I PRESUME that anyone looking at the chart on page 43 of this week's NATURE, or reflecting on the circumstances under which barometric observations at sea are ordinarily taken, will agree with me that it would be wiser to give only two places of decimals, and not indicate a degree of refinement which the observations do not warrant. This point being granted (and even if it is not I shall maintain the same line of argument), I submit that the writer of the article is in error in saying on page 44: “Range corrections for pressure and temperature over the region under discussion are not yet accurately enough known to justify the committee in “correcting” the results on the large chart by hypothetical corrections.”