Leaving Etah soon after midnight of August 16th, the Roosevelt swung out from the harbour of Etah and severed all communication with the civilised world. Below decks the ship was filled with coal until her plank sheer was nearly to the water; on deck were more than two hundred Eskimo dogs; and on the topgallant forecastle, and the tops of both forward and after deck-houses were over half a hundred Eskimos, men, women and children, and their belongings.

The heavy pack ice surging down Smith Sound, past Littleton Island, gave me an opportunity to see what good work the ship could do and as we bored through it toward Cape Sabine, she realised my expectations in regard to her, even though very deeply loaded and her boiler power reduced to one-half. The sharply raking stem was a revelation even to me, though it was my idea. Deep and heavy as the ship was, she rose on the opposing ice without pronounced shock, no matter how viciously she was driven at it, and either split it with the impact, or wedged it aside by sheer weight.

Bartlett obeyed my first orders, to give her full speed and I would be responsible, with some misgivings. The sealing captains are always very cautious with their ships when first going out heavy with coal.

At the end of an hour or two he was enthusiastic, both at the ease with which the most crushing blows could be delivered, and the whaleboat-like facility with which the ship wheeled and twisted through the tortuous passages.

But there were some areas of ancient ice which a thousand Roosevelts merged in one could not have negotiated, and we were soon deflected to the southwest, and only when within some ten miles of Cape Isabella did we find it practicable to work northward again.

Cape Sabine and Payer Harbour, which had been my headquarters for sixteen months in 1901–1902, were densely packed, permitting no near approach, and we bored away to the northeast, till the ice became impracticable for further advance, then retraced our route, and worked towards Bache Peninsula, getting about half-way across Buchanan Bay when we were stopped by large floes barring our passage to open water under Cape Albert. The ice later appearing more favourable to the eastward, we retraced a portion of our route and I very carefully reconnoitred Sabine and Payer Harbour again as I was loath to give up my sub-base there, this being part of my programme as outlined to the Club. But the conditions were entirely impossible, and making a detour to the east, the Roosevelt gained the open water at Bache Peninsula, and steaming to the bight south of Victoria Head, the northwestern headland of the peninsula, landed a depot of boats, coal, and provisions.

The value of this locality for the southern sub-base of an expedition going north by the Smith Sound or “American” route was immediately apparent to me in 1898, and in any future work it should be given preference over Payer Harbour. Its advantages are contiguity to a valuable game region, accessibility during any month in the year, and its less changeable and boisterous climate.

The work of landing the depot occupied about ten hours of the 18th, and while the work was in progress I went away with three Eskimos to a neighbouring valley which I knew, and secured three musk-oxen, a large bull, a cow, and a yearling, the latter being brought aboard alive. This animal was of the greatest interest to the crew and the “tenderfoot” members of the expedition, and the arrival of nearly eight hundred pounds of fine fresh beef created a very agreeable impression on everyone.

Up to this time the rush of getting on board my Eskimos and dogs, restowing the ship and fighting the ice, had left me no time for a thought beyond the demands of each hour. Now as I trod the moss patches beside the murmuring stream whose quieter reaches were crusted with ice, saw the fresh tracks of big game and a little later the shaggy black bulks of the musk-oxen with heads lowered and hoofs stamping, in the way I knew so well, my pulses bounded rapidly and I felt that I had come into my own again.

From Bache Peninsula we steamed for Hayes Point through scattered ice, with the heavy pack close on the starboard hand. Conditions were different from those of 1898, when the Windward was five days crossing the mouth of Princess Marie Bay. The night was fine and I could make out every well-known rock along the Cape D’Urville shore where the Windward wintered in ’98–’99. Looking into the distant depths of Princess Marie Bay, numerous episodes with bear and seals and musk-oxen crowded upon me. We experienced some trouble with ice near Hayes Point and Cape Frasier, and finally dodged into Maury Bay and anchored at noon of the 19th, to escape the large fields of very heavy ice which were moving rapidly southward before a fresh northerly wind, crashing with savage fury against the iron bastion of Cape John Sparrow under which we lay.

Vigilantly watching the ice and taking advantage of every opportunity, we squeezed and hammered our way into Scoresby Bay, hugging the shore closely, and thence to Richardson Bay. Twice we nearly reached Cape Joseph Goode only to be forced back by the oncoming floes to a shelter under Cape Wilkes, close to my “Christmas” igloos of 1898, where on that unfortunate midwinter journey to Fort Conger, during which I froze both my feet, I had spent Christmas and opened a small box from loved ones at home.

Rawlings Bay was packed and the ice along the Grinnell Land shore apparently unbroken. On the Greenland side it appeared less dense. During this time the weather was fine.

The aspect of the ice was so extremely unfavourable, northward on the Grinnell Land side, that I determined to test my belief gained in my last four years of work in this region, that the Greenland side of Kennedy and Robeson Channels offered as a rule more favourable opportunities for navigation than the Grinnell Land side.

Firm in my confidence in the capabilities of the Roosevelt and against all the so-called canons of Arctic navigation in this region, she was headed eastward in the afternoon of the 21st and driven into the thick of the channel pack. The ice encountered was very large and heavy, and its southward drift inevitably swept us down; still we made fair progress eastward and after a severe and protracted struggle, during which Bartlett and the mate remained continuously in the fore rigging and I in the main rigging, we broke out into loose ice off Cape Calhoun and began boring northward towards Crozier and Franklin Islands. The channel between Franklin Island and Cape Constitution was attempted and found impracticable. The main channel pack was then negotiated close under the vertical western cliffs of Franklin Island. We then had fairly good going, interrupted by barriers of heavy but rather loose ice to Joe Island.

Stopped here by an impervious jam, the Roosevelt was made fast to the ice-foot which along the southern end of the island is of most stupendous character, and accompanied by Captain Bartlett I climbed to the summit of the island, from whence we saw the eastern portion of Hall Basin clear of ice to Cape Lupton and apparently to Cape Sumner. The western portion of the basin and as far as we could see north and south along the Grinnell Land coast was densely packed with heavy ice.

With the turn of the tide the tension in the channel pack along the western shore of the island relaxed somewhat, and hurrying back to the Roosevelt, a few hours of severe work forced the barrier, and in the teeth of a strong and bitterly cold north wind which kicked up a very respectable sea and sent the spray flying over our bows, we steamed to Cape Lupton, reaching it at midnight of the 22d. While steaming through this open water we passed Thank God Harbour, the winter quarters of Hall’s Polaris on our right and Discovery Harbour, the winter quarters of the Discovery, and site of Fort Conger, on our left.

A few miles north of Cape Lupton, while smashing through a narrow tongue of ice, a sudden swirl of the current which at times runs like a mill-race in this deep channel, swept the ice together in a way that I can only liken to the sudden scurry of fallen leaves before an autumn breeze, pinched the Roosevelt between the big cakes, and smashing her against the ice-foot, ground her along its vertical face with a motion and noise like that of a railway car which has left the rails and is bumping along over the ties. Very fortunately for us she scraped into a shallow niche in the ice-wall, and was hastily secured with every available line.

The entire flurry lasted less than five minutes, but in that time the steering gear was almost disabled. The back of the rudder was twisted on the stock, the heavy iron head-bands and fittings broken, and the steel tiller rods snapped. Temporary repairs were effected, and as soon as the pressure relaxed we steamed on around Cape Sumner and tied up to the fast ice in Newman Bay under Cape Brevoort, to repair our damaged steering gear and await the opening of a lead across Robeson Channel to Cape Union or vicinity.

The winter’s ice was still intact in the bay, its surface level and granular, and the pools of water upon it covered with ice strong enough to support a man’s weight.

As soon as the lines were secured I walked ashore and climbed to the summit of Cape Brevoort. The crest of the northward-facing cliffs commanded the entire northern approaches to Robeson Channel, from Repulse Harbour across to Cape Rawson and southward along the Grinnell Land coast to Lady Franklin Bay. The Greenland coast south of Sumner was hidden by the cliffs of that cape. The ice all along the Grinnell Land coast, and in the centre of the channel, and to the northward as far as I could see, was densely packed. The water in which we had come north still remained open and the Roosevelt could have worked her way close to the shore along the Greenland coast to Repulse Harbour and possibly to Cape Bryant, had my objective point been in that direction. No indication of lead or crack across the channel was to be seen. While on the summit a school of narwhal came sporting down the shore close to the ice-foot below us.

In Newman Bay we remained five days mending the rudder, replacing the tiller chains with wire cables, and crossing to the south side of the bay, where I took on board the Polaris boat left here by Chester and Tyson of Hall’s party in 1871, at Boat Camp. Then as the northern ice filled into the bay, we were gradually crowded out of our shelter behind Boat Camp delta, and tide by tide forced out to Cape Sumner, sometimes grazing the shore as we dodged a floe. During this time the Bay filled completely with ice and the entire northern part of the channel was packed solid. Captain Bartlett and Marvin made several trips to the top of Cape Sumner to reconnoitre the channel but without satisfactory results.

The turn of the tide the morning of the 28th set us out again, and, impatient of the delay, and encouraged by the behaviour of the Roosevelt in crossing the channel at Cape Calhoun, fires were cleaned, machinery thoroughly inspected, and at 4:30 A. M. the Roosevelt was driven out for another contest with the channel pack in which at the time no pool or lane of water was visible.

Just off the point of Sumner a brief nip between two big blue floes which the swift current was swinging past the Cape, set the Roosevelt vibrating like a violin string for a minute or so before she rose to the pressure.

From this we pushed out and began the attempt to cross to the west side, through ice almost continuously up to our plank sheer and frequently of such height that the boats swinging from the deck house davits had to be swung inboard to clear the pinnacles. The delay and inaction of the past five days had become unendurable.

The Roosevelt fought like a gladiator, turning, twisting, straining with all her force, smashing her full weight against the heavy floes whenever we could get room for a rush, and rearing upon them like a steeplechaser taking a fence. Ah, the thrill and tension of it, the lust of battle, which crowded days of ordinary life into one.

The forward rush, the gathering speed and momentum, the crash, the upward heave, the grating snarl of the ice as the steel-shod stem split it as a mason’s hammer splits granite, or trod it under, or sent it right and left in whirling fragments, followed by the violent roll, the backward rebound, and then the gathering for another rush, were glorious.

At other times, the blue face of a big floe as high as the plank sheer grinding against either side, and the ship inching her way through, her frames creaking with the pressure, the big engines down aft running like sewing-machines, and the twelve-inch steel shaft whirling the wide-bladed propeller, till its impulse was no more to be denied than the force of gravity.

At such times everyone on deck hung with breathless interest on our movement, and as Bartlett and I clung in the rigging I heard him whisper through teeth clinched from the purely physical tension of the throbbing ship under us: “Give it to ’em, Teddy, give it to ’em!”

More than once did a fireman come panting on deck for a breath of air, look over the side, mutter to himself, “By G— she’s got to go through!” then drop into the stoke-hole, with the result a moment later of an extra belch of black smoke from the stack, and an added turn or two to the propeller.

At midnight all that could be said was that we were nearer the west side than the east, and steadily drifting southward with the pack. I quote from my journal: “Slow and heart-breaking work. The Roosevelt is a splendid ice-fighter and if she had her full boiler power she would be irresistible. The ice is very heavy, in large floes, some of them several miles in diameter and their edges sheer walls of blue adamant. I shall be glad when we are through.” In one of her charges the Roosevelt left a considerable piece of the stem just under the figurehead as a souvenir upon the top of a berg-piece which she was obliged to butt out of her path. In another, a blue floe twelve to fifteen feet in thickness was split fairly in two.

Until 4 A. M. of the 29th we continued slowly to near the Grinnell Land side. Then but little progress was made for several hours, then another start which was kept up with occasional interruptions until 4 P. M., when after thirty-five and one-half hours of incessant strain and struggle, we drove out into a small pool of water under the northern point of Wrangel Bay. The battle had been won by sheer brute insistence and I do not believe there is another ship afloat that would have survived the ordeal.

Bartlett and I went to our rooms, worn with the long tension, and I fell asleep instantly.

It was the second birthday of a man-child in the distant home, and in my dreams I saw the round face with its blue eyes and crown of yellow hair, smiling at me from the savage mass of black clouds which shrouded the summit of the cape under which we lay. God bless you, little man.

Soon after getting into Wrangel Bay, and while I was asleep, a piece of heavy ice whirling under the stern twisted the back nearly off the rudder, and the entire night was occupied in temporarily repairing the damage. A hunting party of Eskimos sent out during the night returned the forenoon of the 30th with eleven hares and six musk-oxen. Late in the afternoon an unsuccessful attempt was made to reach Lincoln Bay, from which we were driven back by heavy floes moving southward, and at midnight we were again in Wrangel Bay dodging about to keep clear of the shifting ice, while past the capes the big floes were going south with almost race-horse velocity, the channel and summits of all the land dark with fog. The 31st was spent in the bay keeping clear of the floes which swung into and around it, the night thick with falling snow. Made an early start at 3:30 A. M. September 1st, in fog and a blinding snowstorm, and steamed to the north side of Lincoln Bay, where the unbroken pack again barred our passage and we moored to the exposed face of the ice-foot.

Again I quote from my journal: “A wild morning with snow driving in horizontal sheets across the deck, the water like ink, the ice ghastly white, and the land invisible except close to us as we almost scrape against it on the port side. Summer is at an end and winter has commenced.” Scarcely had we made fast to the ice-foot when the ice filled the bay completely.

With the ebb-tide at night much of the ice inside of us passed out, grazing against our side, but no lead formed at the Cape and no opportunity occurred to get north.

With the turn of the flood, the ice came in again with a rush, and the corner of a large floe caught the stern, bent the back of the rudder over to the other side and forced the ship bodily ashore. Here she hung until high-water, with a heavy berg-piece pressing against her stern and threatening momentarily to press her up the shore beyond possibility of floating again.

Almost unmanageable with her twisted rudder, it was a slow and difficult job to work her through the running ice farther up the bay to a supposedly less-exposed berth, snowing and blowing all this time. Here the back of the rudder was straightened somewhat.

Early in the morning of the 3d, a moving floe forced the Roosevelt ashore again, where she hung until the next high-water, and she was hardly pulled off when another floe jammed her hard and fast aground again. I was very anxious to get out of this dangerous and trying position, where the rapid and vicious movements of the ice were a constant menace, but a reconnoissance from an elevation near the Roosevelt indicated that the channel north of us was simply solid with ice.

Shortly after midnight of the 3d, the Roosevelt floated again and, a southerly breeze forming a little water at the mouth of the bay, we steamed out at 3:30 A. M. and succeeded in getting under the delta of Shelter River just south of Cape Union, and in butting into a natural dock among some stranded berg-pieces. Here the ship had one foot of water under her keel and as we moored her, the slack ice through which we had come jammed tight with floes packing against the barrier at Cape Union. Here we enjoyed a fine day with the temperature in the low twenties and experienced a few hours of peace. The river delta to the north and stranded berg-pieces to the south protected us from the attacks of the heavy floes passing rapidly a few yards outside of us.

Eskimos sent out for hare here obtained thirty-six. We were now only some fifteen miles from the Alert’s winter quarters, and a clear run of two or three hours would enable us to beat the record for ships in this region, and save the game for us.

The 5th of September was a memorable day, one that practically ended my fears and anxieties. At 3:30 A. M. we got under way after about an hour’s backing and butting to get out of our niche. A narrow strip of water close inshore showed as far as Cape Union where a narrow but apparently dense barrier pressed against the Cape. Would it let us through? As we neared the barrier it was evidently only about a mile wide, with water beyond it extending to Cape Rawson. I kept both watches of firemen on, and routed out the chief engineer ahead of his watch, because it was evident that we must get through now. In a few minutes the Roosevelt was in the thick of it, throbbing like a motor, the black smoke pouring from her stack, and successfully forced her way through. Cape Union was passed at 4:30 A. M. South of Rawson the ice ran close in against the shore but was looser outside, and we made a wide detour to the northeast, the captain, the mate, and myself in the rigging. The ice was in large, heavy floes and in rapid motion swinging into the mouth of the channel on the flood-tide. The anxious moments were numerous both as to whether we should get through, and also as to whether we should escape a serious nip.

Soon we opened up the Alert’s cairn at Floeberg Beach and could see a narrow canal of water extending close to the ice-foot, and at 7 A. M. the Roosevelt, racing with the incoming pack, was driven through a narrow stream of ice and fairly hurled into a niche in the face of the ice-foot under the extremity of Cape Sheridan and made fast. The ice was packed heavily against the point of the cape and grinding past it. Before our lines were made fast the ice had closed in upon us and the open water behind us was rapidly disappearing.

We were now about two miles beyond the Alert’s position, moored to the exposed face of the ice-foot, with the nose of the Roosevelt pointing almost true north. I felt now that the risks, the chances of the voyage were past. The ship might be lost by being forced ashore, for our position was an extremely exposed one, but we were not likely to lose provisions and equipment, and with these the remainder of the programme could be carried out, and even should she get no farther she would have done her duty and achieved the purpose of her being.

With my feelings of relief, was a glow of satisfaction that by a hard-fought struggle we had successfully negotiated the narrow, ice-encumbered waters which form the American gateway and route to the Pole; had distanced our predecessors; and had substantiated my prophecy to the club, that with a suitable ship, the attainment of a base on the north shore of Grant Land was feasible almost every year.

Previous to the Roosevelt, only two other ships, the Polaris and the Albert, had completely navigated these channels; and two others, the Discovery and the Proteus, had penetrated them as far as Discovery Harbour.

Our freedom of movement and ability to leave the shelter of the land and cross and recross the channel at will through the heaviest ice, was also gratifying to me.

In the voyage from New York to Etah we had passed the latitudes of the most northern extremities of North America, Europe and Asia.

Since leaving Etah, we had passed the latitudes of the most northern extremities of Spitzbergen and Franz Joseph Land, and now only the northern points of the two most northern lands in the world, Cape Morris Jesup and Cape Columbia, lay a little beyond us. The northern-reaching fingers of all the rest of the great world lay far behind us below the ice-bound southern horizon. We were deep in that gaunt frozen border land which lies between God’s countries and inter-stellar space.