210 



For instance, the drift (body-travel) of arcs was dominantly 

 towards the Magnetic Pole. On occasions of powerful display, 

 however, the arcs appeared, in the first instance, to drift 

 towards the South Magnetic Pole; then apparent attraction 

 gave place to repulsion after the culmination of the intensity ; 

 the later phases including a drift awav from the Magnetic 

 Pole. 



9. Relation to Land Topography. 



Many entries in the journal appear to illustrate a rela- 

 tion between auroral phenomena and topography. Arcs and 

 bands stretching in straight lines over the sea towards Ross 

 Island generally showed deflections where they passed ver- 

 tically above the land. Very frequently curtains hung over 

 Ross Island, showing a relation in form to the topography of 

 the island itself. Curtains whose direction carried them 

 across the highlands of the island often appeared to rise 

 locally as if, in some measure, following the surface 

 contour of the island. Bands extending as if hung 

 over the backbone of Ross Island were very common, and 

 usually, when the luminosity had died out elsewhere, 

 streamers still persisted above the summit of Mount Erebus 

 (13,300 ft.). Another favourite spot for isolated streamers 

 was a location in the direction of Beaufort Island, apparently 

 over it. 



Iii the clearest possible manner this favouritism for cer- 

 tain geographical situations was illustrated. The curve of 

 "Frequency" (text fig. 3), to which reference has already 

 been made, illustrates the great paucity of aurorae over the 

 continent and the abundance over the sea. 



10. Relation to Meteorological Conditions. 



There is certainly no simple relation between meteor- 

 ological conditions and aurorae. Auroral forms of all kinds 

 and varying intensities are to be noted in calm and in bad 

 weather alike. Such evidence, however, does not dismiss the 

 possibility of a connection, for in the first place, as the 

 phenomena for the most part at least operate at very great 

 heights above the earth's surface, meteorological conditions 

 are no criterion of what exists above. In any case, at great 

 elevations in the atmosphere the meteorological conditions are 

 not likely to be subject to the rapid variations of the lower 

 zones. It is likely that the role of meteorology in auroral 

 manifestations is the effect upon earth temperature and 

 atmospheric potential and ionization, caused by extensive 

 regional conditions even remote from the point of observation. 

 For example, extensive calm periods, storm periods, or cloudy 



