CHAP. IX.] TIDES 231 



into hours with vertical lines drawn through them. These 

 hours are those of the moon's superior transit. Directly under 

 the position on the line of the local mean time of the moon's 

 superior transit is plotted, as a dot, the high water next follow- 

 ing that transit, at its proper position for height of the tide, as 

 measured by the side scale in feet. When the next day's 

 similar high water has been similarly plotted, the intermediate 

 high water is plotted half-way between them at its proper 

 height and the two low waters similarly interpolated. 



The high waters following the superior transit are joined 

 by a red line, and the intermediate high waters by a blue line. 



Similarly with the low waters, taking care to note which 

 are the low waters next following the superior transit. Each 

 dot is joined to its next successive high or low water by a 

 black line. 



The lunitidal interval for each high water is plotted in the 

 place provided, and the dots joined. By adding the moon's 

 superior transit to the lunitidal interval as plotted, the mean 

 time of each high water can, if necessary, be ascertained. 



When the curves of the sun's and moon's declination, and 

 of the moon's parallax, are plotted, the general movement of 

 the tide, and its relation to the positions of the sun and moon 

 in declination, can at once be seen. 



The mean between each high and low water height will give 

 roughly the mean water-level. 



To obtain the true mean water-level in a few days, other True 

 observations must be taken, and we subjoin an extract from 5!.®^'^ ^ 

 the " Instructions to H.M.S. Challenger,'''' which contains full level, 

 directions, only adding that these observations must be made 

 when the tide is moving normally, that is, when there are no 

 strong winds to raise or depress the water-level. 



" A good determination of the mean-level sea by the simple 

 operation of taking means may be made, in less than two days, 

 with even a moderate number of observations 2^operly distri- 

 buted so as io subdivide both solar a7id lunar days into not less 

 than three equal parts. Suppose, for example, we choose 

 eight-hour intervals, both solar and lunar. Take a lunar day 

 at twenty-four hours forty-eight minutes solar time, which is 

 near enough, and is convenient for division ; and choosing any 

 convenient hour for commencement, let the height of the 



