300 petteusson, ok water and ice. 



strata, which will tend to bend the ice upwards. Still the ice 

 will not generally break, on accoiint of its toughiiess, iintil 

 the temperature has siink below the point of inflection of the 

 curves V or VI. Then the ice will begin to assume the pro- 

 perties of a härd body and, being unable to resist the strain 

 on the upper strata any longer, will break np into innumer- 

 able rifts. The violence of the catastrophe announces itself 

 by sudden cracks, which by some witnesses have been com- 

 pared to the explosions of thunder or the report ofacannon. 

 By externa! pressure of the surrounding pack-ice the dislocated 

 pieces will be piled np against each other, or on each other, to 

 those hummocks or torosses, so eloquently described by arctic 

 travellers, which make sledging excursions över the ice a 

 thing next to impossible and cansed one of the members of 

 the 1860 expedition of Hayes to stigmatize the difficnlty of 

 penetrating farther to the north över the gigantic torosses of 

 the ice of the polar basin thus: »you might as well try tO' 

 cross the city of NewYork över the tops of the houses.» 



I think the resnlt of the present research can be briefly 

 expressed thns: we must henceforth attribute the frequent 

 ruptures of the sea ice and the heaping of its fragments into 

 hummocks or torosses by pressure to an expansion instead of 

 a contradion of its volume. The sole cause of this pheno- 

 menon we have found to lie in the peculiar modilication of 

 the physical properties of the solid water by a slight quantity 

 of dissolved salt. We must of course acknowledge the fact, 

 that sea-ice begins to contract its volume, like any other kind 

 of ice, at a sufficiently low temperature, but it must be observed, 

 that the coefficient of the regtdnr expansion of sea-ice is less 

 than that of pure water and widely inferior to the great 

 changes in volume of the reverse order, which take place 

 nearer to the melting point. 



We may also draw the .conclusion from the numerical 

 data, that every kind of sea-ice is remarkably inferior in vol- 

 ume to fresh-water ice and cousequently sinks deeper into the' 

 water. 



On plate 23 the reader will find a graphic representation 

 of the volumes of the ice-waters from samples V and VI at 

 different temperatures. The volumes of IV are recorded in 

 table IV, B, but the curve belonging to this water could not be 

 traced on plate 23 with distinctness, because it would be hardly 

 discernible from the red line representing the volumes of pure 

 water. The slight saltness of sample IV (only O.ou % of chlor- 



