56 
H. T. FERRAR. 
observations in this connection, and have proved a great amount of variation. Our 
observations show that the variation is even greater than they have recorded. We 
may mention here that the average amount of salt in dry sea-ice is about 4 • 3 grams 
a litre, whereas 32 *8 grams a litre is the average salinity of the sea. 
Ice met with at sea is more variable, and the amount of salt contained 
in it depends upon the previous history of the ice. 
Near the shore, where floe-ice has buckled below the level of the sea-surface, 
shallow ponds form, and a gradual concentration of the dissolved salts takes place. 
Solutions containing as much as 266*6 grams of salt a litre have been found in 
such pools. 
When the open sea first freezes, part of the concentrated solution left yields 
well-crystallized rosettes (ice-flowers) on the upper surface of the ice. The rosettes 
are usually two to three 
inches across and about an 
inch high and are scattered 
thickly over the surface ; 
they are always extremely 
saline. If the sea-ice be- 
comes depressed by a great 
accumulation of snow, the 
original upper surface is 
still always capable of 
identification through this 
local first salt-concentrate. 
The Snow on the sea- 
ice in McMurdo Sound does 
not often accumulate to a 
thickness greater than two 
feet. Where any object 
interrupts the uniform level of the ice, snow-drifts form. The amount of the drift 
depends upon the magnitude of the object ; such an object as our vessel produced 
a drift some 30 feet thick. The accumulation of local drifts has little ultimate effect 
upon the depth of the lower surface of the ice, for the snow, as it accumulates, pushes 
down the underlying layers into a region where the temperature of the water 
approaches that corresponding to greatest density ; the lower ice then melts and is 
transported by the currents of the water (Fig. 49, p. 85). The equivalent of more 
than 18 feet of snow has been observed to be removed by this means during two 
consecutive years. From this observation it would seem impossible that a very thick 
sheet of ice could be entirely built up through continued deposition of snow on 
sea-ice or oceanic ice. 
At a hole near the ship the upper surface of the snow was three feet above 
Fig. 30. — The Pack-ice, seen from the Crow’s Nest of the Ship. 
