153 



from horizon to horizon, as they drifted away from the zenith, 

 always appeared, on account of perspective, to become bowed 

 with convexity away from the observer. This and other 

 optical effects had at all times to be carefully borne in mind 

 when recording the trend of fragments of arcs at a distance 

 from the zenith. 



It is certain that many faint auroral phenomena escape 

 observation altogether, except at the hands of an experienced 

 observer. Often when the luminous phenomena have so 

 faded as to be invisible to the eye in the normal way, 

 something may still be seen by looking towards a point 60° 

 or 90" away and then staring sideways. By doing so the 

 focus is thrown upon a fresh (more sensitive) portion of the 

 retina . 



Reference in the text is frequently made to topographical 

 features of the locality about Cape Royds, the observing 

 station. The position of these can be noted by reference to 

 plate xix. (facing page 151). 



2. Explanation of Terms employed in the Text. 



The manifestations recorded as aurorae are luminous 

 phenomena of diverse form resembling closely the appearances 

 produced by electric discharge in rarified gases. The accumu- 

 lated labours of contributors to this subject show that the 

 incandescence observed in auroral displays is due in some way 

 to electric disturbances in the higher rarified atmosphere. 



In logging the manifestations a number of types have 

 been recognized and agreed upon by various authors. The 

 majority of the contributions to the literature of the subject, 

 however, have not been sufficiently scientific, and published 

 works are overloaded with a redundancy of terms, often over- 

 lapping and ill-selected. In this place, a discussion of the 

 subject will not be attempted. It will suffice to explain what 

 is meant by the terms actually used in this journal. 



Curtain. — This is the idealized form of aurora phe- 

 nomena, and contains all the elements of auroral manifesta- 

 tions. Curtains are in the form of somewhat convoluted 

 (draped) bands and bars across the sky. Like the article of 

 furnishing from which they take their name, they are narrow 

 in lateral width, but the vertical extension is always con- 

 siderable. A fine illustrative example, though somewhat 

 more than usually convoluted, is that figured by Dr. Wilson 

 as plate 10 in the Aurora Report of the British National 

 Antarctic Expedition of 1902-1904. The elements of a 

 curtain are vertical filaments, or streamers, as they are- 

 referred to in this journal. The streamers are arranged just 



