Apparatus f»r affording a Standard Meafure hy the Pendulum, 35 



of the commonly fuppofed meafure of 39,2 inches*. He thinks the (hortnefs of his meafure 

 a proof, that the maintaining power did not afFeft the penduljm. But, on this it may be re- 

 marlced, that this power may either accelerate or retard the time of vibration, according tocir- 

 cumftances, and that there were other modifying events that would tend to (horten his pendu- 

 lum, fuch as the unfteadinefs of his houfe, his flight frame, and the mode of fixing, as well as 

 the loofe connexion of the piece, I VV, fig.3, with that frame, the extent of his arc, and the 

 flight fpring of his wire at the place of flexure. 



Notwithllanding all the obj.eftions which may be brought againft this method of meafuring' 

 the difference of lengths of two pendulums, inftead of the direct length of one, it obvioufly 

 poflefTes confiderable advantages over this laft method ; and though the experiment requires 

 uncommon precautions before unexceptionable refults can be expected, yet, upon the general 

 confideration of the limits of error in a meafure obtained by this procefs, and that which might 

 be had by deduction from the aftronomical obfervations of angles, it appears more than proba- 

 ble, that Hatton's procefs would prove the moft accurate. This was, no doubt, a leading 

 motive with Dr. Fordyce to improve the machine of Whitehurft, and is the chief reafon why 

 I have thought it might be acceptable to the reader, to poffefs an account of fadls which are 

 only to be met with in books of confined circulation. 



One of the moft confiderable, if not the greateft impediment to accuracy, in Mr. White- 

 hurft's experiments, muft have confifted in the variations of temperature to which his apartment 

 was undoubtedly fubjedted. Dr. Fordyce amended the confl:ru£tion of the clock fo far, that 

 thefe variations became of no confequence during its adjuftments for time and the fubfequent 

 trials of rate. The wire N O, fig. 3, muft lengthen by heat and contrail by cold ; in confe- 

 quence of which, the ball O muft rife and fall, and the length of the cfFedive pendulum be- 

 tween I and V will be more changeable than that of a fimple pendulum. The remedy of 

 Dr. George Fordyce was founded on the confideration, that if the piece G G were fupported 

 upon a metallic rod or other piece of fuch a length as that the expanfion of this piece upwards, 

 by any increafe of temperature, fiiould be precifely fuch as to counterafl the expanfion of the 

 wire andpreferve the length of the efFedlive pendulum unvaried — the frame-plate, S S, would, 

 when the adjuftment was completed, occupy that ftation, with regard to thebrafs bar, when at 

 60 degrees, or any other ftandard temperature, as it would, have occupied if that temperature 

 had been permanent during the whole of the adjuftment; and, confequently, that the marks 

 drawn on the brafs while at that temperature, according to the previous diredlions, 

 would not be afFefted in their truth from changes of this kind. In fig. a, themafs 

 AAA, reprefents the wooden fupport of the clock : P, reprefents the head of the fcrew, 

 which is denoted by the fame letter in fig. 3. The fcrew, and confequently the piece 

 E G N of fig. 3, is fupported entirely by the extremity of the lever N H P. This lever, 

 the unflexibility of which is fecured by the triangular framing M L K, refts upon a brafs 

 tube I I, and is provided with a counterpoize, O, equal in weight to that of the pendulum, 

 and ferving to fecure the tube 1 I from any ftrain or flexure that might arife from the 



+ George Graham's numerous experiments, loofely quoted, without reference, hy Defaguliers, in his<39«r/> 

 of Lellura, I. 437, gave the meafure between 39,133 and 39,1*5 inches by compatifon wiih the ftandard yards 

 at Guildhall and at the Exchequer. 



Vol. III. — April 1799. F weight 



