ON THE SEICHES OF LOCH EARN. 379 



7th July and 8th August 1905. Six-inch wells were used in every case. In the 

 original limnograms the seiche was multiplied by four, and the time-scale was "2 inch per 

 minute. In the reproduction the scale is f . 



In Fig. 16, the upper limnogram AB was taken with 6 feet of |--inch tube; the 

 part C D of the lower with two access tubes each ^-inch diameter ; the part D E with 

 6 feet of |-inch tube. It will be seen that the part C D is much smoother than the 

 simultaneous limnogram above it, and that there is well-marked damping and consider- 

 able lag. On the other hand, the part D E goes pari passu with the simultaneous 

 portion of A B above it. Anything like exact conformity could not be expected, as the 

 curves are plotted from half-minute eye observations. Owing to accidental wind dis- 

 turbances, accurate quantitative determinations of damping and lag from such an 

 observation are impossible. 



In fig. 1 7 is given the result of a similar pair of simultaneous observations. A B 

 was taken with 6 feet of ^-inch tube ; C D with 1 2 feet of ^-inch tube ; and E F with 

 12 feet of ^-inch tube. 



Experimental Determination of the Seduction Constant x- 



The constant x> which plays so important a part in the theory of the ordinary 

 limnograph, could scarcely be calculated with much accuracy from the dimensions of 

 the apparatus. The best method would be to determine the lag and damping of 

 oscillations of known period and range artificially produced, which could be done in a 

 laboratory without much difficulty. 



A rough determination of x at the lake side could be made on a day when there is 

 no seiche,* by observing the times that the level in the well takes to fall through given 

 distances after it has been disturbed by pouring water into the well. 



Suppose that at time t = 0, the level of the well is h cm. above the normal height; then, 

 there being no seiches, A = ; and we have from (6) x = d + ~Le~ xt . When t = 0, x = d + h ; 

 hence L = h. Therefore x — d = he~ xt . Let the level at times t' be h' ; then we have 

 V = he- xt '. Whence x = log.(A/A')A'. 



Theory of the Statolimnograph. 



For simplicity, we suppose the instrument so adjusted that the mean level 

 of the water in the well is the same as the mean level of the lake, and that 

 the pressure inside the cylinder of the statoscope is equal to the atmospheric 

 pressure when the water in the well is at mean level. And we also suppose 

 that neither the temperature of the air inside the statoscope nor the atmospheric 

 pressure outside varies. 



* Or even when there is a small regular seiche giving a smooth limnogram. 



