INTRODUCTION. 607 



relation to the winds. The clearness of the tables and graphical 

 illustrations contribute not a little to the value of the record of 

 the Expedition to East Greenland under Captain Koldewey, as 

 published by the Committee of the German Arctic Expedition. 



Since all descriptions of phenomena which maybe called appear- 

 ances, such as the forms and character of clouds, and especially the 

 display of the aurora, depend not only on the phenomena, but also 

 on the sense of relief felt by the observer in having something to 

 interest him beyond mere routine, it would be well that the descrip- 

 tion of such appearances should be accompanied by sketches. For 

 clouds reference should be made to the excellent sketches of clouds 

 in the Report of the Proceedings of the Conference on Maritime 

 Meteorology, 1874; sketches of the Aurora similar to these would 

 be of very great service. Good sketches of Aurora may be made 

 in Chinese white and indigo on grey paper. (Such sketches have 

 been made by M. Capron, who suggests that drawings should be 

 made through a grating of wires or threads upon ruled paper.) 



The excellent papers of M. Lemstrom and M. Wijkander, taken 

 in connexion with the discoveries of M. Angstrom and other ob- 

 servers on the Spectrum of the Aurora, seem to show clearly that 

 the Aurora is an Electrical effect, closely resembling the discharge 

 in a vacuum tube ; whilst the important results of M. Wijkander 

 for determining the electrical state of the air point to some very 

 important questions with regard to the conduction of electricity 

 by air at different temperatures. Some light would probably be 

 thrown on these questions by measurements of Earth currents in 

 the Arctic regions. 



It does not appear, from M. Wijkander's results on the Electricity 

 of the Air, that there is any distinct relation between the conducting 

 power of air and its temperature : for we find that the air insulates 

 well at — 15° and — 16° ; whilst at another time at — 17°, or between 

 — 17° and —13°, the charge instantly disappears; we also find 

 that although at lower temperatures air usually insulates well, yet 

 at times the insulation is not good at — 27° or again at —31''. 



When the temperature of the air is or has been increasing, the 

 electrometer may be somewhat colder than the surrounding air ; in 

 such a case moisture will be condensed on the insulating supports 

 and the charge will be lost ; whereas when the temperature of the 

 air is diminishing, and the electrometer warmer than the air, no 

 moisture will be condensed upon the supports and the insulation 

 will be good. 



This may explain M. Wijkander's results, and is in accordance 

 with the results of experiments by Sir W. Thomson, that Air, 

 whether dry or moist, always insulates perfectly at all ordinary 

 temperatures. 



If, as M. Wijkander supposes, Air at a low temperature becomes 

 a good conductor of electricity, then it will be well to employ his 

 method of using the Quadrant Electrometer to detect it ; the 

 quadrants may be kept at a constant difference of potential by 

 employing Mr. Latimer Clark's constant cells. It might also be 

 possible to test the direct eftect of the Aurora on the needle of a 



