Reply to Note on Determination of Heavy Water
The only criticism made of the determinations of Wirth, Thompson and Utterback was “A difference of 0.1 × 10−6 in the density of water at 25° corresponds to a temperature difference of 0.0004° and it is difficult to see how the temperature could be controlled and determined to the required degree of accuracy in an air-bath at 25 × 0.02°, although the authors state that through insulation the essential part of their delicate apparatus remained constant within 0.001°.” I find that I had not misread their paper and am still of opinion that they have not shown in their paper, or in the above note for that matter, that the temperature was in fact controlled sufficiently accurately to justify determinations of density differences as accurate as their results would imply. Amplification of my brief criticism is apparently necessary, though it clearly infers that differences of temperature and of density are involved, not actual temperatures and absolute densities. The accuracy of the density differences determined by Wirth, Thompson and Utterback is dependent on the minuteness of the difference in temperature between the water always retained in one side of their apparatus and samples introduced from time to time into the other side. The difficulty of attainment of temperature equilibrium, particularly in unstirred liquids in an air-bath, is very great and is well known. It is consequently unfortunate that in the paper the only reference to temperature was as follows: “The apparatus was kept in an air-bath at 25 ± 0.02°. It was further insulated so that the temperature of the vertical tubes remained constant to within ± 0.001° after temperature equilibrium was attained.” The close agreement of “exact duplicates” does not necessarily prove that temperature equilibrium was attained.
If the approach to equilibrium is 0.001° an experimental error of ± 0.25 × 10−6 in the density is introduced. If this error were shown in Table II in which the vertical distribution of isotopic waters is recorded, all the values of density difference would be found to diverge from a constant value by less than 0.1 × 10−6 and the only valid conclusion would be that within experimental error there is no divergence from constancy.
I wish to make it quite clear however that I do not contend that the experimental error could not have been less than 0.25 × 10−6 but that Wirth, Thompson and Utterback in their excellent and valuable paper have in my opinion not shown that it was less. It is quite possible that temperature equilibrium was attained much more effectively than they have indicated. Their apparatus and method offer possibilities of investigating the difficult problem of the attainment of temperature equilibrium.
With reference to the final paragraph of the note the very low values of heavy water in the sea found by all observers is confirmatory of the mixing of ocean waters as found by other oceanographic studies. In undisturbed and tranquil deeps, if such exist, a much higher content of heavy water is to be expected. My experiments were carried out solely for the purpose of looking for concentrations of heavy water greater than 10 p.p.m. indicative of undisturbed deeps.
F. P. Worley.
University College, Auckland.