388 



On the Reduction of Tidal Observations. [Dec. 15, 



Accordingly the computation form has log sec , room being left 

 for the syllable " co " if necessary ; underneath this is written log , 

 and the computer will insert A or B as the case may be. 



There is usually required also a numerical factor l/f or Jf, or both ; 

 the logarithms of the l/f are found amongst the astronomical data, 

 and the logarithms of the constant Jf's are printed in the forms in 

 their proper places. 



The treatment of the 21 harmonic components derived from the 

 harmonic analysis of the 5 J^'s and ^'s is shown in the forms. 



The tides of long period. 



The processes involved in the evaluation of these tides are suffi- 

 ciently shown in the forms. 



Postscript. December 17, 1892. 



Correction to previous paper. 



An error has been detected on p. 333 of my paper " On the Har- 

 monic Analysis of Tidal Observations of High and Low Water," 

 'Boy. Soc. Proc.,' vol. 48 (1890). 



In the example of reduction certain tabular values, extracted from 

 Baird's ' Manual,' have the wrong signs attributed to them. Since 

 in 1887 the longitude of the moon's node lay between and 180, 

 the signs of i>, , i/, 2v" should be positive, instead of negative as 

 stated on p. 333. When this error is corrected the reduction leads to 

 the following comparison : 



