64 Trans. Acad. Sci. of St. Louis. 



in the Rowland effect. Tliey indicate a swaying of tlie 

 lines of force, due to variations in permeability. They 

 indicate that ionized molecules of air in the earth's mag- 

 netic field tend to set in definite position with respect 

 to the lines of force. This wonld cause local variations 

 in the position of the magnetic needle, and local and 

 rapidly varying changes in intensity of the field. The 

 local electrical condition of the air at the station of 

 observation and at the moment when the mnd-gust 

 occurs is also involved. 



Some \asible evidence of auroral streamers have been 

 obtained around the poles of a steel magnet. Satisfac- 

 tory photographic evidence has not yet been obtained on 

 account of the flickering character of the phenomena, 

 and the inductive effects of the walls of the bell-jar in 

 which the magnet is hung. 



Plans are being made to repeat some of the work 

 described in this paper, at the summer home of the writer, 

 at the north end of Lake Huron, where the surroundings 

 are those of Nature. 



These observations were made in Eads Hall, a two- 

 story building with granite walls, and floors of reinforced 

 concrete. No vibrations of the building due to wind 

 could have had any effect upon the observations. The 

 only disturbances which complicated the results were 

 those due to electric cars, and railway trains. Their 

 effect was observed when that of the wind was inappre- 

 ciable, and no effects comparable with those shown in 

 Plates XIX and XX have ever been observed. 



The tracks of these railways were about 650 feet dis- 

 tant. 



Reversal of the Rowland Effect. 



In a former paper in these Transactions, No. 3. Vol. 

 XXI, this subject was discussed. The results there de- 

 scribed have been obtained by suspending the wire shown 

 in Fig. 1, p. 81, on three or four long silk fibers. The 

 wire was placed in contact with vertical strips of hard 



