102 G. H. KNIBBS. 
to remark that this correction must be not only well ascertained | 
but also small, for accurate measurements of viscosity, since there 
is ample evidence that it is variable in amount even under con 
ditions where one would anticipate constancy. In illustration of 
this we take for data the results of several series of measurements 
of pressure-heads and velocities at constant temperatures made 
by Jacobson,' with his brass tube C, whose radius and lengths 
were respectively 0-14328, 62-04, and 15-76 c.m., and which was 
so fitted that its influx terminal was flush with the inner surface 
of the large reservoir which supplied the effluent.?2. For measure 
ments of pressure-head and mean ae | we have, sinceU/ = q/(7# E 
— ater fi 
in which c’ =8 ZL n/(pg R?), and is oes constant while the 
temperature is so. The value of m was found by the graphic 
method, the above being a linear equation.’ 
Values of m deduced from Jacobson’s experiments. 
No. of experiments in series 5 i 7 5 9 
Deduced value ofm .... 1:38 1:25 1-44 1:30 158 
Another result from a series of five measurements with a tube of 
0°12516 radius and 15-77 c.m. length,* was m= 1-08. 
The fall in pressure at the entrance of the tube was also directly 
observed by Jacobson,® his experiments being made with 4 brass 
tube of 0-2545 radius and lengths as shewn, and the temperature 
maintained constant for the measurements of any one series. 
tube was bored through at a point 0°15 c.m. from its infos 
terminal? and the fall in pressure ascertained from the differen 
1 Archiv. fiir on u. Phys., 1860, p. 93 and 1861 p. 327. 
2 Thid. 1860, p. § 
3 Jacobson’s alle t. p. 94, said to be obtained from the sheet 
tions by the method of least squares, should, when multiplied by ee 
give values sensibly equal to these. They are nearly so. 
* Ibid. 1861, p. 327. 5 Thid. p. 309. 4 
6 It should be mentioned that the coéfficient ¢ p. 322 of the 1861 volume i 
Archiv. f. Anat. &c., when multiplied by g and expressed in the one : 
units 18 our m, and that its values for Poiseuille’s tubes, p. 326, are are ver : 
nearly identical with those hereinbefore given. - 
os 
