On the night preceding the 31st of August, as we steamed past Svjatoinos, a peculiar phenomenon was observed. The sky was clear in the zenith and in the east; in the west, on the other hand, there was a bluish-grey bank of cloud. The temperature of the water near the surface varied between +1 deg. and +1.6 deg., that of the air on the vessel between +1.5 deg. and +1.8 deg.. Although thus both the air and the water had a temperature somewhat above the freezing-point, ice was seen to form on the calm, mirror-bright surface of the sea. This ice consisted partly of needles, partly of a thin sheet. I have previously on several occasions observed in the Arctic seas a similar phenomenon, that is to say, have observed the formation of ice when the temperature of the air was above the freezing-point. On this occasion, when the temperature of the uppermost stratum of water was also above the freezing-point, the formation of ice was clearly a sort of hoar-frost phenomenon, caused by radiation of heat, perhaps both upwards towards the atmosphere and downwards towards the bottom layer of water, cooled below the freezing-point. The whole day we continued our voyage eastwards with glorious weather over a smooth ice-free sea, and in the same way on the 1st September, with a gentle southerly wind, the temperature of the air at noon in the shade being +5.6 deg.. On the night before the 2nd September the wind became northerly and the temperature of the air sank to -1 deg.. Little land was seen, though we were still not very far from the coast. Near to it there was a broad ice-free, or nearly ice-free, channel, but farther out to sea ice commenced. The following night snow fell, so that the whole of the deck and the Bear Islands, which we reached on the 3rd September, were sprinkled with it. Hitherto, during the whole time we sailed _along the coast_, we had scarcely met with any fields of drift-ice but such as were formed of rotten, even, thin and scattered pieces of ice, in many places almost converted into ice-sludge, without an "ice-foot" and often dirty on the surface. No iceberg had been seen, nor any large glacier ice-blocks, such as on the coasts of Spitzbergen replace the Greenland icebergs. But east of Svjatoinos the ice began to increase in size and assume the same appearance as the ice north of Spitzbergen. It was here, besides, less dirty, and rested on a hard ice-foot projecting deep under water and treacherous for the navigator. The ice of the Polar Sea may be divided into the following varieties:--