ELECTRICITY OF THE BLOWPIPE FLAME. 395 



Iron and copper acted as zinc, but rather more feebly. 

 As in these experiments the wires of zinc, iron, and copper 

 respectively were stouter than the platinum wire, I attribute 

 the superior amount of deflection when the oxidable metal 

 was in full flame to the greater cooling effect of the thicker 

 wire reducing the antagonising thermic current. 



As the above experiments seemed to show that there was a 

 proper flame-current irrespective of, and even overcoming the 

 thermic flame-current, I was led to expect that by uniting in 

 direction these two currents I might get more marked results. 

 The following experiment, it will be seen, realised this expec- 

 tation. 



I formed a little cone of platinum foil of ^ths of an inch 

 in depth, and the same width at the widest part ; I suspended 

 this in a ring of platinum wire and substituted it for one of 

 the coils. 



Being placed in the full flame, the coil being at the root, 

 it was filled with water, and water dropped into it from a 

 pipette to supply that which was boiled away. I now readily 

 obtained a deflection of 20 in the same direction as in my 

 original experiment, and frequently the needles deviated to 

 30. When the cone filled with water was placed at the root 

 and the coil in the full flame the deviation was only 5. 



In all the above experiments with the blowpipe it will be 

 seen that the direction of the current was, as far as a com- 

 parison can be instituted between the blowpipe flame and ordi- 

 nary flame, the reverse of that indicated in the experiments of 

 Hankel, excepting his experiment with the flame of hydrogen ; 

 and also the reverse of the greater part of the experiments of 

 Buff, which he rightly attributes to thermo-electricity. 



One result of M. Buff ('Archives d'Electricite,' vol. xvii. 

 p. 275), when he places one wire in the centre and the other 

 at the outer margin of the flame, is probably dependent on a 

 similar cause to mine ; though when he places the second wire 

 in the full flame the current is in a contrary direction to that 

 which I obtain. 



My experiments prove, I think, distinctly that there is a 

 voltaic current, and that of no mean intensity, due to flame, 

 and not dependent on thermo-electricity. 



