Experiments with Rotating Viscous Liquids. 257 



Again, when for instance the large drum contained 89 c.c. 

 of castor-oil, a load of 150 grms. gave initially a very slow, 

 seemingly uniform, motion, until about a revolution of the 

 drum had heen completed : at this stage the accelerated 

 motion began and continued thereafter in the normal manner. 

 An effect of this kind frequently occurred with small loads, 

 but it doe- not appear to be possible to represent the effect 

 by points on such graphs as we have here. 



The graphs representing the relation between the total 

 friction and volume of liquid in each drum, with a specified 

 load hung on the cord, are shown in figs. .'» and G. In order 

 not to confuse the graphs, the ordinates here represent the 

 total friction = axle friction + fluid friction. 



The axle friction is very nearly that for the empty drum. 



The liquids were poured into the drums from a measuring- 

 cylinder. With the oils especially it was very difficult to 

 measure out exactly a definite volume, so that there was an 

 accumulation of small errors in measurement when the drum 

 was full. Due to this cause are the discrepancies between 

 the right-hand terminations of the graphs in figs. 5 and 6. 



These graphs show clearly that (/) the type of graph is 

 the same for the two oils used. They indicate an increase — 

 at one stage it maybe a very rapid increase — of friction with 

 quantity of liquid in the drums. This is followed afterwards 

 by a decrease, more or less rapid, until the drums are full ; 

 (g) for small quantities of liquid, and the same load, the 

 friction is less with castor-oil than with the lamp-oil. There 

 is a certain volume for which the friction is the same with 

 both oils ; (/*) in the case of the castor-oil, the effect referred 

 to above, where the friction is equal to the load, is brought 

 out 1 >y these graphs in a very marked way. Thus, the graph in 

 fig. 6 marked for 100 grms. load, with castor-oil in the small 

 drum, rises up to the 100 grms. ordinate ; that in fig. 5 for 

 300 grms. load, with the same oil in the large drum, rises up 

 to the 300 grms. ordinate. 



With the large drum the effect in question was made still 

 more striking by beginning with the large drum quite full ; 

 a load of 300 grms. caused a rapid acceleration. About 

 one-eighth of the contents of the drum being withdrawn, 

 there ensued a slow creep-down of the load. 



Aberdeen, 



31st October, 1904. 



Phil. Mag. S. G. Vol. 9. No. 50. Feb. 1905. 



