128 Mr. A. Mallock. [Dec. 13, 



These experiments give directly the force which a known angular 

 twist of the wire exerts at the known radius of the divided circle. 



These experiments are recorded below, and it will be seen that the 

 results are very fairly consistent. 



The dimensions of the various parts of the apparatus are given in 

 the computation of the viscosity constant. 



In making the experiments on viscosity the velocity of the cir- 

 cumference of the cylinder A was made to vary from 0*5 to 50 metres 

 per minute. 



It was found that at all these speeds the force tending to turn the 

 inner cylinder B could be represented by the sum of two terms, one 

 varying as the velocity and the other as the square of the velocity ; 

 the latter being small compared to the former, even at the highest 

 speed. See Diagram 1. 



The cause of the square term seems to be that, owing to the action 

 of the bottom of the revolving cylinder, a circulation is set up in the 

 Quid in the annulus, the flow being up the side of the revolving 

 cylinder and down the side of the stationary one, the result being that 

 the fluid having the velocity due to a position near the outer cylinder 

 is by this circulation continuously carried towards the inner one, 

 thus making the variation of velocity in the neighbourhood of the 



