140 G. H. KNIBBS. 
those in the last column—the means—we obtain 
Tap © 00 06 1 2 3 4 5 Oe 
Dif, O-4 04+13411+104+11-34+6 8 
that is the differences are generally positive and are 8 
The curve best representing the mean values must therefore t 
the form 
: F's = 1000 (1 + ar - f’r”) 
between these limits, and since from 5° to 45°C. the 
coéflicient 8, was positive—formula (37)—there is an inflexion : 
the curve. To determine the point of inflexion, extremely accum® 
measurements of the viscosity would be necessary. The chang 
in the form of the curve from 0° to 5° somewhat vitiates 
extrapolation—though it be for only a half degree—by bs 
efflux time for 0° C. was obtained in § 20. 
26. The variation of the fluidity with presswre.—The sna 
pressure is to increase the fluidity or to diminish the viscosity, ¥ 
least for temperatures under 30°C.!_ This is best shewn by 
method of expressing the value at each temperature for 
atmosphere as unity: in this way the results become indepe! 
of the relative fluidities for the different temperatures” 
values for a pressure of 24 atmospheres are from Rantgen's ® 
those for higher pressures from Cohen’s experiments.” 
Effect of Presswre on Fluidity at 15° C., according to Cone 
Experiments. : 
Atmos. = 100 200 300 400 500 600 700 
I. +0072 0127 0150 0219 0208 —— 0280 
IT +-0051 0126 0161 0223. 0258 —— 024 
ete. 60162 — 0239 = 
Mean +-0057 0127 0154 0221 0233 0239 0237 © 
oe -—The flui idity under a pressure of one pers sigs is unity: # 
at any other pressure is 1+ the decimal in the t 
es A 
1 Under 40° C. according to Cohen— Wied. Annal. Bd. 45, P- oak 
2 The relative fluidity cannot be accurately determined from the ¢ . 
* The figures are not alwa this h 
ys identical with Cohen’s: 
~ beta to the — of the small correction to the tim 
