Note on Determination of Heavy Water in Ocean Waters
In a recent article entitled Heavy Water Content of Deep Sea Water (Trans. and Proc. Royal Society of New Zealand, 66, 97–98, 1936, F. P. Worley questions the results obtained by the undersigned (Jour. Amer. Chem. Soc., 57, 400–404, 1935) on the basis of the temperature control employed by us and the accuracy with which temperatures could be maintained during the experiments. It is rather unfortunate that Dr. Worley has evidently misread our paper. No reference is made to measurements of absolute density. All of our data are based on differences in density. We made use of a differential method for the determination of density, in which the density of the sample of water to be measured was compared with that of a reference water retained in the apparatus. While it is true that the density varies with the temperature, it was not essential, in the method employed by us, that the temperature be known as accurately (±0.02° C.) as was actually the case. However, it was essential that the temperature not vary during a single comparison by an amount which was greater than 0.001° C. This was done. The apparatus was so constructed that it served as an indicator of changes in temperature and thus enabled us to make comparisons with the temperature not varying by as much as 0.001° C. That conditions for accurately measuring differences in density to 0.1 p.p.m. obtained is proved by the agreement between duplicate determinations. The average difference in duplicate determinations was 0.1 × 10−6 grams per cubic centimeter and the maximum difference was 0.27 × 10−6 grams per cubic centimeter.
Gilfillan (Journ. Amer. Chem. Soc., 56, 406–407, 1934) found sea-water from the Atlantic (surface and 4500 meters) to be more dense than Cambridge tap-water by 2.3 p.p.m. Greene and Voskuyl (Jour. Amer. Chem. Soc., 56, 1649–1650, 1934) found surface sea-water to have a density 1.8 p.p.m. greater than tap-water. Washburn and Smith (Bur. Standards J. Research, 12, 305, 1934) showed sea-water to be about 2 p.p.m. more dense than ordinary water. Ball (Biol. Bull., 65, 371, 1933) found no difference in density in sea-water from 5000 meters and the surface within the accuracy of his determinations (±5 p.p.m.). Investigations of the density of other naturally occurring waters have rarely disclosed differences greater than 6 p.p.m. In view of these results, and of those of the undersigned, it is difficult to see how significant variations could be expected to be found, using a method sensitive to differences of only 10 p.p.m.
H. E. Wirth.
T. G. Thompson.
C. L. Utterback.
Oceanographic Laboratories, University of Washington, U.S.A.