determined by five different thermometers, namely, the one used for the investigation, one standard by Baudin graduated to 0'02 C., one by Delaunay, and two by Becker, these last three graduated to O'l C. All these instruments were calibrated in our previous investigation, and have subsequently been submitted again to the same process with the same results.
In each case two independent determinations of the water and acid contents were made, so that either four or two results were obtained according as all the determinations or one in each pair were different. The following results were obtained for the density at 24'2/4 :
 | A table should appear at this position in the text. See Help:Table for formatting instructions. |
Thermometer. I. Baudin.
II. III.
Casella. | Delaunay.
IV.
Becker (i).
V.
Becker (ii).
Densities. . . <
1 -26730
1 -26731
1 -26732 ! 1 -26735
1 '26734 j 1 -26734
1 -26736
1 -26737
1 -26737
1 -26732
1 -26737
1 -26734
1 -26739
1 -26730
1 -26730
1 -26732
"Mean
1 -26731
1 -26733 ! 1 -26735
1 -26735
1 -26730
The mean of all the above determinations is 1-26733, with a probable
error of 0-00002 calculated from the ordinary formula
0-674
though it would appear probable from the results that the thermo- meters I, II, and V gave identical readings, but different from III and IV ; this difference is probably equal to about 0'02 C. However this may be, the errors accumulated by the variations of one instrument would not affect any conclusions, and the method, as regards tempera- ture, gives accurate results.
A number of duplicate experiments were made in which one or more of the following were varied : (i) stock acid ; (ii) pyknometer ; and (iii) standard alkali ; the difference in the values obtained did not exceed one or two units in the fourth place of decimals, which would cause no appreciable error in the contractions.
Further, it is most probable that the divergence in values obtained for the contractions of acidic and saline solutions, to which attention has recently been drawn,* arises rather from errors of analysis and unequal amounts of impurities in materials, than from density deter- minations with all their consequent errors.
- Wade, ' Journ. Chem. Soc.,' 1899 (Trans.), p. 254.
