caught in the ice on this side of the pole near the Greenland Sea are carried southward, and that the crews run, as a rule, no great risk. Not so on the other side of the Pole north of Bering Strait; ships caught in the ice there drift northward and often disappear, some with few and others with many men on board; most of them probably are destroyed in high unknown latitudes. These facts must lead the thoughtful observer to the conclusion that there are differences in the sea currents which may be utilised in favour of a polar expedition. Let us, therefore, examine the question more closely.

The most important polar current is, without doubt, that which runs southward along the east coast of Greenland. This has a considerable speed, and carries an immense quantity of water out from the polar basin. It fills the whole opening between Greenland and Spitzbergen, with the exception of a narrow belt along the coast of the latter, and it runs over the deepest known bottom in the Arctic regions; there are ascertained depths of 2600 fathoms. The depth of the actual current itself cannot, however, be so much; I do not think that we are entitled to assume that there is any current of importance deeper than 300 fathoms ; and in order to be within the mark, let us say only 200 fathoms. It might be expected that under this polar current another current was running northward. From what we know of the water, we seem, however, to be fully entitled to say that such cannot be the case. On the contrary, water at a much greater depth probably comes from the unknown north. The breadth of the polar current on the surface is 250 nautical miles, and at the depth mentioned it seems to be about 170 nautical miles. To calculate the average speed of the current is very difficult; it runs probably more rapidly at the surface than in its deeper parts, and, on the other hand, the speed is nowhere constant during the whole year. Sometimes, especially in the summer months, it is very rapid, but at other times it seems to have a much slower course. Taking everything into consideration, I do not think we are entitled to estimate the average speed of the whole current for the year at more than two nautical miles a day. By this calculation we arrive at the conclusion that the polar current between Greenland and Spitzbergen carries southward between 80 and 120 cubic miles of water every twenty-four hours.

Whence is all this water derived? It cannot originate at the pole itself; the place of the water that flows out from the polar basin must be supplied by water running in. It is also evident that the influence of a current so considerable as this cannot be limited to a small area; it must affect the polar basin like an immense pump, sucking the water even from the shores of Siberia and Bering Strait. This is the more certain, as the polar basin is found to be unusually shallow wherever it has been sounded. There are only a few currents known which run into the polar basin. A small branch of the Gulf Stream is known to

run northward along the west coast of Spitzbergen. This current is, however, too insignificant to be of much value in this connection; to some extent it certainly also rounds the north coast of Spitzbergen, and returns southward again towards its eastern coast. The main body of the Norwegian Gulf Stream passes eastward to the north of Norway, and enters the polar basin north of Novaya Zemlya. This current is considerable; our knowledge of it is, however, not sufficient to enable us to form any certain idea about the quantity of water which it carries along; but according to the calculation of Prof. H. Mohn, in his important memoir on the Northern Ocean,* and according to information from the sealers, I think we may assume that it carries at least 60 or 70 cubic miles of water every twenty-four hours into the polar basin. A third current running into the Polar Sea is that which runs northward through Bering Strait. This cannot be of great importance, as the strait is so narrow and shallow, but from the latest descriptions of the current we are perhaps entitled to assume that at least 10 or 14 cubic miles of water are here running northward daily.

The currents certainly furnish the most important supplies of water to the polar current along the east coast of Greenland. Another addition comes from the American, and especially from the Siberian rivers that run into the polar sea. The drainage area of all these rivers is very considerable, embracing nearly the whole of northern Asia, or Siberia, besides the principal part of Alaska and British North America. The rain and snow of this region are not, however, very considerable, and the whole quantity of moisture falling over Siberia I have calculated to be no more than about 626 cubic miles in one year if the Russian meteorological data on Siberia are correct. On account of evaporation we cannot assume that more than a certain part of this water reaches the Polar Sea; perhaps not more than one cubic mile daily during the year. This is not much, compared with the size of the ocean currents, but this addition is of special importance, as it consists of fresh and comparatively warm water which principally runs out into the basin during the summer, and which for a very long time keeps at the surface of the sea on account of its lightness, and thus produces surface currents running northward from the Siberian coast. This is also the reason why there is so much open water along this coast every summer. Το this stream of fresh water the evaporation from the melting of ice in the Polar Sea contributes very little. The moisture of the air over the area draining into the Polar Sea must consequently originate mainly in the Atlantic and Pacific Oceans. This constant addition of fresh water must evidently be the principal reason why the water of the polar current between Greenland and Spitzbergen contains somewhat

* Prof. H. Mohn, The North Ocean; its Depths, Temperature, and Circulation. The Norwegian North Atlantic Expedition, 1876-78.' Christiania 1887.

less salt, even at considerable depths, than the water of the North Atlantic seas.

We thus see that the polar basin is daily receiving a large inflow of water. As little evaporation takes place from its ice-covered surface, there must necessarily be a corresponding outflow, and the most natural outlet is the broad and deep opening between Spitzbergen and Greenland. According to what has already been said the water running out here seems very nearly to correspond in quantity to the inflow mentioned.

Currents also run southward through Smith Sound, Jones Sound, and Lancaster Sound, in the Arctic Archipelago of North America, but as these sounds are very narrow and shallow, the body of water which their currents carry off is of little importance in this respect. The current running southward between Spitzbergen and Franz Josef Land is also insignificant when compared to the east Greenland current. By considering the contributions of water already referred to which this last current probably receives, it may be possible to form some idea of the approximate course of this current through the unknown regions. The waters of the North American rivers form, very likely, a portion of the currents through the Arctic Archipelago of North America; a small part of the current through Bering Strait, perhaps, runs also in this direction. We have left then, for the formation of the east Greenland polar current, the Novaya Zemlya current, the Siberian rivers, a part of the current through Bering Strait, and the moisture falling over the polar basin.

It seems quite natural that these sources should converge, and to some extent unite to form the Greenland current. We must expect, therefore, to find the main body of the current which is formed in this way, lying somewhere to the north of the middle of that extended area from which it receives its converging sources, and this place must consequently be somewhere in the neighbourhood of the New Siberian Islands. Here we also have the mouth of the Lena River, which carries a considerable body of comparatively warm water northward into the Polar Sea. From this region the current must naturally run in a northerly direction by the shortest route to the outlet between Spitzbergen and Greenland, and this must be to the north of Franz Josef Land and near or across the North Pole. But the direction of the current may perhaps, to some extent, be disturbed by the winds. Unfortunately we do not know much of these in the Arctic regions; from the little we know it would appear, however, that the winds should be favourable for such a current, and that their average direction during the year is very nearly the same as that which we have assumed for the latter. This we can also conclude from the observations made during the drift of the Jeannette.

I have tried to convince you that, from what we know about the

ocean currents and the winds along the "threshold of the unknown regions," we are entitled, in fact are obliged, to assume that these regions are traversed by an ocean current. But is there no direct evidence of the existence of such a current? I think there is.

The American Jeannette, under the command of De Long, was on September 6th, 1879, caught in the ice to the east of Herald Island (north of Bering Strait at about lat. 71.30° N., long. 175° W.). This is the only case in which we know exactly the drift of a ship caught in the ice north of Bering Strait. Like all other ships caught there, the Jeannette drifted towards the north and north-west, but her course was at first very irregular, and to a great extent dependent on the varying winds. Her main drift, however, had a decidedly northwesterly direction; this course grew more and more direct the more the ship advanced westward, and during the last half year of her drifting she kept a pretty straight course, advancing north-westward with an average speed of no less than two miles daily, until she was crushed in the ice and sank on June 13, 1881, north of the New Siberian Islands (at about lat. 77° 15' N., long. 155° E.). This, drift of the Jeannette shows that a current must be running north-westward from the sea north of Bering Strait. This current does not seem to have been very strong during the first part of her drift, but as she approached the region to the north of the New Siberian Islands the current seemed to get a stronger and stronger hold of her, and the last week before she sank she often drifted along with a speed of more than eight miles daily. This was just near the region where I have already supposed the Greenland current to originate.

More remarkable, however, in this respect than the drift of the Jeannette itself, is the fact that a number of objects belonging to her or her crew were found on an ice-floe near Julianehaab, on the south-west coast of Greenland, just three years after she had sunk (June 18th, 1884). These objects, fifty-eight in number, were found by some Eskimo, and were afterwards collected by the director of the colony of Julianehaab, Mr. Lytzen, who has described them in a paper in the Geografisk Tidskrift (vol. viii., 1885-86, pp. 49-51), which is issued by the Danish Geographical Society in Copenhagen. Among these objects the following may be mentioned here:-1. A list of provisions with the signature of De Long, the leader of the Jeannette expedition, in his own handwriting; 2. A written list of the boats of the Jeannette; 3. A pair of trousers made of oiled linen, marked "Louis Noros," which is the name of one of the men saved from the Jeannette.

There seems to be some doubt as to the genuineness of these relics, and a well-known American traveller has even maintained that if the articles were found, it would seem more reasonable to trace them to the Proteus, which was wrecked in Smith Sound, about 1000 miles north of Julianehaab (in July, 1883). But how a list of provisions with De

Long's signature, a list of the boats of the Jeannette, a pair of trousers marked Louis Noros, &c., were brought on board the Proteus he does unfortunately not inform us. If the articles had still existed it would have been very easy to identify them, but unfortunately they are now lost. Mr. Lytzen sent them to a friend in Copenhagen, who had them for some time, and sent them to the International Exhibition at Amsterdam. After the close of the latter they were returned to the friend in a wooden case, and were placed in the cockloft in his house. A few years ago this man died, and his wife, not knowing what the articles were, allowed them to be destroyed, and I came just two months too late to save them. But I cannot see why these relics should not be genuine. What could the natives of Greenland know about the Jeannette and her sad fate? Or can it be supposed that Mr. Lytzen, a respectable official of Greenland, should have invented a story like this, and given a minute description of articles never found? Nobody who has read his description can easily believe such a thing, and that there should be any mistake or deceit is certainly much more improbable than the drift of a floe from Siberia to Greenland, which is certainly not at all improbable seeing that a great many other objects known must have drifted, and constantly do drift, the same way. The above-mentioned facts seem to speak for themselves, and need no further testimony. We have consequently to deal with the fact that an ice-floe, with objects from the Jeannette lying on it, was found near Julianehaab. These objects must have been left on the floe either near the place where the Jeannette sank or somewhere on the route of her crew towards the Lena delta. It is quite impossible that this floe could have come to the south-west coast of Greenland through Smith Sound, Jones Sound, or Lancaster Sound, as the currents through these sounds run southward along the west side of Baffin's Bay and Davis Strait, along the east coasts of Baffin Land and Labrador towards Newfoundland. No ice or objects coming that way reach the south-west coast of Greenland, along which a current runs northward, coming from the east coast (of Greenland) and round Cape Farewell, carrying along all the floe-ice and foreign objects which are found on the southern part of the west coast. There can consequently be no doubt that the floe which carried these relics from the Jeannette was borne along by the East Greenland current.

The question therefore arises: by what route did it travel all the way from the New Siberian Islands to the east coast of Greenland? The shortest and most natural route would, of course, be across the region to the north of Franz Josef Land, i.e. across the region round the pole; nay, we are obliged to assume this route as the only possible one. There is no probability that the floe should have been able to travel against the before-mentioned branch of the Gulf Stream, running eastward into the polar basin through the sound between Novaya Zemlya and Franz Josef Land, and that it should thus have passed westward to the south