TEMPERATURE OF THE ROSS SEA 13 



but while we do not claim that these temperatures are absolutely accurate, we are of 

 opinion that the following considerations suggest that there is a slight inversion of 

 temperature in much of the water of Ross Sea. Firstly, there can be no doubt that 

 the sea ice of a preceding winter season is melted from below as summer approaches, 

 so that 7 feet (2*13 metres) of sea ice towards the end of the summer may be less 

 than 5 feet (I'o metres) in thickness. In other words, half a metre may be lost 

 annually to the sea ice near the latitude of Mount Erebus as the result of melting 

 from below.* Secondly it was noticed that there was a tendency for open pools of 

 water to form and keep open even during the winter at any point in Ross Sea where 

 there was any obstacle, such as a submerged ridge or moraine which tended to 

 deflect currents upwards. The pool of sea water which remained unfrozen until late 

 in May to the north of the Cape Barne Glacier, the pool off" the submerged moraine 

 to west of Cape Armytage, and possibly the large strip of open water on the north 

 side of the Drygalski Glacier,t may be due to the upward deflection of Avarmer but 

 more saline water, so that locally it rises above the surface colder but fresher water. 



This explanation of these pools seems more reasonable than the liypotbesis that 

 they may be due to the local warming efl'ect of hot springs situated along lines of 

 faulting. In a region of such modern volcanic activity such springs are almost sure 

 to exist, but their heating effect on the vast body of water in Ross Sea would probably 

 be negligible. 



o 



* Unfortunately we have only approximate data on this subject. Tlie above statement is based on 

 the fact that whereas we found that the sea ice acquired a thickness at Cape Royds, in 1908, of 7 feet 

 (2-13 ni.) during autumn, winter, and spring, in January its thickness had been reduced to about 5 feet 

 (15 m.) with numerous intervening corrosion hollows where the thickne.^s was less than 1 foot {'Sm.). 



t That these open pools are connected with upward deflected currents seems highly probable. What 

 is now needed is a seiies of accurate vertical temperatures at type localities. These will no doubt in part 

 be supplied by the future reports of the recent British Expedition under Captain R. F. Scott. 



