Melville was by no means recovered from his long exposure, and his frozen limbs caused him great suffering, but nevertheless he went back over the track of Nindemann and Noros step by step. On November 10, the natives who had accompanied him announced they must return as the provisions were exhausted, but Melville commanded them to go on, declaring they would eat dog as long as the twenty-two lasted, and when these gave out he should eat them. Such determination won the day, and they proceeded to the settlement of North Belun. Here a native brought him one of De Long’s records, left on the march. From these natives he learned in which direction the records had been found, and pressing on, in spite of his frozen feet, which were in such a condition he could no longer wear his moccasins, he reached, November 13, the hut where De Long’s first record had been left, a distance from North Belun of thirty-three miles. Could De Long’s chart but have shown the native settlement of North Belun, the whole party would doubtless have been saved. On November 14 following the northeast bank of the river he came to the shores of the Arctic Ocean and found the flag-staff where articles from the first cutter had been cached. Loading his sled with all the articles found there, including logbook, chronometer and navigation box, he returned to North Belun. With fresh dog teams he set out again November 17, in an endeavour to find the hut where Erickson died. Fierce storms and lack of food forced Melville to take refuge in a snow-hole dug about six feet square and three or four feet deep. “The storm continued to blow,” writes Melville, “the whole of that night, the next day and the next night. It was impossible to move until the next day morning, when it cleared up a little, but in the mean time, we had nothing to eat. It was too stormy to make a fire to make tea, and the venison bones which the natives had dug out were full of maggots. We chopped this up in little cubes and swallowed it whole, which made me so sick after it warmed up in my stomach that I vomited it all out again.” Melville reached Ku Mark Surka November 24, and at Belun three days later, after an absence of twenty-three days, in which he travelled no less than six hundred and sixty-three miles over the tundra of Northern Siberia in the face of an Arctic winter. Upon reaching Yakutsk December 30, 1881, where Danenhower and his party had preceded him, Melville retained Nindemann and Bartlett to assist him in the spring search, and instructed Danenhower to proceed with the other nine men to Irkutsk, distant over nineteen hundred miles, from thence to America. The spring search was made under the following instructions from the Navy Department at Washington:—