138 Professors Ayrton and Perry on the Measurement 
well as at all points in the glass tube, is the same as measured 
by the electrometer. 
The next set of experiments, K and L, were made under 
exactly the same conditions as A, B, C, D, E, F, and G, with 
the exception that the terminal platinum plates were now per- 
pendicular to each other, the plate towards which the wire was 
moved being parallel to the long side of the trough. 
Distance 
between the 
platinum 
Time after 
putting on 
Galvanome- 
ter-deflection. 
Electrometer- 
deflection. 
Temperature. 
wires, in 
battery. 
centimetres. 
m s 
f80 
1 
706 
51 
1 
K 60 
K 1 40 
30 
701 
38 
L 13° C. 
50 
699 
27 
[20 
2 
697 
15 
J 
f80 
L 60 
h \ 40 
1 
705 
51 
] 
ii 30 
55 
701 
698 
40 
27 
13° C. 
[20 
2 20 
697 
15 
J 
The resistance of the longer column of the water as mea- 
sured by the electrometer is about the same as before, whereas 
that of the shorter is even greater ; so that the resistance for 
the 80-centimetre column is even still less than four times 
that for the 20-centimetre one. But since the platinum plate 
near the stationary platinum wire was in these last two sets of 
experiments K and L kept parallel to the trough (that is, 
parallel to the mean direction of the lines of flow), it follows 
that any want of parallelism of the lines of flow to the edge 
of the trough at the point where was the stationary wire 
would be exaggerated by this mode of placing the plate ; and 
since we observe that the error in the proportional law for 
distance is also increased, we may conclude that the explana- 
tion given above of the want of perfect accuracy in the pro- 
portional law being due to want of perfect parallelism in the 
lines of flow is the correct one. 
In all the previous experiments the distance between the 
electrometer-wires only was altered ; but in the next set 
the distance between the platinum plates as well as that be- 
tween the platinum wires was altered, the distance between 
each plate and wire being kept constant. Further, the r. -i-r- 
ance determined from the electrometer was calculated, not, as 
before, by comparing the electrometer-deflection when its 
electrodes were attached to the platinum wires with the deflec- 
tion obtained when its ends were attached to a known resist- 
ance traversed by the same current, but by first determining 
the absolute value, in volts, of the electrometer-scale with the 
