Density of Water with the Temperature. 123 



the different series of observations often present differences to 

 the amounts of ten-thousandths. 



12. Weidner (Pogg. Ann. 1866, cxxix. p. 300) applied the 

 volumetric-therm ometric method for the determination of the 

 expansion of water between 0° and —10°. He determined 

 the coefficient of expansion of glass, by means of mercury, 

 between 0° and 90° for vessels blown out of the same glass, 

 and obtained very discordant results from two determinations, 

 viz. 0*00002625 and 0*00002424. He only adopted the former, 

 but this gave larger volumes than found by other observers, 

 which is especially apparent at 0°. 



13. Kosetti (Pogg. Ann. Erganz. Band, v. p. 265), in 

 1869, published a fresh series of determinations for the density 

 of water, which were made by a combination of the volumetric 

 and gravimetric methods. He determined the coefficient of 

 expansion of glass according to Regnault's data, and found it 

 to increase with a rise of temperature. His observations were 

 expressed by a formula of the form : 



Y t =l + a{t-4:) 2 -o(t-4:) 2 - Q + c(t~4;)% 



or else in a formula where the last term (£ — 4) is not raised 

 to the cube, but to the 3*2 power. In its latter form, 

 Posetti's formula recalls that given by Hagen. 



The figures given in Table I. are taken direct from the final 

 results of the different observers, and without doubt contain 

 some errors which in course of time will be capable of correc- 

 tion, so as to render the values for the volumes of water more 

 accurate. Such corrections, or a revision of the mean of 

 equally trustworthy determinations, have been undertaken 

 more than once, and the results thus obtained are brought 

 together in Table II. Although I consider it right to cite 

 these results, and even myself proposed, in 1884, a similar 

 revision for corrected averages, yet at the present moment, 

 after having studied the subject more closely and recognizing 

 the insufficiency of many of the corrections, 1 do not think it 

 necessary to dwell especially upon this question, as, in the 

 absence of new determinations, notably for the expansion of 

 mercury and glass, it is impossible to hope to add to the 

 trustworthiness of what is already known. 



In Table II., in the first line, Biot's figures are given 

 for their historical interest. He calculated them from an 

 aggregate of the data extant at the beginning of the present 

 century. I cite them from Geliler's Physik. I J r orterbuch ( 1 825, 

 i. p. 616). 



The following line is occupied by the figures calculated by 

 Hallstrom in 1835 (Pogg. Ann. xxxiv. p. 24), when he became 



