4 Mons. C. Despretz on the Maximum Densii^ of Liquids. 



which latter number only differs by unity in the third figure 

 from that obtained by Dulong and Petit. The dilatation of 

 glass increases therefore from 0° to 100°. But the increase 

 assigned by Hallstrom is evidently too great, not only since 

 this increase differs widely from the one we have found, but 

 because it is in opposition to the general course of dilatation 

 established by the two physicists above-mentioned. 



The process just described has also the advantage of being 

 the only one applicable to water at low temperatures, and to 

 solutions whose maximum is below their freezing point in a 

 state of agitation, which is the case in all solutions somewhat 

 concentrated. The maximum may also be determined by a 

 process independent of the dilatation of thr glass. This pro- 

 cess is founded on the fact that in a liquic' mass the strata of 

 which are at unequal temperatures, those molecules having 

 the temperature of the maximum sink while the others rise. 

 I have considerably modified this process, previously em- 

 ployed by Hope, Tralles, Rum ford and Hallstrom*. The 

 latter thought, on account of the discordance of his results, 

 that this method ought to be rejected. In fact it is, as it was 

 employed by them, more adapted to prove the existence of a 

 mascimuvi, than to determine its temperature with accuracy. 

 The following is a short description of the process employed 

 by me : 



We took a porcelain vessel capable of holding above thir- 

 teen pints; — a larger vessel would have required too much time, 

 for in five hours this did not become more than a few degrees 

 colder in the air; — the temperature of the fluid was several 

 degrees below zero. A smaller vessel, an eprouvette for in- 

 stance, would have become cold too rapidly, and the conse- 

 quence would be that the thermometers would indicate a tem- 

 perature too much at variance with the real temperature of 

 the liquid. 



Four thermometers, the stems of which were passed through 

 the side of the vessel, were placed horizontally in a vertical 

 plane, two on the one side and two on the other. The 

 distance between the first thermometer and the bottom of the 

 vessel amounted to 54- millimetres, this being also the distance 

 between each two thermometers. The entire height of the 

 vessel was 270 millimetres and the diameter 160. The ther- 

 mometers alternated, so that No. 1 and No. 3 came out on 

 one side. No. 2 and No. 4 on the opposite one. The middle of 

 the reservoir of each thermometer was in the axis of the vessel. 

 The vessel was suspended by three cords of equal length, 

 with its axis in a vertical position. As soon as it was filled 



* Poggcndorff's Annalcn, vol. ix. p. .5oO.- (W. F.) 



