2 Mons. C. Despretz 07i the Maximum Detisity of Liquids » 



offered much interest to physicists on account of the phaeno- 

 mena of the temperature of the polar and equinoxial seas. I 

 was led to make experiments on pure water, a question which 

 is also very important from its connection with the determina- 

 tion of the Gramme, because I noticed that those who had 

 occupied themselves with this intricate subject had every one 

 of them left it in considerable uncertainty. Hallstrbm, to 

 whom we are indebted for the latest work on this subject, 

 found by each process a different number, he therefore as- 

 signs as limits 4°'85 and 3°*4. The number upon which he 

 fixed after a detailed discussion of the known results is 4°'l 

 + 0°'3*. We see the uncertainty in which this memoir leaves 

 us. I shall not speak of Rudberg, who gave in to the Academy 

 of Stockholm the same number as that which I made known 

 to the Academy at Paris rather more than a year before. 



Four methods have hitherto been employed in this kind of 

 research ; apparently the most simple consists in weighing a 

 body in water of different temperatures. The necessity of 

 agitating the fluid in order to distribute the heat uniformly 

 renders this method of difficult execution, because this agita- 

 tion necessarily sets the scales in motion. Lefebvre-Gineau, 

 Hallstrbm, and others have employed this. 



In the second the same vessel is weighed full of water at 

 temperatures near to the maximum. Blagden and Gilpin 

 made use of this method. I also have tried it. I have even 

 subjected it to numerous trials; but it is not sensitive enough. 

 This objection may also be applied to the first method. 



We should naturally think that refraction would furnish 

 a very delicate mode of determining this point, but we know 

 from the experiments of Arago, that water, while it is dilated 

 by cold, reflects light more and more powerfully; this fact, 

 which is not less singular than that of the maximum itself, ex- 

 cludes the use of refraction from our researches. We might 

 also determine the temperature of the maximum by the help 

 of the relation discovered by Savart between the temperature 

 and diameter of sheets of water {nappes), but this process 

 would require great practical dexterity in experiments on the 

 efflux of liquids. 



The process which appeared to me to be the most suitable, 

 was to compare the progress of a water thermometer with a 

 mercurial thermometer. For this purpose I constructed six 

 water and four mercury thermometers. All these instruments 

 were divided into parts of equal volume. In order to get rid 

 of the error which arises from the conical form of the tubes, 



* M. Despretz is evidently unacquainted with the last memoir which 

 appeared in Poggendorff 's A7ina/en, vol. xxxiv. p. 220. — (W. F.) 



