Density of Water with the Temperature. 107 



this decrease only extends to a temperature approaching 63°, 

 and that beyond, for water as for other liquids, ja increases. 

 viz. : — 



By experiment. By formula No. 7. 



0°. /*10 6 = 50*3 ^10 6 =s50-5 



10°. 47-0 4:7-3 



20°. 44-5 44-6 



30°. 42-5 42-4 



40°. 40-9 40-7 



50°. 39-7 39-6 



60°. 38-9 39-0 



70°. 39-0 38-9 



80°. 39-6 39-3 



90°. 40-2 40-2 



100°. 41-0 41-7 



In order to be able to deal with experimental data for tem- 

 peratures exceeding 100°, I have expressed the variation of //, 

 by a parabola : — 



p = 10- 6 (5O49 -0-348 t + 0-0026 **)•■ . . (7) 



Since the quantity p represents some hundred-thousandths 

 of the volume, it is evident that in determining the density of 

 water (just the same as for all other liquids) it is impossible 

 to obtain results agreeing to a millionth, even at ordinary 

 atmospheric pressures, if (as is usually done) we neglect the 

 influences of compressibility and pressure. Supposing, for 

 example, that the experimenter had determined the volume 

 for water at 100° as 1*043212, operating at a pressure of 1 at- 

 mosphere; and on another occasion with this pressure decreased 

 by one tenth of an atmosphere, he ought to find 1*043212 x 

 1*0000042 or 1 •04321b', if his determinations attain an accu- 

 racy of a millionth. The influence of compressibility is par- 

 ticularly notable when determining the density or volume of 

 water beyond 100°, because at such temperatures the pressures 

 are unavoidably considerable. In this respect the first, so to 

 speak, reconnoitring determinations were made by Sorby * 



* Sorby (Phil. Mag. 1859, xviii. p. 81) made his determinations in a cylin- 

 drical sealed tube, and compared the volumes of water and of some 

 saline solutions. His data refer to vapour-pressures p of water at the 

 temperature indicated. 



f=120° 140° 160° 180° 200°. 



F = 1-913 3-57 6-12 9*93 15-88 Atm. 



V = 1*05988 1-0796 1-10186 1-12676 1-1543. 



Judging from the mode of observation, the degree of accuracy hardly 

 exceeds +0-005. 



In 1860, being unaware of Sorby's determinations, I made a series oi' 

 determinations of the expansion, above their boiling-points, of water, 

 ether, and alcohol (Mendeleeff, Liebig's Ann. cxix. p. I). My experiments] 



