Maximum Density of Water. 389 



4°*07 C. by Rosetti, at 4°-l C. by Hallstrom, at 4°'08 C. by 

 Kopp, and at 4 o, by Despretz ; but here the maximum point 

 of the irregularity in the curve is at 4°*7 C. There can be 

 no doubt that such is the case, as several separate cooling 

 determinations gave this temperature as the maximum, it 

 being also corrected for the thermometer, which gave the 

 freezing-point of water as +*08° C. Also the rate of cooling 

 of oil was taken by the thermometer, and was found to be 

 quite regular ; so the difference could not have been due to a 

 thermometric error. 



In these experiments the water was not stirred ; and it is 

 to this fact that the great irregularity in the curve at 4°'7 C. 

 is due. The cooling effect produced in the outer layers of 

 water in the flask mainly reach the inner layers of water and 

 the thermometer-bulb by convection-currents set up by the 

 change in the density of the water with change of temperature. 

 The inside layers will therefore always have a slightly higher 

 temperature than the outside layers. When the temperature 

 of the outside layers has cooled to near 4° C, when the density 

 no longer increases with fall of temperature, the convection- 

 currents will be stopped, and at a slightly lower temperature, 

 when the density decreases, will begin to be set up in an 

 opposite direction. This explains why there is a sudden 

 pause in the curve for the rate of cooling, and also why 

 directly after this the temperature sinks rapidly for a degree 

 or tw T o and then sinks much less rapidly ; for this is due to 

 the cold layers, which had been accumulating while the cooling 

 almost stopped, being suddenly brought into contact with the 

 thermometer-bulb. This experiment, therefore, merely serves 

 to show the change in the density of water at about 4° C, 

 and that the cooling effect reaches from the outside to the 

 inside of the flask mainly through convection-currents, and not 

 by radiation or conduction. 



Experiments were then made on the cooling of water when 

 stirred. A beaker containing 100 cub. cent, distilled water was 

 suspended in another large beaker surrounded by a freezing- 

 mixture. The smaller beaker was closed by a tightly-fitting 

 bung, through which passed the thermometer and a light 

 glass-rod stirrer. This was moved up and dow r n regularly at 

 intervals of about four seconds, during the whole period of 

 cooling, by means of a cord passing over two pulleys. The 

 second curve in the figure represents the rate of cooling. The 

 curve above 14° is quite regular ; and this curve, produced on 

 the supposition that the curve is regular, gives the dotted-line 

 curve in the figure. It will be seen that there is no sudden 

 change in the curve at any point, but that the irregularity 



