78 



BJ0RN HELLAND-HANSEN 



[REP. OF THE "MICHAEL SARS" NORTH 



mauer Hansen". It obviously did not e.xist at the former 

 station, as the salinity at the bottom was as high as 34-98 

 "/oo. At the other station it may have existed in a rela- 

 tively thin sheet between 3200 metres and the bottom, 

 but that is not probable when we consider that this station 

 (62) is situated much farther to the east than Stat. 6. In 

 this case, and provided that considerable temporal changes 

 do not take place in the deep water, the bottom-water of 

 northern origin appearing in the eastern part of the North 

 Atlantic does not come from due north, but from regions 

 more to the west. 



During the "Ingolf" E.xpedition in 1895 [Knudsen, 

 1898], 5 stations were worked along a curved line from 

 SE to W of Cape Farewell, at a distance of between 120 

 and 160 nautical miles from land. The following table 

 gives the results of observations near the bottom at these 

 stations: 



At the time of the "Ingolf" Expedition the methods 

 of determining the salinity were not nearly as exact as 

 they are nowadays, and the salinities cited above cannot 

 be expected to satisfy our demands with regard to accuracy 

 in dealing with the bottom-water of the oceans, although, 

 on the whole, Knudsf.n's determinations are remarkably 

 good. The salinity of the bottom-water found at Stat. 

 21 should probably have been about 34-90, as 34-94 was 

 observed at 1695 metres with a temperature in situ of 

 2-9° C. Similarly, the salinities found at Stats. 38 and 

 37 are certainly too low. 



The temperatures from the "Ingolf" were determined 

 by means of reversing thermometers. In general they 

 are probably too low, because they are not corrected for 

 the heating of the severed column of mercury, and the 

 temperatures of the surface waters and the air were much 

 higher than those of the deep water. At any rate, the 

 bottom-temperatures recorded at depths below 3000 metres 

 are much lower than those found in the deepest bottom- 

 water of the North Atlantic. 



The table above shows temperatures (/,„ as well as 

 H) which decrease fairly regularly with the depth. 



WOST [1928] suggests that a ridge, wliicli he calls 

 the Newfoundland Ridge, connects the Newfoundland Bank 

 with the central longitudinal ridge of the North Atlantic, 

 and thus forms the southern limit of a deep called "The 



Labrador Deep". This does not appear from our bathy- 

 metrical chart on p. 63-', but is indicated on Fig. 26. 

 In his sections (/. c. Pis. XXXIIl and XXXIV) Wosi indi- 

 cates that the depth of the Newfoundland Ridge below 

 the surface is about 3000 metres. It is very probable 

 that the deep water below this depth in the sea south 

 of Greenland is shut off from the deep water in lower 

 latitudes. In this case we must look for the chief nor- 

 thern source of the bottom-water of the North Atlantic 

 Ocean at levels above 3000 metres. The observations 

 from the "Ingolf", cited above, seem to show that such 

 water existed between 2500 and 3000 metres in the sea 

 south of Cape Farewell. 



In his above mentioned paper in 1913 N.ansen repro- 

 duces two charts (/. c. Figs. 8 and 9) illustrating the con- 

 ditions at the surface of the sea between the British Isles, 

 Greenland and Newfoundland in April. He defines an 

 area where the conditions seem to be specially favourable 

 for the formation of bottom-water, with temperatures of 

 about 2° C. and salinities of about 34-9 "/oo. The Danish 

 Meteorological Institute has constructed isotherms for the 

 mean monthly temperatures of the surface of the sea 

 north of 50° N., based upon observations from 1876 to 

 1915. There are only a few observations made in winter 

 in the northwesternmost parts of the Atlantic, so the 

 isotherms can only be drawn approximately. There is 

 no doubt, however, that temperatures of about 2° C. 

 appear at the surface over quite long stretches outside 

 the drift-ice along southern Greenland in winter, when 

 no ice-melting takes place and the salinity is, presumably, 

 relatively high in these regions. From what may now 

 be deduced it seems very probable that salinities of about 

 34-9 "/ou here coincide with temperatures varying between, 

 say, r and 3° C. 



We have, unfortunately, no vertical series with obser- 

 vations of temperature and salinity from the areas in 

 question in winter, but we may safely conclude that an 

 e.xtensive vertical convection must take place where the 

 surface waters have attained such high densities as result 

 from temperatures of 1° — 3° C and a salinity of 34-9 %o. 

 The heaviest, /. e. in this case the coldest, water will 

 sink deepest. We must then assume that the coldest 

 water follows the deepest parts of the sea-bed and collects 

 in the Labrador Deep (an excess being carried away by 

 mixing with the water above) while water with a poten- 

 tial temperature of around 2' C. occurs at higher levels 

 and spreads southwards. It inay move across the New- 

 foundland Ridge to the western part of the North Atlantic, 

 or eastwards, south of the Reykjanes Ridge (see the ba- 

 thymetrical chart, p. 63*) to the eastern part. It is very 

 probable that considerable annual variations may occur 

 in the temperature as well as in the quantity of the 



