VISCOSITY OF WATER BY THE EFFLUX METHOD. . 135 
the application of Couette’s method when an inequality in the 
radii is known to exist. Let, as in § 12 and § 14, R, denote the 
‘mean radius of efflux’ of tube 1, etc.—i.e. the radius of a right 
circular cylinder of equal discharge—then putting, similarly to 
our previous notation, 
4 4 Ri ZK Ri 
R* =} (R{ +R?) ande = RigRt 
we obtain 
8 0 P,- P,- Pith et sey | (43) 
8q L,-L, 
that is, the correction / is eliminated, but not y. The term 2ey 
is probably however always negligible, but not so (2: +P,)e. 
This result is important as shewing that when the radii are 
unequal the deduced value of the viscosity depends not merely 
upon the difference of pressures in the reservoirs, but also upon 
their absolute amounts. 
I have not from Couette’s data reduced afresh the values of the 
viscosity, but accepting them for the particular temperatures of 
his own observation, reéxpressed them for other temperatures by 
formula (37). 
Cohen, 1893.—Cohen’s extended series of experiments, by the 
efflux method, to ascertain the effect of pressure on the viscosity 
of liquids, cannot be used to obtain absolute measures, as only the 
approximate dimensions of his tube,! and of the fall in pressure 
therein,? are given. In his own reduction his times are uncorrected 
ee although this does not sensibly affect the results, for the 
Suestead pointed out in referring to Réntgen’s experiments, the 
Boy of experiments seemed well worth rigorous reduction. For 
~ purpose of the reduction of the times, the volume of efflux— 
oumeh Was not given—may be computed from the known value of | 
the Viscosity, so that instead of finding 7’, by (5) we must use® 
Sole d. Annal Bd. 45, p. 666. 2 Ibid. p. 669. 
cca formala is of course derived by substituting for Q? in (5) its 
Only iin - equivalent in terms of the other quantities, supposed known 
Sufficient accuracy for the purpose of the correction. 
