6. The choice of the most appropriate season is another important consideration in ice-navigation; for this period does not fall at the same time in all seas, and the disregard of season was a common cause of the failures of the expeditions of earlier centuries. Since the frozen sea remains unbroken and almost unaffected by the action of the sun even in June, and at that time extends far to the south, it is evident that all attempts to force a passage in that month are labour thrown away. The ice-barrier retreating northward, or the transformation of pack into drift-ice, leaves free navigable water four or five weeks later. The month of August is the best time for ice-navigation in Baffin’s Bay; the end of July or beginning of August on the East Greenland coasts; the second half of August and the beginning of September in the Spitzbergen waters; and in the region of the Parry Islands the favourable opportunity ends about the beginning of September. In general, it seems that the time most propitious for all the coast-water routes, begins some weeks earlier than the corresponding period in the frozen seas remote from land. But since, even in the first weeks of September, the most promising conditions are often succeeded by a sudden reaction due to storms, to cold setting in rapidly, or to excessive falls of snow, navigation in the land-remote frozen seas, in itself so extremely hazardous, becomes specially critical, just when the ice-sheet at its minimum appears to promise the greatest results. 7. The help of steam power is an indispensable requisite, as by it a vessel is able to defy the capricious changes of the wind. The movements of a ship amid the ice are made in interminable curves, and the power to describe an arc with the least radius enables a vessel to follow up narrow and often blocked water-ways. As it is incessantly exposed to severe shocks from the ice, a paddle-wheel steamer is useless; and even in screw-steamers care must be taken to protect the propeller by a special construction. 8. The rate of speed of a vessel in the ice must necessarily be moderate. From three to six miles an hour are sufficient: and a rate of eight or ten miles would soon render her not seaworthy. But even with this reduced rate, her whole frame-work is shaken and loosened at last by the incessant shocks she sustains; and this condition of the ship becomes apparent when concussion with the ice is followed not by a noise as of thunder, but by a low, dull, groaning sound. The larger a vessel, the less her capacity to withstand these shocks, and the sooner will these signs of her diminished strength betray themselves.