VISCOSITY OF WATER BY THE EFFLUX METHOD. 103 
in the height of the water in the reservoir, and in the manometer 
attached at the bore near the terminal. I have computed the 
coéfficient by the formula 
g 
m= 2%. (h-h 
J, (hho) 
h being the height, above the axis of the tube, of the water in the 
reservoir, and fA, that in the manometer: the derivation from 
formula (4) is evident. 
Values of m deduced from Jacobson’s measurements of the 
differences of pressure between the reservoir and within the tube 
near its entrance : 
b= 134". 71 7.912 10]. 10). 2a ee 
m=1-06 1:09 1:22 1:10 1:38 91 93 1-09 
O01 245.219. -3é 2-28-1700: Bh 6 
131 £26: 04 (1-54.06 079 1288 
89) «1-41 “82 
1-40 1-42 
The mean of these 28 results is 1-10, of these with the 6 pre- 
ceding, 1:14; this average gives a general confirmation to 
Boussinesq’s theory, but the individual results shew how, even 
under circumstances in which uniformity might be expected, it is 
not realized. 
The differences are certainly not due to mere errors 
of observation. 
We see therefore that Hagenbach’s correction 
must be increased 41%, Couette’s 12%, and that Reynolds’ dictum — 
is negatived : and further, that if this correction be of sensible 
magnitude, the deduced viscosity is to the extent of the uncertainty 
therein unreliable.1 
9. The correction to the length of the tube.—In the preceding 
article it has been shewn how the product of the pressure and 
time of efflux of a given volume of fiuid may be found, when the 
loss of head at the entrance of the tube vanishes or is eliminated. 
Following Hagenbach it has usually been assumed that this 
Correction is the only one necessary, or that the corrected differ- 
ences of pressures may be regarded as the differences at the 
‘erminals of the tubes; and Reynolds tacitly admits this view 
