252 Dr. Buchanan and Mr. Malcolm : 



liquid be determined. The results obtained by us in the 

 time, and by the means, at our disposal, appear to us to be 

 new and of sufficient interest to warrant publication. 



Apparatus used. 



Two drums were tried, a small and a large. These were 

 made o£ brass, and for convenience of cleaning can be sepa- 

 rated circumferentially into two portions. When either drum 

 is in use, the two parts can be pressed together by a brass nut 

 which engages with a screw-thread cut on the steel spindle 

 passing axially through the drum. Leather washers keep 

 the joints tight. For admission of the liquid to the interior 

 of the drums each has provided on it a small nipple closed by 

 a nut; a dummy nipple and nut are so placed as to balance 

 the inertia of the other. 



The same steel spindle and its bearings serve for one or the 

 other drum. The spindle has coned ends which turn on fixed 

 set-screws of steel furnished with lock-nuts, so that the 

 spindle can turn freely and yet have scarcely any sensible 

 end play. The other parts, comprising side-brackets and base, 

 are made of stout cast iron for the sake of rigidity. 



The principal approximate dimensions are as follows: — 



Mean depth of small drum inside — 3-5 cms. 



Mean depth of large drum inside = 19-0 „ 



Mean internal diameter of both = 1204 „ 



Diameter of spindle to top of thread = 1*25 „ 



Diameter of axle on which cord wraps, to 1 _ -i k 

 centre of cord J 



Theory of the Experiments. 



In the case of such extremely complex internal motions as 

 we have to deal with here, it appears to be hopeless to attempt 

 quantitative measurements other than those of a kind which 

 will serve to bring out the average effect of such motions. 



It would seem that the most convenient means of expressing 

 this average effect is to take it as measured by the amount of 

 retardation of the falling load over and above what the retar- 

 dation would have been if the drum and its contents ivere 

 moving as a solid. 



As will be seen presently, this quantity — the additional 

 retardation due to the fact that the interior of the drum 

 contains a liquid, and not a ri«id material — can be measured 

 without much difficulty by allowing the falling load to run 

 down to the full extent of the cord, and then permitting the 

 inertia of the rotating mass to rewind the cord on the axle, 



