Mons. C. Despretz on the Maximum Density of Liquids. 3 



I disposed them so that the variation in the size of the diame- 

 ter took the one direction and the other alternately. In the 

 first experiments the instruments were placed in the midst 

 of a liquid which was gradually made cold; and after it had 

 passed the apparent maximum, the apparatus was left to the 

 calorific influence of the surrounding bodies ; it then became 

 warm, and arrived at the point of departure. By performing 

 the experiment so that the heating kept equal pace with 

 the refrigeration, the error occasioned by the want of coinci- 

 dence between the water thermometer and the mercury ther- 

 mometer was avoided ; the first being always behind the se- 

 cond. I also lessened very much this cause of error by taking 

 the mean of the results obtained; nevertheless I preferred ob- 

 serving in the statical condition. 



After various essays, which need not be enumerated, I 

 adopted the following apparatus. 



It consists of a cylindrical copper vessel, similar to a large 

 eprouvette In this vessel two water thermometers and three 

 mercury thermometers are suspended, the two first alternating 

 with the latter; all the reservoirs are at the same height; 

 the vessel is corked so as to hinder the access of external air. 

 It is then placed in a large earthen vessel filled with a mix- 

 ture at various temperatures, from +16° Cent, to the freezing 

 of the water, which takes place at times at —5°, at times at 

 — 10°, sometimes at —15°, and even at— 20° Cent. We should 

 remember that Gay-Lussac had previously observed water to 

 remain in a liquid state at —12° Cent. 



Each experiment lasted from eight to ten hours, during 

 which from eight to ten numbers were taken. 



The curve of the apparent dilatation is drawn, and then a 

 tangent parallel to the line of the dilatation of the glass is 

 drawn to it ; for the maximum is that point where the abso- 

 lute dilatation of the water is zero, /. e. where the apparent 

 dilatation observed is equal to the effect produced by the con- 

 traction of the glass. The maximum might also have been 

 fixed by the method of calculation which Biot has followed in 

 his Traite de Physique, but we preferred the graphic method, 

 which perhaps indicates better the course of the results. 



The determination of the absolute maximum requires the 

 knowledge of the dilatation of the glass. As the composition 

 of this substance varies more or less, we had to find the dilata- 

 tion of those tubes of which our thermometers were formed; 

 I found it to be : 



Between 28° and 100° Centig. = 0*0000258 



28 = 0-0000255 



— ^ 100 = 0-00002575 



B 2 



