120 Prof. D. Mendeleeff on the Variation in the 



St. Petersbourg, t. i. p. 249), in Heidelberg, made a determi- 

 nation of the expansion of numerous liquids by the volumetric 

 or thermometric method, which was afterwards adopted by 

 Despretz, Kopp, Pierre, and many others. Although 

 Muncke's results were communicated to the Academy in 

 1828, they only appeared before the public in 1831, ?. e. 

 simultaneously with the investigations of Stampfer. Muncke 

 determined the expansion of the vessel by means of mercury, 

 taking the value given by Dulong and Petit (J^). Muncke's 

 results for low temperatures (0° to 30°) are greater than the 

 actual values, and those for higher temperatures (40° to 100°) 

 are less. This is due to the insufficient accuracy of calibration 

 and to the determination of the coefficient of expansion of glass. 

 It should be observed that most of the results of Muncke's 

 researches appear inaccurate when compared with recent 

 researches. 



3. Stampfer (Pogg. Ann. xxi. p. 116) in 1831, in Vienna, 

 determined the expansion of water hydrostatically by weigh- 

 ing a brass cylinder, whose linear expansion w~as previously 

 determined and found to be 0*001920 between 0° and 100 6 . 

 The determinations were conducted between —3° and +40°, 

 and were expressed (taking the volume at 4° as unity) by the 

 formula S^ = S + a£— bt 2 + cf — dtf. By the method of least 

 squares the constants w-ere : S = 999887, a = 60'932, b = 

 8*4236, c = 0-0580, and d = 0*0001207, on multiplying by 10 6 . 

 Temperatures below zero appear for the first time in Stamp- 

 fer's researches. For —3° he gives the volume 1*000373 ; 

 and since the difference of the volumes at —3° and —5° 

 equals 0*000275, I have introduced the number 1*000648 for 

 — 5°. The coefficient of expansion for brass given by 

 Stampfer is too large, and hence the volumes exceed the true 

 values. 



4. Despretz (Ann. de Chemie et de Phys. t. 70. pp. 23, 

 47), in Paris, 1837. His determinations made at temperatures 

 from —9° to +15° had chiefly in view the study of the 

 densities near 0°. Despretz made but few determinations for 

 temperatures from 20° to 100°, and only gave them to 

 hundred -thousandths. But even these must be considered 

 as among the most trustworthy up to the present date. The 

 coefficient of expansion of glass was only determined for a 

 portion of the dilatometers by means of mercury with Dulong 

 and Petit's figures. From 0° to 28°, for glass, £ = 0*0000255, 

 and from 0° to 100% was equal to 0*0000258. For a long time 

 Despretz's figures were in general use ; and if they have been 

 eventually replaced by more recent ones, such a change has 

 not really any firm foundation. Generally speaking, our 



