THE PRESSURE UPON THE POLES OF THE ELECTRIC ARC. 
127 
The observation that coiivexion currents from the poles occasion a reduction of the 
pressure between them is in accord with the experiments of Dewar referred to 
elsewhere. He states that “ the effect of hot poles upon the registration of the 
<u 
a 
to 
• rH 
4 -> 
O 
Amperes 
Fig. 21. 
manometers was to produce a small negative i)ressure when the arc was stopped, due 
to the passage of currents of hot air.” 
llie Elimination of Convexion Effects. 
The curve D in fig. 22 is the same curve as in fig. 21, but drawn to the appro¬ 
priate scale, and shown below the current axis instead of above it. A is the 
curve representing the observed total pressure upon the anode taken from 
fig. 14. The corrected pressure is given by the curve A referred to D as datum 
line, or by the curve B referred to the original axes. It wdl be seen that 
correctiiiP' for convexion results in a rather more linear curve than that obtained 
from direct observation, and that the curve now more nearly passes through the 
origin. Remembering that the effects of convexion currents have been under¬ 
estimated, and therefore that the convexion current curve should have been rather 
lower, we see that it is probable that the corrected curve is linear, and that 
the reaction varies directly with the current; it is unfortunate that information is 
very difficult to obtain in the crucial part of the curve where the current is small; 
there is no evidence that the pressure when corrected for convexion ever becomes 
negative. It is clear therefore that, without much error, we may take the oiigiii of 
the oliserved curve at the point at which the straight portion when produced back¬ 
wards cuts the vertical or pressure axis. This provides us with the simplest means 
for correcting for convexion currents, and all previous diagrams should bo so treated, 
T 2 
