SCIENTIFIC RESULTS 29 



current data and drift records, except the Framh, so far secured in 

 polar regions have been confined to the shelves inside the continental 

 edoe. The circulation of shelf water, as confirmed by studies otf 

 Newfoundland (Smith. 19'i4a) are controlled by the Avind, by the 

 tidal currents, and by othei- dynamic forces in irregular succession, 

 (xradient currents, on the other hand, consistently hug the steepest 

 part of the continental slopes, and their influence, we have found, 

 ceases with astonishing abruptness as shallow depths are entered. 

 Since the total collection of data to date, except the Fra?n''s, has been 

 made on coastal shelves, they furnish little or no evidence for or 

 against the presence of a gradient current of cyclonic direction in 

 the north polar sea. On the other hand, we have every reason to 

 believe, that Avith the large amount of land drainage from Siberia, 

 such a current is developed there. The fact that the Frani remained 

 on the edge of the continental slope— the seat of gradient currents — 

 at only two points and briefly, weakens the evidence favoring the 

 nonexistence of a current. It is of interest to note that the two 

 points where the Fram approached nearest the slope, viz, («) into 

 tlie basin north of the New Siberian Islands, and {Ij) north of Franz 

 Josef Land, coincide with the two most tortuous parts of her drift. 

 These irregularities may indicate the efi'ect of the gradient current, 

 which is prol)ably not strong enough to overcome the wind-driven 

 movements of the pack. 



The main escape of polar cap ice toward the Atlantic is through 

 tiie opening between Spitsbergen and Greenland. Ice streams much 

 more attenuated and distinctly secondary in size emerge around the 

 southern side of Spitsbergern ; through the Ellesmere-Greenland 

 Strait to Baffin Bay; and by still more minor passages through the 

 sounds of the American Archipelago. The aggregate volume of 

 these >everal discharges and their relative jiroportions are still 

 matter> of conjecture. (J)bservations on the rate of drift of ships 

 and other objects together with a consideration of the age of the ice, 

 leads to an estimate of four to five years for the average period that 

 a given sample of cap ice remains in the Arctic Ocean. 



Few who have visited the north polar regions can fail to appre- 

 ciate the great magnitude of the forces that are constantly at work, 

 forcing the cap together in some places but rending it apart in 

 others.^" The momentum often attained by the fields, hundreds of 

 s(juare miles in area, driven forward by a gale is great. Small won- 

 der that in meeting the edges are tossed high aloft by the impact, 

 and that along the line of collision veritable embattlements are 

 formed in a confusion of ribs and ridges. The rugged features found 

 in one sector, contrasted with the flat sheets prevailing in another, 

 mirror the effects of wind and current. Whether the polar cap ice 

 suffers its greatest dynamic defoi'nuition in the Avinter when the ice 

 is knitted tiie tightest, or during the autunm when it is loosest, is an 

 open question.^" 



''Makarov (1901) has carried out some very interesting investigations on the state of 

 concentration of the cover and estimates that approximately 10 per cent is continually 

 in the process r)f opening, closing, or freezing. 



'■'' Transehe ( 1!»2H. p. 97) states that tlir hummocked. telescoped condition is best 

 developed in late summer and aiilunin. Open water is most plentiful then, permitting the 

 fields to travel across leads with maximum momentum. It is claimed that during winter 

 and spring the liattering and liuffeting diminishes on account of the clo.se. tightly knit 

 condition of the cover. Sverdru]) (19L'.S. p. l(it;i regards the sub.iect in a different light, 

 pointing out that when the ice cover is open it affords one floe which is being struck an 

 opportunity to give way to the next floe, and so on. The ensuing shocks being propor- 

 tionately reduced. 



