958 Mr. Beer and Prof. Tyndall on Manometric 



observed by connecting it to one limb or! a Chattock Fry 

 tilting bubble gauge of the usual type. In general, the other 

 limb was open to air — that is to say, to some point in or 

 near the body of the arc at which the pressure was that of 

 the surrounding atmosphere. The particular instrument used 

 responded to a difference of pressure of about 1/1000 mm. 

 of water. 



An image of the arc was focussed on a screen so that the 

 distance between the carbons could be maintained at 

 required values and simultaneous readings taken of current 

 and manometer pressure for various lengths of arc. At 

 intervals the carbons were adjusted by filing to a standard 

 shape. The Dewar effect was readily obtained — namely, an 

 excess pressure at the anode for a silent arc and a com- 

 paratively negligible pressure which was sometimes negative 

 at the cathode ; also there was a greatly reduced value of 

 anode pressure when the arc was hissing. 



But the effects were variable and difficult to correlate with 

 one another, even when all readings for which the luminous 

 patch on the electrode was not centred round the hole, were 

 discarded. An observation that the anode pressure seemed 

 to decrease with time suggested either that the Dewar effect 

 was not an electrical wind-pressure phenomenon at all or 

 that there was some superposed disturbing effect. 



It is well known that carbon contains occluded gases. 

 If the end of the carbon is raised to incandescence these 

 gases will be gradually evolved, and if any of this evolution 

 occurs from the walls of the hole leading to the manometer 

 an excess pressure will be developed in the hole and at the 

 gauge, the value of this pressure being dependent upon 

 the rate of evolution of gas and the amount of constriction 

 between the heated region of the hole and the outside air. 



This view was easily put to test by using carbons bored 

 as before, except that the last centimetre of carbon at the 

 arc end was pierced with a hole 1*0 mm. diameter instead 

 of 3*5 mm. If the effect was due to the evolution of gas, 

 a much greater resultant pressure at the gauge was to be 

 expected, because the constriction between the inner limit 

 of the heated region and the outside air is approximately 

 inversely proportional to the 4th power of the radius of the 

 hole and the heated surface area of the inner wall to the first 

 power. 



This was found to be the case, pressures as much as a 

 hundredfold greater being observed in certain circumstances. 

 Since we should expect that in time the end of the carbon 

 would be denuded of its absorbed gas by incandescence, the 



