144 



MR. LUBBOCK ON THE TIDES. 



Tlie correction for parallax 57' being considered as given by the semimenstrual in- 

 equality, Bernoulli's correction for the influence of the moon's declination is equally 

 erroneous ; but his accurate determination of the law of the semimenstrual inequality 

 is one of the most important results ever obtained a 'priori by means of the theory of 

 universal gravitation. 



The theory of the tides is now, as Mr. Whewell remarks, in the state which that 

 of the motions of the moon and planets presented about a century ago ; and unless 

 considerable exertions be made, it may so continue for many years to come. The 

 tables of the planets have only acquired their present accuracy through the liberal 

 encouragement of learned bodies, and of some of the governments of Europe ; nor 

 can tables of the tides, adapted to the present state of science, be now constructed, 

 unless a very considerable expense be incurred, from the immense labour required. 



In discussing tide observations, when the greatest possible accuracy is desired, and 

 when the correctness of the observations appears to warrant such a nicety, in order to 

 obtain any particular correction, all known approximate corrections of a different 

 nature should first be separately applied to each observation with a contrary sign. 



So, in the determination of the correction for the calendar months, it would be well 

 to correct all the observations first for parallax and declination ; and this considera- 

 tion should be particularly attended to, where it is practicable, in attempting to de- 

 termine, from a few observations, the establishment of any port and the semimenstrual 

 inequality. The best method of verifying my tables would be to determine by their 

 means the times and heights of high water at the London Docks for nineteen years, 

 and then to classify the transits of the moon with the errors of those determinations, 

 according to calendar months, and according to parallaxes and declinations: the 

 average error corresponding to each calendar month would be the error of the Table* 

 for the calendar months, and so for the rest. 



* I allude now more particularly to that which accompanies this paper, not to any given previously. 



