1893.] 



The Electrolysis of Steam. 



103 



change takes place in the action of the coil, and an excess of oxygen 

 begins to appear ; this gradually wipes out the accumulation of 

 hydrogen, and if it goes on long enongh makes the residual gas in 

 the tube entirely disappear ; then the oxygen begins to accumulate, 

 only, however, to be wiped out later, when another change in the 

 action of the coil has caused hydrogen to appear in excess in this 

 tube. Thus, in this case, the residual gas in the tube does not, as 

 before, steadily increase with the time of sparking, but is continually 

 waxing and waning,-sometimes being oxygen and sometimes hydrogen. 



Long Sparks. 



When the spark length is increased beyond the critical value the 

 excess of hydrogen, instead of appearing as with shorter sparks at 

 the positive electrode, changes over to the negative; the excess of 

 oxygen at the same time going over from the negative to the positive 

 electrode. Thus the gases when the spark length is greater than 

 its critical value, appear at the same terminals as they do when 

 released from an ordinary electrolyte, instead of at the opposite ter- 

 minals, as they do when the sparks are shorter. 



The length of spark at which this reversal takes place depends, to 

 a very great extent, upon the current sent through the steam : the 

 smaller the current the shorter the critical spark length. By 

 diminishing the current by inserting a liqnid resistance I reduced, on 

 one occasion, the critical spark length from 11 to 8 mm. This critical 

 length, too, seems to depend upon a number of small differences not 

 easily specified ; it will even vary greatly in the course of one after- 

 noon, though apparently nothing has been changed. I have found, 

 however, that this capriciousness disappears either altogether, or to 

 a very great extent, if Ley den jars — very small ones will do — are 

 attached to the terminals of the steam tube, or if an air break is 

 placed in series with that tube. Under these circumstances the 

 critical length will, if the same coil is used, remain constant from 

 day to day. 



It will be noticed that my results, when the length of the spark is 

 greater than the critical length, agree with those obtained by Perrot 

 and Ludeking, as these observers fonnd that the hydrogen appeared 

 at the negative, the oxygen at the positive, electrode. Ludeking 

 worked with long sparks only, so that his results are quite in accord- 

 ance with mine. In Perrot's experiments the spark length was abont 

 6 mm. I have never been able to reduce the critical length quite as 

 low as this, even though I diminished the current to the magnitude 

 of that used by Perrot ; I have, however, got it as low as 8 mm., and 

 it is probable that the critical length may not be governed entirely 

 by the current. 



