678 Prof. Fleming and Mr. "Richardson on the Effect of an 



with practical radiotelegraphy has been noted by one of 

 us (J. A. Fleming), but its use in purely scientific mea- 

 surements is an advantage, as shown in the following 

 paper. 



If a large induction-coil has its secondary terminals con- 

 nected to a pair of spark-balls, or to the outside pair of a 

 series of balls, so arranged that each gap is not more than a 

 millimetre in width, and if a condenser having a capacity say 

 of 0*005 mfd. is connected across the outer balls, then when 

 the coil is in action intermittent sparks pass at the gap. If 

 these sparks are examined in a revolving mirror or photo- 

 graphed on a moving plate, they present themselves as bright 

 images set at fairly equal distances. If a jet of air under a 

 pressure of 16 or 18 inches of water is thrown on the gap by 

 a glass nozzle, the images are then seen to have a ragged 

 tail or aureole which is blown away from the spark, and this 

 tail is generally reddish in colour and easily distinguished 

 from the bright condenser spark. The tail is the image of 

 the arc-discharge superimposed on the oscillatory spark. To 

 determine the effect of this air-blast the following experiment 

 was tried. A ] 0-inch induction-coil had its secondary circuit 

 connected to brass spark-balls 3 cms. in diameter set with a 

 gap of 1 mm., and the balls were also connected to a rectan- 

 gular circuit of round copper wire, the diameter of the wire 

 being 0*162 cm. and the sides of the rectangle respectively 

 142*1 cms. and 34\17 cms. The ordinary or steady resistance 

 of this rectangle is 0*046 ohm and its high frequency resis- 

 tance to currents of a frequency of the order of 1*25 X 10 6 is 

 0*31 ohm. The inductance of this circuit (calculated) is 

 5012 cms. In series with this circuit was placed a condenser 

 consisting of metal plates immersed in paraffin oil, the 

 capacity of which was 0*002645 of a microfarad. 



In contiguity to the long side of the above rectangle was 

 placed the bar of a Fleming Cymometer, two such instru- 

 ments being used in the experiments, called respectively 

 No. 2 and No. 3. The cj-mometer circuit can have two short 

 fine wires of constantan each about 5 cms. long, inserted in 

 it at pleasure, against one of which a bismuth-iron ther mo- 

 junction is attached. By passing measured small continuous 

 currents through this fine wire and connecting the ends of 

 the thermo-junction to a low resistance single-pivot Paul 

 galvanometer, the arrangement can be calibrated as a hot- 

 wire ammeter to indicate directly the mean-square value of 

 the oscillations, which when passed through the fine wire 

 cause a certain deflexion of the galvanometer attached to the 

 ends of the thermo-junction, The other fine wire can be 



