492 PROFESSOR CHRYSTAL 



and almost or altogether absent at the windward end. It also depends on the amount 

 of shelter at the point of observation. 



In most cases the occurrence of embroidery is accompanied by the characteristic wind 

 blurring on the microbarogram, or else by fluctuations of very short period and very 

 small range. In some cases the fluctuations could be counted ; and in one or two their 

 period seemed to coincide with the period of the lake vibrations. The sensibility of 

 the microbarographs used and the number of interpretable cases were not sufficient, 

 however, to justify any general conclusion. 



Attempts were made to connect the periods of the lake vibrations with the periods 

 of the wind fluctuations, as indicated on the anemogram, but without success, possibly 

 owing to the fact that the time scale of the anemograph was so short that it was 

 impossible to count the wind fluctuations with any certainty. 



The simultaneous limnograms taken on Earn and Tay during October and November 

 1905 were examined to see whether there was any connection between the vibrations 

 on the two lakes pointing to a common atmospheric cause. It was found that the 

 average of the maximum ranges and of the periods was much the same for both lakes ; 

 but there seemed to be no connection between the occurrence of a particular range or 

 a particular period in the two. The range might be high in both lakes and the periods 

 different ; or the periods nearly the same and the ranges different ; or there might be 

 vibrations of considerable range on one of the lakes, and none, or only the merest tremor, 

 on the other. 



Several suggestions have been or may be made regarding the nature of these lake 

 vibrations. 



I. They might be longitudinal seiches of very high nodality. This was the sug- 

 gestion put forward tentatively by Forel, after trying in vain every other explanation 

 that occurred to him. 



If the period of l'47 m were due to a longitudinal seiche, the number of the nodes 

 would be 12 or 13. It is easy, by regarding Earn as a symmetrical rectilinear lake,* 

 to calculate roughly the positions of the nodes. It would therefore be possible, by 

 means of careful experiments with two or more self-registering instruments, such as the 

 statolimnograph, to obtain positive or negative evidence regarding the truth of the 

 hypothesis that the vibrations are wholly or partially plurinodal longitudinal seiches. 



In the present state of our knowledge the balance of evidence seems to be against 

 this hypothesis. A plurinodal seiche is a simultaneous oscillation of the whole lake. 

 If, therefore, a vibration were a plurinodal seiche, it should be apparent simultaneously 

 at both ends of the lake ; whereas we know that it may be present at either end and 

 apparently absent at the other. Also, if it be a plurinodal seiche, it should be present 

 simultaneously at nearly opposite points on the two sides of the lake. We made 

 repeated attempts to detect correlations of phase, by stationing observers on the two 

 sides, and signalling the maxima or minima of the vibrations, but were quite unable to 



* See my memoir on the " Hydrodynamical Theory of Seiches," Trans. B.S.E., vol. xli., p. 639 (1905). 



