i 



245 



Maskelyne's experiment on Schiehallion, but with the superior ad- 

 vantaires arising from the perfect homogeneity of the attracting mass, 

 and from the circumstance that all the observations might be made 

 at a single station. The experiment might, in short, appear to unite 

 the advantages both of Maskelyne's and Cavendish's methods of de- 

 termining the earth's density. 



The suggestion was Dr Robison's, and Dr Chalmers had it from 

 him. It is contained in the Elements of Mechanical Philosophy, 

 Edit. 1804, page 339, and is given in the following words: — " Per- 

 haps a very sensible effect might be observed at Annapolis-Royal in 

 Nova Scotia, from the vast addition of matter brought on the coast 

 twice every day by the tides. The water rises there above 100 

 feet at spring tide. If a leaden pipe a few hundred feet long were 

 laid on the level beach, at right angles with the coast, and a glass 

 pipe set upright at each end, and the whole filled with water, the 

 water will rise at the outer end, and sink at the end next the land 

 as the tide rises. Such an alternate change of level would give the 

 most satisfactory evidence. Perhaps the effect might be sensible on a 

 very long plummet, or even a nice spirit-level." 



It is needless to observe that the methods proposed by Dr Robison 

 are not the best which might be suggested ; but that, in consequence 

 of the extreme simplicity of the observation, considered as a purely 

 astronomical one, a deviation of the direction of gravity of only a 

 very few seconds could be ascertained within small limits of error.* 

 I thought it worth while to make the calculation approximately 

 for an assumed height of the tide-wave. Had the result been at all 

 encouraging I should have taken pains to ascertain, on good autho- 

 rity, the exact rise of the tide, and the circumstances of the locality 

 whence the rise is greatest. 



I have calculated the horizontal attraction of a semicylinder of water 

 100 feet thick, and of about two, four, and eight miles radius upon 

 a point at the extremity of the axis of such a semicylinder ; because 

 these conditions can easily be reduced to calculation, and because 



* The micrometric observation of a plumb-line, as in a zenith sector, would 

 be sufficient ; or, as Professor Smyth has suggested to me. the view of the wires 

 of a transit instrument, with a collimating eye-piece, as reflected in a mercury 

 trough,— an observation, the accuracy of which may. he states, be brought within 

 ,*j of a second. 



VOL. II. 



T 



