VISCOSITY OF WATER BY THE EFFLUX METHOD. 101 
Tubes A, and A,, B, to Bs, C to C,, D, to Dy, H to L,, and F 
give no satisfactory indication as to the value of this factor, For 
tubes A and D the following results were obtained, viz., 0-4 or 
0:5, and 1-2, these values however are entitled to but little con- 
fidence inasmuch as they depend on lines the directions of which 
are determined by the positions of three points only, two of these 
being close together so far as the abscisse are concerned, but in 
regard to the ordinates, not sufficiently concordant to warrant 
confidence. For tubes A, and B the results depend respectively 
on 2 and 3 points: in the other cases on from 7 to 10. 
It is difficult to assign a reason for the values 0°82 for /’, and 
F; ; possibly the flow was turbulent ; the values for the viscosity 
—see the table hereafter—point to that. But one would have 
expected that A, would also have given similar indications, which 
it does not excepting for the last member of the series. The 
value is of course that for the fall of pressure known to occur in 
efflux through a short adjutage. The value for C’, may perhaps 
be accounted for by the presence of a particle at the entrance of 
the tube interfering with the flow. The general result however 
agrees remarkably well with Boussinesq’s deduction from kinetic 
theory, and since in the instances of some of the tubes showing 
discordant values the amounts of the corrections are much smaller 
than the discrepancies among the observations themselves, no 
doubt need be entertained as to the propriety of its application 
where there is no evidence of the particular value of m. 
It may here be noticed that in illustrating his theory of the 
Correction to the pressure Couette selects series A, and A, to 
exhibit the advantage of his formula over Hagenbach’s'. For 
these tubes m is respectively 1-02 and 1-08. For the former the 
factor is sensibly unity and therefore the agreement with Couette’s 
wealt is Satisfactory ; in regard to the latter the progressive 
merease of the corrected values of » is clearly a consequence of 
the fact, that the correction factor is too small. This leads me 
DUO eos ce ee 
1 Annal. de Chimie, 6 sér. t. 21, p. 502, 1890. 
