No. 6.] DAWSON — FLUCTUATIONS OF LAKES. 311 



of rainfall and cyclones, and for the southern hemisphere, by a 

 discussion of his own and Mr. Meldrum's results. In the table 

 (p. 333) I have arranged the more accurate numerical observa- 

 tions of the height of the lakes from registers kept for the last 

 few years, in a method similar to that there adopted. 



Prof. Kingston's observations of Lake Ontario were taken at 

 Toronto, and measured upward from an arbitrary mark. They 

 extend from the year 1854 to 1869, and include the minimum 

 periods of 1856 and 1867, and the maximum of 1860. Taking 

 the mean annual level for each minimum and maximum epochal 

 year, and one year on each side of it, as is done by Mr. Meldrum, 

 and deducing a mean from each of three tri-yearly periods, the 

 agreement is close between the solar periods and those of fluctua- 

 tion in the lakes. 



The remaining observations are those of the U.S. Lake Survey, 

 and include only one period each of maximum and minimum in 

 solar spots. The measurements of the U.S. Survey are reckoned 

 doivnwards from a mark representing the high water of 1838 in 

 each of the lakes, but in the table here given they have been re- 

 duced so as to read upwards from an arbitrary line chosen 4 feet 

 below that datum. They are thus rendered more intelligible 

 and made to agree in sense with Prof Kingston's measurements. 



The result is the same in each of the lakes, only differing in 

 amount by a few inches. A mean deduced from the U.S- Lake 

 Survey observations in Lakes Superior, Michigan, Erie, and 

 Ontario, gives a difference between the years surrounding the 

 maximum of 1860 and the minimum of 1867 of 14.64 inches in 

 favour of the former. 



2. Diagram of Curves. — The curve representing the fluctua- 

 tion of Lake Erie from 1788 to 1857 inclusive is constructed on 

 a careful discussion of the evidence collected by Col. Whittlesey 

 and given by him most fully in the *' Smithsonian Contributions 

 to Knowledge" for 1860. 



From 1788 to 1814 there are no accurate measurements to 

 any well-recognised datum line, and I therefore give below the 

 measurements and approximations on which the general curve 

 for these years has been constructed. The description of the 

 fluctuation of the lake will be seen in many cases to apply with 

 verbal accuracy to the sun-spot curve. 



*' 1788— 1790. By tradition derived from the early settlers, 

 very high • according to some as high as 1838, but this is doubt- 

 ful. 



