CONSTRUCTION OF THE CHARTS 289 



three semicircular charts covering different parts of the Antarctic, and overlapping 

 each other by 6o° of longitude. Thus, for instance, Plates LXXXIII, LXXXIV and 

 LXXXV show all records made in January in whatever year, but the day of the month 

 and the year are shown against each record. For example, in Plate LXXI, which shows 

 September observations, a record of the ice-edge is plotted in 20 E and marked 30.38. 

 This means 30 September 1938. 



The latitude of the ice-edge varies to some extent in different years, but there is a 

 much greater variation in different seasons of the same year, and it is for this reason that 

 separate charts are given for each month and not for each year. 



THE OBSERVED ABSENCE OF ICE 



On each chart a line is drawn in red in what appears to be the mean position for the 

 ice-edge for that month. In the interpretation of these charts it is important to re- 

 member that they show only the observed presence of ice and not, except in very few 

 instances, its observed absence. In a really thorough determination of the mean position 

 of the ice-edge full account should be taken of the occasions on which ships have 

 steamed southwards and reached the land or turned away to the north without seeing 

 any pack. Such negative observations would, however, be very difficult to trace. The 

 charts do indicate one or two points (Plates LXXXIV, LXXXVI and LXXXIX) where 

 the continental coast or barrier was found to be clear of ice, but there have, for example, 

 been occasions when factory ships approaching the South Shetland Islands in October 

 and November have reached the Bransfield Strait without encountering ice, although 

 Plates LXXIV and LXXVII show that on other occasions the ice-edge in those months 

 lay a considerable distance outside the islands (see p. 292). However, the scarcity of 

 negative observations is perhaps less important elsewhere in the Antarctic, for whereas 

 the South Shetlands always become ice-free in certain summer months the coasts of 

 other parts of the Antarctic continent are probably seldom clear, and ships from which 

 our records are taken are believed to have generally cruised southward until they found 

 and recorded the ice-edge. 



MONTHLY MEAN POSITION OF THE ICE-EDGE 



The monthly mean position of the ice should, then, be placed a little farther south 

 than the average position of the plotted records of ice observed to be present. The red 

 line drawn on the charts is in any case only tentative. Where it is based on actual ob- 

 servations it is drawn as a continuous line, and where it is interpolated or filled in by 

 analogy with other months it is drawn as a pecked line. In placing these lines we have 

 been guided here and there by Hansen's charts, and the distribution of surface 

 temperature, such as shown by Deacon (1937, p. 29), has occasionally been suggestive. 



In Plates LXIX and LXX the mean position shown on the other charts are assembled, 

 the summer and winter months being shown separately, and these give a general picture 

 of the seasonal advance and retreat of the ice-edge. It will be understood that they repre- 

 sent the normal positions of the ice only so far as the available data indicate them, and 

 are subject to modification in the light of additional observations. 



1-3 



