

 hole, the ball was firmly supported in its proper place. 



The ball was placed in such a position that the end of 

 the conical projection was immediately over the cork cup, 

 at the distance of i^ inches above the level of its brim 

 and consequently \ of an inch above the upper part of 

 the bulb of the small thermometer which lay in this cup. 



The quantity of cold water in the cylindrical vessel 

 had been so regulated beforehand that when the conical 



o 



point was entirely submerged, the surface of the water 

 was on a level with the base of this inverted .cone, so 

 that the whole of the cylindrical part of the projection 

 was out of the water. 



I knew that the particles of ice-cold water which were 

 thus brought into contact with the conical point could 

 not fail to acquire some small degree of heat from that 

 relatively warm metal, and I concluded that if the par- 

 ticles of water so warmed should in fact become heavier 

 than they were before, in consequence of this small 

 increase of temperature, they must necessarily descend in 

 the surrounding lighter ice-cold liquid, and as the heated 

 metallic point was placed directly over the cork cup, 

 and fixed immovably in that situation, I foresaw that 

 the descending current of warm water must necessarily 

 fall into that cup and at length fill it, and that the pres- 

 ence of this warm water in the cup would be announced 

 by the rising of the thermometer. 



The result of this very interesting experiment was 

 just what I expected: the conical metallic point had not 

 been in contact with the ice-cold water more than 20 



