144 G. H. KNIBBS. 
p the density of the mercury therein, and p that of the water, both — 
at their own temperatures at the time of the measurement.’ 
As already pointed out no correction for change of dimensions 
in bulb or capillary is needed, but these must be given for 4 
common temperature. 
The only further corrections then are :—(c) the correction for — 
fall of pressure at the entrance of the capillary, (d) that for fall 
in pressure producing the flow in the tube J A and B Z, and(@)a — 
possible correction to the length of the tube to equate the terminal ‘ 
resistances—§ 9. Of these (c) and (e) have been already fully 
discussed: (d) only requires attention. Let R and S§ denote — 
respectively the radii of the capillary and supply tube, and Z and : 
Z their lengths, then the fall in pressure p, in the tubes JA + BE 
: a= ey =. (46) 
This must be so small as to be practically negligible. ; 
29. Dimensions of apparatus.—The general principle which — 
determines the relation of the several parts of the apparatus is 
that the resultant precision should be uniform throughout. The 
following measurements are merely suggestions as toa suitable — 
size for the essential parts : ‘they serve also the purpose ¢ of illus 
trating the application of the above principle. We shall assum E 
for g the value 980°6 c.m. and for p 13-5958. : 
Pressure in manometer H = 100 cm. mercury: hence P= 1333200 : 
Radius of capillary R=0 ‘Olem. Length 2=50 cm. 
Bulb V in Fig 4 Q=50cem. Rad. bulb 2:28 cm, about 
Supply tube not less than S=0-lem. Totallength Z= 25 cm. say a 
In apparatus of these dimensions the fall in pressure in the | 
supply tube J A+B a Z will be not more than = 
pein ea ‘ 
3 i ae Sey aD le 
(Fu)* x ae 133320 = 67 or 1/20000 th. 
formula (46). The rate of flow even for 0° C. would be sulticiently 
ee 
1 The variation in the loss of pressure at the entrance to the tube is? 
small that it may be computed from this mean pressure, without ici 
