A change rarely came unexpectedly. Often for days beforehand “mare’s tail” clouds, with a hard wavy outline, would float up against the faint moonlight in the southern sky, and spread themselves into wings and fingers over Robeson Channel. Then, with a sudden gust from the south, and a mist of flying snow from the land, the temperature would rise. Mercurial thermometers would thaw, and soon register as faithfully as spirit instruments beside them. After a while the wind begins to come more and more from the westward. The thermometers remain high, but the wind feels piercingly cold wherever it can find a way inside our sealskins. While the storm lasts, it is impossible to go outside the ship. Whirling snow hides everything. Even on deck exercise is uncomfortable, for powdery snow floats in through every chink in the carefully-closed tent-like awnings. Notes on the instruments on shore have to be suspended, for no one could force a way as far as the beach through the darkness and whirlwind of drifting snow; and if they could, they would find the observatories so buried that it would take several hours to dig out their doorways. Even the thermometers within seventeen feet of the ship were not always easily registered. Upon one occasion the officer in charge of the meteorological work had to confess himself beaten, after two determined attempts to reach and register them. In twenty-four hours or more the storm lessens, and gradually dies away to a gentle breeze from the northward; and with it the temperature declines, until it is as cold or colder than before. A striking change of this sort came in December. From thirty-five degrees below zero, the thermometers rose rapidly with a gusty southerly wind till the temperature reached the freezing-point. This strangely warm wind cannot have travelled far in contact with the frozen earth, for it was being rapidly cooled. The quick changes, with every puff of wind, suggested the advisability of trying what the temperature was in the air overhead, and it was discovered that the higher we climbed up the rigging the warmer it got. The main-top was three degrees warmer than the deck at the same instant, and a thermometer secured high aloft in the cross-trees actually registered + 36°—a temperature which can hardly be accounted for by supposing that the wind was warmed by passing over pools of open water in Robeson Channel or Smith’s Sound. At times, when the air was undergoing rapid changes of this sort, it was striking to find that, by boring a hole into the ice with an auger, it was possible to get down past zero, and reach the temperature of yesterday or last week before coming to + 28°.3, the steady temperature of the Polar Sea beneath.