202 AxMONGST THE SOUTHERN ICE. 



February lolh, in a latitude nearly corresponding to that of the 

 Shetland Islands and Christiania in Norway, in the northern 

 hemisphere. 



The temperature gradually fell as we went southwards, and 

 on February 9th went down for the first time to just below 

 freezing point in a snow squall. 



At first, all the icebergs seen were numbered each day, and 

 their positions noted down ; but when we came to have 40 in 

 sight at once this plan was abandoned, and we subsequently 

 had more than a hundred in sight on several occasions. 



The typical form of the Antarctic iceberg, as seen above 

 water, and apparently the form which it always has when first 

 set free on its wanderings, is very simple. The top is a nearly 

 flat expanse of snow, and this is bounded all around by per- 

 pendicular cliff's. The boundary lines of the expanse are no 

 doubt always in the first instance nearly straight lines, since 

 they must be produced by the splitting off of the berg from the 

 parent mass, and the previous splitting of similar bergs from its 

 own outer border when still attached. 



A considerable number of the undecayed bergs seen by us 

 were almost rectangular in outline. Some few were irregularly 

 oval, and the weathered ones of course of all possible irregular 

 outlines. 



Since ice requires about nine times its volume to be im- 

 mersed in order to float it above sea water, the portion of an 



iceberg which shows 



above water is a very 

 small proportionate part 

 of the mass. Mr. Bu- 

 chanan made an accurate 

 estimate of the specific 

 gravity of samples of the 

 berg ice, and calculation 

 of amount of immersion 

 i^.- 1 of icebergs. The pro- 

 portionate depth of a 



DIAGRAM TO SHOW THE PROPORTION OF AN DCrg DelOW WatCr Will OI 

 ICEBERG IMMERSED, AND ABOVE WATER. COUrSC UepenU On thC 



form and on the rela- 

 tive density of the upper and lower strata of the mass. 

 Usually, no doubt, the mass below water is far less than 

 nine times the vertical depth of the height of the part above 

 water, from two considerations. Firstly, the sides of the 

 berg are not perpendicular, but long ledges run out from the 

 base of the cliffs below water, the immersed part being thus 

 much larger in figure than the exposed ; and, secondly, the 



