1892.] facilitating the Reduction of Tidal Observations. 383 



M, 2SM, MS, when the computing strips are written for the third 

 time, we must remove strips 59, 60, 61 (which have numbers written 

 on them) and may leave the remaining strips of that writing which 

 are blank. When the strips are written for the fourth time strips 

 0, 1, 2, 3 will be blank, but we must remove strips 4 to 13 inclusive. 

 When all the strips are used in a complete year there are 369, and 

 this is the divisor used in obtaining the harmonic constants, but 

 when there is this supposed hiatus we do not use 15 strips of the 

 third writing and 14 strips of the fourth writing, so that the divisor 

 will be 340. 



Again, when we are evaluating in the third writing, strips 57, 

 58, 59, 60, 61 must be removed, and in the fourth writing strips 

 4 to 9 inclusive. In a complete year the divisor is 369, but we now 

 do not use 17 strips of the third writing and 10 of the fourth writing, 

 so that the divisor becomes 342. 



Again, in evaluating N, L, J, Q, in the third writing we remove 

 strips 45 to 61 inclusive, and in the fourth writing strips 4 to 25 

 inclusive. The divisor is reduced from 358 to 303. 



Lastly in evaluating i/, in the third writing strips 43 to 61 inclu- 

 sive are removed, and in the fourth writing strips 4 to 31 inclusive. 

 The divisor is reduced from 350 to 287. 



Any hiatus, be it long or short, may be treated in this way, but it is 

 clear that if it be short enough to treat by interpolation, it is best to 

 adopt that method. 



INSTRUCTIONS FOB USING THE COMPUTING 

 APPARATUS. 



The apparatus for the reduction of tidal observations, together with 

 computation forms, can be purchased from the Cambridge Scientific 

 Instrument Company afc a price (as far as can be now foreseen) of 

 about 8. 



In case of any insufficiency in the following instructions recourse 

 must be taken to the preceding paper. 



On the degree of accuracy requisite in the hourly heights. 



It will usually be sufficient if the heights be measured to within 

 one-tenth of a foot, and the decimal point may, of course, be omitted 

 in computation. 



This gives amply sufficient accuracy at a place where the semi- 

 range of the principal lunar tide is 2 ft., and where spring range is 

 from 6 ft. to 7 ft. 



At some places with small tides a smaller unit might be necessary, 

 and at others with very large tides a unit of 2 in., or of a fifth of a 

 foot, might suffice. 



