46 



On the Maximum Density oftVater. 



The point of maximum density indicated by the above 

 series will be situated at about 39°" 134. 



We now pi-oceeded to apply a severer test to our method. 

 In the next series of experiments we arranged matters so that 

 the temperature of the water in one vessel was only a degree 

 and a half higher than that of the other vessel. The expan- 

 sion of water increasing as the square of the temperature from 

 that of the maximum density, it was obvious that the current 

 in the trough would be much more feeble than in the former 

 experiments. We therefore allowed the vessels to be in per- 

 fect repose for six minutes before we introduced the glass 

 ball, and we afterwards watched its motion for four or five 

 minutes. 



Series 4. 



The position of the point of maximum density according 

 to the above fourth series of experiments will be at 39°'091. 



I. Point of maximum density of pure water 39*102 

 II. ... ... ... ... 39-078 



III. ... ... ... ... 39-134 



IV. ... ... ... ... 39-091 



Mean ... 39-101 



Although in the different series of observations there are 

 several irregular results, there is on the whole sufficient con- 

 sistency among them to enable us to receive 39°-l, the mean 

 of the four sets of observations, as the actual point of maxi- 

 mum density. We think it highly probable that this tem- 

 perature is within one-hundredth of a degree of the truth : 

 it certainly cannot be more than one-twentieth of a degree in 

 error. We were prevented by the mildness of the season 

 from extending the experiments further, but we doubt not 

 that by repeating them more frequently we should be able to 

 bring the determination of the point to any required degree 

 of accuracy. The result arrived at by Despretz from a very 

 extensive series of experiments, with an apparatus similar to 

 that employed by Hope, is 39*^-176 *, which agrees veiy well 



• Annalcs de Chimic, 183[), t. Ixx. p. 45, 



