a Sa ities hn a calamea 
sae Saenees : 
pap y es Soap ial ee on 
ES ee Oe ee ee Be ee eng on Lee Weert eR) ST Eee Re ean rm et Mater te Pee) th SN See eee am eR wey ey: 
VISCOSITY OF WATER BY THE EFFLUX METHOD. 131 
ge Ry 7 
= — = —_-  —=E +7 . 
"Soh ocd cli ore eg 
& being the coéfficient of cubic dilatation. Thus, for measurements 
of dimensions at all ordinary temperatures we have with ample 
precision 
4 . 2 
ho T(1l+7ré+p'— p) - oo , ae \ Be ae (41) 
The correction represented by the term ré is rejected by Meyer, 
in his reduction of Rosencranz’s observations, as negligible!: its 
maximum value is however about 1 /500, and it affects the com- 
puted values of » from 4 to 7 units in the last place of figures, viz. 
the fourth, 
The pressure registered by the manometer was throughout 59°3 
centimetres of mercury at 18°C. Meyer takes this to be of 13-415 
density and equivalent to a column of water of 795-6 c.m. height : 
to this has to be added 9-0 c.m., owing to the disposition of the 
apparatus, thus making a total of 804-6 c.m.?. But this value for 
the density is, I suspect, an error, and I have therefore recomputed 
the pressure, obtaining the result 812-59 grammes per square 
centimetre. The bulb measuring the influx contained 111-03 
grammes of water, which [ take to represent a volume of 111-178 
cubic centimetres at 18° C., the laboratory temperature.* The 
capacities of the tubes were found by filling with mercury and the 
mean radius so determined.‘ The reduced results can be expressed 
in the form> 
0 
Ie Tear 
but they do not appear to be satisfactory either intrinsically, when 
Compared with Graham’s observed values, or with those obtained 
by extrapolating the curve given by Poiseuille’s measurements. 
1 . 
in ag Annal. Bd. 2, p. 399—* doch kann diese Correction ihres ger- 
2 Ths Bes Wegen unterbleiben.” 
M =< P. 397. 3 Ibid. p. 397. 
; See also Pogg. Annal. Bd. 148, p. 36, 1873. Tube ut. 
To WAS “01854 and « -04635 
