106 FRIDTJOF NANSEN. M.-N. Kl. 
freezing process was given to the land and atmosphere) but these small 
quantities of ice are of no importance compared with the enormous . 
Polar ice-masses formed in the North Polar Basin itself, and the cooling 
thus produced, is naturally negligeable compared with that caused by 
radiation from the surface of the Polar and Arctic Seas during the long 
winter. 
On an earlier occasion! it was pointed out that it is the buoyancy 
of the layer, 200 metres thick, of diluted light water which, in 
connection with the wind, is of essential importance in the forma- 
tion of the East Greenland and East Iceland Polar Current and 
not the thin surface layer of diluted water formed by the ice- 
melting. It is therefore not necessary to repeat these arguments 
here; there remains only to discuss Pettersson’s theory in its new form. 
It may first be pointed out that he still overlooks the fact that the 
Polar ice floats in a layer of diluted Polar water of low salinity, 100 
or 200 metres thick; and in the northern seas which he mentions, there 
is, according to the writer's knowledge, not a single place where this ice 
comes into direct contact with Atlantic water. Distinction has, however, 
to be made between Polar ice and the thinner ice, of much wider distri- 
bution, which is formed during the winter in the northern parts of the 
Norwegian Sea itself (cf. above p. 80). It is the latter kind of ice 
which has the wide extension eastwards mentioned by Pettersson, . be- 
tween Jan Mayen and Spitsbergen, and between Jan Mayen and Iceland, 
in the spring and early summer; and by the melting of this ice the 
same quantity of heat is of course consumed as was disengaged on its 
formation. 
It was above (p. 78) pointed out that no appreciable quantity of 
heat can be conducted to the underside of the polar ice from the under- 
lying warmer water?, through the cold polar water with a temperature- 
minimum at 50 or 60 metres, 
Pettersson also appears to have forgotten that, the Atlantic Current 
in its way from the Shetlands to Spitsbergen is cooled about 10° C. 
near the surface and at least 4 or 5° C. in the deeper layers without 
any contact with ice (and the amplitude between summer and winter- 
1 Oceanography of N. P. Basin. See also Nyt Mag. for Naturvid. vol. 39, 1901, p. 157. 
2 If there be such an effective conduction of heat through these water strata, as Petters- 
son is prone to believe, the temperature-minimum at 50 metres must have a cooling 
effect upon the overlving water-strata; and if the water at 50 metres be cooled dowm 
to freezing point by contact with the ice it must naturally cause a formation of ice in 
the overlying water strata with a lower salinity. A conduction of heat to the ice- 
floes from below, through this temperature-minimum is of course excluded. 
