256 Dr. Buchanan and Mr. Malcolm : 



The position o£ these graphs relatively to one another, 

 depending so intimately on temperature as it does, can, 

 however, only be considered as on the whole substantially 

 correct. 



The most complete observations were made with the small 

 drum. In figs. 1, 2, and 3 (PL IV.) are shown the graphs 

 connecting the liquid friction and the load on the cord, with 

 a specified constant volume of liquid in the drum. These 

 graphs have been reduced to approximately a common zero by 

 subtracting from the total friction the friction of the empty 

 drum. 



When the drum was completely full of liquid, it will be 

 noticed that (a) the graphs for all three liquids are approxi- 

 mately straight lines ; (o) the liquid friction for the same 

 load is greatest for water and least for castor-oil ; (c) the 

 approximately horizontal straight line for castor-oil shows 

 that the drum when full of this oil behaves almost as would 

 a solid body. 



When the drum was not full these graphs show clear! y 

 that the normal result is that the liquid friction for castor- 

 oil diminishes continuously as the load increases ; for water 

 the rise is continuous and nearly linear ; tor lamp-oil it rises 

 to a maximum, then diminishes. 



In fig. 4 the graphs refer to similar experiments with the 

 large drum. Inspection shows (d) the resemblance in form 

 between those for water and for lamp-oil ; (e) this form is 

 similar to that of the left-hand portions of the lamp-oil 

 graphs of fig. 2. 



The experiments with the large drum were projected in 

 order to discover whether the results would present any 

 fresh points not shown by the small drum. The graphs 

 show, however, that the general results are of the same 

 nature with the two drums. The large drum gives results 

 which appear to be merely exaggerated versions of those 

 given by the small drum. 



The behaviour of castor-oil under the action of small 

 loads is somewhat striking, and is indeed a consequence of 

 the fact that the liquid friction with this oil decreases with 

 increased acceleration. For example, with 320 c.c. in the 

 small drum, loads of 50 grms., 75 grms., 100 grms. respect- 

 ively gave simply a slow, apparently uniform, creep-down of 

 the load to the bottom of the descent, where it stopped dead. 

 Here, of course, the total friction was equal to the load. 

 This action is indicated by the straight line AB in fig. 1 ; it 

 is brought out still more fully in figs. 5 and 6, 



