Velocity of Electric Waves in Air. 131 



now changed. The preceding table shows some of the 

 changes. 



Similar changes took place at other points. 



Next the capacity of the coherer circuit was altered, and as 



consequence, at those points where the original readings 

 were taken the readings now were different. Finally, the 

 capacity (i. e. the tinfoil plates) of the coherer circuit was 

 left out entirely, and so also was the self-induction, for at this 

 stage of the experiment the graphite broke. The readings 

 were now of all sorts, the needle moving very capriciously 

 whenever the sparking began or continued at the oscillator. 

 No regular increase and decrease in the readings could be 

 determined as the coherer was moved farther and farther 

 from the reflector. At one time the needle would be deflected 

 four or five whole divisions of the milleamperemeter scale, 

 while in the next second, at the same place, there would be 

 only a displacement of the needle of "2 or - 3 of a division in 

 the forward direction, or else a deflexion in the opposite 

 direction, and even at times the needle trembled as if acted 

 upon by almost equally opposing influences. These irregular 

 motions of the needle made a great contrast with the regular 

 motions when the coherer and oscillator were in tune. This 

 experiment confirms the truth of the damping theory of 

 Poincare and Bjerknes. 



Sarasin and de la Hive's theory of a multiplicity of waves 

 of different amplitudes originating at the same oscillator 

 was substantiated by the erratic readings obtained when the 

 coherer proper alone was employed. In fact the truth of 

 their theory was readily seen when the ordinary metal filings' 

 coherer or Branly tube was used, since the irregularity of the 

 readings was evidently such as would be produced by quickly 

 succeeding waves of different kinds acting upon the tube. 

 The point which might be the position of a node due to one 

 wave would be the antinode, or at least not always the node 

 due to the next following wave. 



I desire here to express my gratitude to Prof. Trowbridge 

 for his valuable suggestions, his assistance, and his very 

 great kindness in placing at my disposal the resources of 

 the Physical Laboratory in order to have this experiment 

 carried to a successful completion. 



Jefterson Physical Laboratory, Harvard University , 

 Cambridge, Mass., April 1899. 



K2 



