IN THE NORTH AMERICAN LAKES. 7 



when the rise occurred. Tliis bclicjf sliows the tendency to liasty generalization, 

 and the superstitious proneness to attribute to the number seven a peculiar appli- 

 cability to the recurrence of natural phenomena. 



By examining the table we have given, containing observations that have been 

 made since 1819, there will appear a continual rise until 1838, a period of nine- 

 teen years, without any decline. Other tables show an uninterrupted decline from 

 1838 to 1841, three years; in 18-11, a slight rise; from 1842 to 1851, a regular 

 decline of eight years. During a space of thirty-two years, there is no instance of 

 a return of high water in the period of seven years. For the years since 1838, I 

 am able to offer a much more satisfactory exhibit. To simplify the result, I have 

 constructed a diagram of curves whose ordinates are the monthly average of the 

 surface reduced to the Buffalo zero for such months and years as have a good mean. 

 There are four years complete, the means of which are consolidated into one curve, 

 which is placed, to prevent confusion, below the other curves. See Plate I, No. 1 . 



The regularity of the annual rise and fall is evident from an inspection of the 

 form of the curves. The months of June and July are high as compared with other 

 months, whatever the general level may be. A depression follows immediately, 

 which reaches the lowest points in the months of December and January. This is 

 the law, to which there are exceptions, arising from variations of the seasons. In 

 fourteen of the best ascertained years high water occurred in June and July ten 

 times ; in ten years, the annual decline reached the lowest point in the months of 

 December and January six times. There is, therefore, a spring flood and a winter 

 ebb, the same as in the Mississippi and other large rivers or ponds. The surplus 

 water due to melting snows and spring rains causes an accumulation of water. In 

 winter the frost and drought, by diminishing the supply, causes the surface to settle 

 below that of summer. The amount of fluctuation within the year, deduced from 

 sixteen years' observation, is as follows: — 



Cleveland, greatest average montlily difference of high and low water . 

 Detroit, " " " " 



Buffalo, " " " " 



Mean annual difference of highest and lowest months . . . . 1 H 



Lakes Huron and Michigan have not received much attention; but are known 

 to have been high in 1838 and low in 1819. It docs not necessarily follow that 

 the highest or lowest level of different Lakes will occur at the same time, nor 

 that the quantity of rise and fall should be the same. There should be, however, 

 in all of them an annual flux and reflux, and also secular fluctuations. As the 

 lower Lakes receive more water from those above during years that are high than 

 they do when there is a depressed surface, there should be a greater range between 

 high and low water in them than in those nearer the source of supply. Lake 

 Superior is the only one of the chain that exliibits the effects of conditions strictly 

 its own. 



