64 J. A. EWING AND J. G. MACGREGOR ON THE 
down to the solution of maximum conductivity ; density 1:2891. To shew its 
accuracy we give the following table :— 

Specific Resistance. Specific Resistance. 
foe Ta he a aoe 
Observed. above Kysation, | Observed shore Bauation 
182°9 : : : 204°6 33°7 " : : BPA 
1405 , : ‘ 156:9 32:1 : 2 : 31°8 
iilet : 4 : 110°5 30°3 : ; 5 30°1 
63°8 : : ‘ 64-6 29°2 ; ; 28°9 
50°8 : : : 50°6 28°5 : ‘ A 28°5 
42:1 : é ‘ 42-1 28:3 < : : 28:2 
It will be seen that the first and second points do not agree at all well with 
the formula. With regard to them two things should be noticed. They are at 
a part of the curve near the axis of y, where an excessively small alteration in 
the density produces an enormous change in the resistance. Hence the liability 
to error is very great. Ifthe curve got by this equation, however, were plotted 
on the plate, it would lie very close to the experimental curve, even at these 
points where the divergence is greatest. Further, at these points the experi- 
mental curve would be slightly to the left of the hypothetical one,—that is, a 
little nearer to the axis ofy than the latter. Now the above formula assumes that 
the resistance of pure water is infinite. This is not absolutely the case, and 
hence the axis of y should not be a true asymptote, but should meet the curve 
at a finite distance from the origin. We might, therefore, expect that the curve 
should incline towards the axis of y, where its form is determined by weak solu- 
tions, more than it would if it were truly hyperbolic. Possibly this consider- 
ation of the finite resistance of pure water may account for the divergence in 
the case of the first two points. The other points all agree with the formula 
very exactly, the slight deviations being quite accounted for by the difficulty of 
keeping the solutions absolutely at the standard temperature. 
Since the curve is symmetrical about a vertical axis passing through the 
point of maximum conductivity, we have the means of forming an equation for 
the part to the right of that point. Considering the axis of y as transferred 
parallel to itself to the right to a distance, equal to twice the excess over 
unity of the density of the solution of maximum conductivity, we must write 
(578 — x), instead of z in the above equation. 
Then Mere PE Nik) eee 0 
y = 144416 (378-2) + 
or 
2°66 
= 9236 LOW Feo se 
Writing D for the density, and R for the sp. resistance, at 10° C., as before, we 
have 
