of the Acceleration of Gravity for Tokio, Japan, 47 



were using, was by accident omitted in our final calculation ; 

 and we are glad that Major Herschel's having charged us with 

 neglecting the corrections for arc, for buoyancy, and for re- 

 sistance, all of which we had attended to, has enabled us to 

 add another correction which had been omitted. Applying 

 this latter to the data we have given, it is easily seen that our 

 experiments lead to a value of g in Tokio equal to 97 9 '82 

 centims. per second per second. 



Not only is this correction extremely difficult to apply, with 

 any degree of accuracy, to a Kater's pendulum, but it is differ- 

 ent for the two axes of suspension. It would, however, like 

 all the other corrections, be of a nature to make the real value 

 of g greater than the apparent ; so that its employment or non- 

 employment will not explain why our results with the rever- 

 sible pendulums were unsatisfactory. 



To avoid having to apply this air-inertia correction, ifc is of 

 course not unusual to swing the Kater's pendulum in a va- 

 cuum ; but even then the correction for viscosity will not be 

 diminished, since it has been shown by the late Prof. Clerk 

 Maxwell and by Mr. Crookes that it is not until the very high 

 vacua of a good Sprengel pump are reached that the viscosity 

 of the air is sensibly diminished. The corrections, then, ne- 

 cessary to apply to a reversible pendulum swinging in an ordi- 

 nary vacuum will be more numerous than in the case of a long 

 wire pendulum like ours swinging in the air, not to mention 

 the far greater facility with which experiments can be made 

 with the latter kind of pendulum. 



III. We beg to thank Major Herschel for drawing our 

 attention to a misprint of X for 2 X in the formula for calcu- 

 lating g given by us. This misprint, however, which escaped 

 our notice when correcting the proof, is of but little conse- 

 quence, seeing that it was obviously the correct formula we 

 employed in the calculation itself ; for the formula as it stands 

 would lead to the value only 978*57 for the latitude 35° 39'. 

 But your correspondent says the formula itself is et wrong 

 numerically/' meaning, of course, that the numerical coeffi- 

 cients are erroneous. These coefficients, however, are, figure 

 for figure, those given by Dr. Everett on page 21 of his ' Units 

 and Physical Constants,' published last year. Now, although 

 the latest pendulum-experiments make it probable that the 

 numerical coefficients mentioned by Major Herschel are still 

 more accurate than those given in the book we have quoted, 

 he is a little unfortunate in taking this occasion to object to 

 Dr. Everett's formula, seeing that the latitude of Tokio is 

 just that which gives to cos 2 X such a value as to make the two 

 formulae yield identical results, or at least results differing by 



