160 Dr. K. A. Houstoun on the Damping of 



the vertical acceleration into consideration k becomes smaller,, 

 instead of greater. 



The difference appears to be due to dissipation of energy 

 in the surface-film. All the other conditions being the same,. 

 k increases with the time. The following figures taken for 

 h = 5 cms. exemplify this : — 



July 2nd, 3 p.m k = '022 



„ 3rd, 10 a.m -036 



„ 4th, „ -040 



„ 6th, „ -049 



„ „ „ (surface cleaned) '025 



On July 2nd the trough was cleaned, filled with clean tap- 

 water, and k determined. The water was then left standing 

 in the trough, and k determined on the 3rd, 4th, and 6th. 

 Immediately after the first determination of k on the 6th the 

 surface was " scraped," the film removed, and k fell to half 

 its value. 



The values of k determined from a fresh surface were- 

 never very consistent ; the motes were drawn in different 

 directions owing to the film forming. After the surface had 

 stood some time the values of k were more regular. 



Since k increases so much with the time — an increase 

 which we can explain only by the formation of a surface 

 film — it seems reasonable to suppose that the difference 

 between the value of k for a fresh surface and the theoretical 

 value is due to a film which forms with great rapidity. In 

 other words, none of the surfaces were perfectly clean. This 

 view is strengthened by the fact that the addition of minute- 

 quantities of lubricating oil or linseed oil to a fresh surface 

 had no appreciable effect on k. Apparently there was a film 

 already there. 



The impurity seemed to come from the atmosphere of the 

 laboratory and from the putty at the corners of the trough... 

 No improvement was obtained on cleaning the trough with 

 acid or on using distilled water. When a smaller trough, 

 which could be kept clean more easily, was used, the waves- 

 were not suitable for accurate measurement. 



According to the table the effect of viscosity on the period 

 agrees with its calculated values ; the fourth and fifth columns 

 agree. Some other measurements of the effect of viscosity 

 on the period are given below. For these h was calculated 

 from the volume of the water, and the surface was not at alL 

 so clean. 





