6 Mons. C. Despretz on the Maximum Density of Liquids. 



Hallstrom indeed obtained by the process of Tralles a 

 smaller number in heating than in cooling, but as the differ- 

 ence amounted to 1°*3, he thought this method offered litde 

 exactness. 



If instead of taking the mean of the higher temperatures 

 at 4-°, and that of the lower temperatures, we had taken the 

 mean of all the temperatures relative to one curve, we should 

 have obtained 3''-988 instead of 3°-982 : difference 0°-006. 



The results obtained with the water thermometers, that is 

 to say, by the first method, were as follows : 



Seven experiments with one tube 3°'99 C. 



Seven another 4* 02 



Two a third 4*01 



Two a fourth 3- 96. 



The mean from these eighteen experiments is 4° C, which 

 agrees within two hundredths with the result of the former 

 process. 



Before and after each experiment the zero of the thermo- 

 meter was examined. This verification is absolutely neces- 

 sary, because the zeros of thermometers, even of those which 

 have been constructed for a long time, differ when these in- 

 struments are kept for some time at a low or high tem- 

 perature. We shall have occasion to return to this important 

 fact on another occasion. 



So many contradictory results have been obtained on the 

 subject of the maximum of density of pure water, that it is 

 quite unnecessary to mention here in what these experiments 

 may be regarded as more nearly approaching to the truth. 

 They occupied me for a year. I constructed, I graduated 

 all the instruments myself. The weighings were performed 

 with the greatest care. Fearful of partial errors, all the re- 

 sults were represented by drawings on a very large scale. 

 I lay before the Academy a few only of the numbers and 

 curves. Although we cannot answer for the hundredth part 

 of a degree, considering the extreme mobility of glass in- 

 struments, we yet remark that the difference of the single re- 

 sults with 4°, a difference which in general has amounted to 

 some hundredths, never surpassed 0°*1, and that the two pro- 

 cesses, which have not the least relation with one another, have 

 furnished sensibly the same result. Nevertheless, on account 

 of the importance of the subject, I shall have the honour in 

 a short time of laying before the Academy some experiments 

 made by a process not yet described, and employed at very 

 low temperatures. 



This memoir closes with a table of the dilatation of water 

 from degree to degree, from the maximum to the boiling 



