102 ROYAL SOCIETY OF CANADA 
of internal friction have been accurately determined, but I have not found 
the necessary leisure for this investigation. In the following table is given 
a curve for water between 20° and 45° C. The number of seconds for 
delivery of 50 che. with nozzle A, I 44. and also the viscosity co-efficients 
X 10° are taken as ordinates, and the temperatures as abscisse. The 
absolute viscosity numbers are from a table by Thorpe and Rodger. 
































































































(Phil. Trans., 1894, 449) : a 
s S 
N 
à 
Curves showing that the McGill instrument indicates a viscosity & 
greater than the true interior molecular friction, in the case Ÿ 
of water. os 
= 
125 Sec $ 
124 À 
123 à 
122 1000 
121 975 
120 950 
€ 119 925 
118 900 
II 11786875 
NOS 116 850 
AUS 115 825 
ES 114 800 
os Es 1s 775 
ry NE 
f a 112 750 
= 725 
tS 110 700 
109 675 
108 650 
107 625 
| 106 600 
HHH f05SeC.975 
90 
I propose to investigate the following points : 
1. The relations to each other of the curves of viscosity as obtained 
with nozzles I., IJ. and III. 
2. The interpretation of these curves in terms of interior friction, ete. 
3. The limiting temperatures at which the viscosity of oils, ete., may 
be profitably made for practical purposes. 
4. The advantages, if any, in expressing results by a curve instead of 
by absolute values obtained at definite temperatures. 
5. The desirability of working with a suite of nozzles when carrying 
on viscosity tests through a long range of temperature. 
6. The maximum and minimum deviation of single tests, with a view 
of rendering errors of experiment in reading, change of temperature, 
volume delivered and time of flow, so small as to be negligible. 
