out of bays and inlets during the previous summer, but have not had
time to get north before the winter set in; (3) of comparatively
heavy ice formed over the Ross Sea early in the last winter; and (4)
of comparatively thin ice which has formed over parts of the Ross
Sea in middle or towards the end of the last winter.

'Undoubtedly throughout the winter all ice-sheets move and twist,
tear apart and press up into ridges, and thousands of bergs charge
through these sheets, raising hummocks and lines of pressure and
mixing things up; then of course where such rents are made in the
winter the sea freezes again, forming a newer and thinner sheet.

'With the coming of summer the northern edge of the sheet decays
and the heavy ocean swell penetrates it, gradually breaking it into
smaller and smaller fragments. Then the whole body moves to the north
and the swell of the Ross Sea attacks the southern edge of the pack.

'This makes it clear why at the northern and southern limits the
pieces or ice-floes are comparatively small, whilst in the middle the
floes may be two or three miles across; and why the pack may and does
consist of various natures of ice-floes in extraordinary confusion.

'Further it will be understood why the belt grows narrower and the
floes thinner and smaller as the summer advances.

'We know that where thick pack may be found early in January, open
water and a clear sea may be found in February, and broadly that the
later the date the easier the chance of getting through.

'A ship going through the pack must either break through the floes,