042 PART IV. : EXPERIMENTAL VERIFICATION OF HYDRODYNAMICAL THEORY. 



measuring the depth of each layer and of "002 in p — p ', we find the possible error 

 introduced in the calculation is 1 per cent. The error in observing the period may 

 amount to I per cent., and leaves therefore an error of only V6 per cent, in the 

 correction applied (1) on account of the ratio h/l being large, and (2) on account of the 

 variation of the normal curve from the parabolic form. This is as close an agreement 

 as could have been expected. 



§ 52. On comparing the observations in the plain trough with those in the truncated 

 trough, we find that the period of oscillations in the former is generally the greater, 

 but the difference is always small, and within the limits of error the periods are the 

 same. In calculating the period of the temperature seiche in lakes, an end correction 

 may be omitted without introducing any considerable error. There is no apparent law 



Fig. 7. 



Fig. 8. 



for the difference between the two periods, but the period of the plain trough is 

 seemingly about 1 per cent, longer than that of the truncated trough. 



§ 53. A few tentative observations were made with a thin layer of olive oil inter- 

 posed between the water and the paraffin, olive oil being heavier than paraffin, but 

 lighter than water. 



The following are the observations made with the layer of olive oil in the truncated 



parabolic basin : — 



Z = 43-26 cm., h= 15*35 cm. (water), /j/ = 7"43 cm. (paraffin). 



1. Olive oil layer -1 — -2 cm. 



T = 3-02 ±-02. 



2. Olive oil layer - 5 cm. 



T = 306±-02. 



3. Olive oil layer 1 cm. 



T= 3-07 ±'01. 



There is a slight lengthening of the period caused by the interposition of the oil which 

 bears out the assertion made in § 29 that the effect of lack of sharpness in the dis- 

 continuity of temperature is to lengthen the period of the temperature oscillations. 



