420 
ON THE EXPANSION OE SEA-WATER BY HEAT. 
we see that the three values of the volume we obtained by three independent series of 
experiments were 1-00212, 1-00211, and 1-00206, and that the general formula gives 
1-00209, which is, to within a fraction in the fifth place, equal to the mean of the three 
observations, and certainly cannot be affected with so great an error as 17 in the fifth 
place, which is, however, the difference between it and Prof. Hubbard’s number. 
We may also point out that it appears from Prof. Hubbard’s Table that the point of 
maximum density of average sea-water is 28°-5 F. or — 1°-94C. If we compare this 
result with other observations on record, we find that it is probably between 2° and 3° C. 
too high. Excluding Muncke’s result of — 5°-25, which is probably too low, we find 
that Hespretz obtained — 3 0, 67 (which is probably too high for average sea-water), 
Neumann — 4°-74 in the case of a mixture of sea-water from Trieste, Genoa, and Heli- 
goland, and Rossetti — 3°-90 in that of Adriatic water of specific gravity 1-02814 at 0°. 
The number obtained from our general formula, viz. — 5°T4, is probably too low 
by at least 0 o, 5 C. ; but it is necessary to remember that the formulge we give are based 
on observations between 0° and 36°, and are only applicable to temperatures between 
these limits. 
It is quite evident from the foregoing observations that the laws of the thermal 
expansion of all ocean sea-waters cannot be assumed to be sensibly the same. The 
difference at 15° between the volumes of sea-waters of specific gravities 1"024 and 
1-0278 is 18 in the fifth place; at 30° it is 26; so that even the use of the formula 
corresponding to the mean of the two numbers would, for high temperatures and 
relatively weak or strong solutions, introduce errors in the fourth place of decimals, 
whence it follows that the alterations in the law of thermal expansion depending on the 
salinity of the solutions cannot be neglected. 
