380 Roijal Society, 



which the pendulum might be composed. But M. Besscl has shown 

 that a quantity of air is also set in motion by tlie jiendulum (varying 

 according to itsform and construction), and thus a compound pendulum 

 is in all cases produced, the specific gravity of which will be much 

 less than that of the metal itself. M. Bessel's principal experiments 

 for establishing the accuracy of tliis principle, were made with two 

 spheres, about two inches in diaaieter, differing from each other very 

 considerably in specific gravity, one being of brass, and the other of 

 ivory, and each suspended by a fine steel wire. The author of the 

 present paper, however, pursued another and a very different course 

 for obtaining the same end : namely, by swinging the same pendu- 

 lum first in free air, and afterwards in a highly rarefied medium, 

 nearly approaching to a vacuum. From the difference in the results, 

 he deduces a factor (denoted by n), by which the old, and hitherto 

 received, correction must be multiplied in order to obtain the new and 

 more accurate correction indicated by M. Bessel ; and which, in the 

 case of the two spheres above mentioned, is found by that author to be 

 equal to IDS. 



But Mr. Baily, instead of confining himself to spheres of this size, 

 and composed of these two substances only, has extended his inqui- 

 ries to pendulums of various magnitudes, substances and forms. His 

 first recorded experiment is on Borda's platina sphere, the diameter 

 of which is 1'44 inches ; and he found that the old correction must in 

 this case be multiplied by 1"SS in order to obtain the true and accu- 

 rate correction ; or, in other words, that the old correction was but 

 little more than half what it ought to be. The author then tried 

 three other spheres of precisely the same diameter, but differing con- 

 siderably in s])ecific gravity : namely, lead, brass, and ivory, all of 

 which gave nearly the same result; the mean of the whole being 

 n = l-8(). He next proceeded to spheres of the size used by M. Bes- 

 sel, made of three different substances, viz- lead, brass, and ivory. 

 These gave a result (agreeing very well with each other,) somewhat 

 smaller than the former; the mean of the whole being 7j=l*75: 

 thus showing that the factor for the additional correction is due to 

 the form and magnitude of the moving body, and not to its weight 

 or specific gravity. This last value, as the author observes, differs 

 from that deduced by M. Bessel as above mentioned ; but the cause 

 of the discordance, does not appear. 



The author then shows the effect produced on cylinders of various 

 kinds, both solid and hollow, and suspended in different ways, — on 

 lenses, on cylindrical rods, on bars, on tubes, on convertible pendu- 

 lums, and on several clock pendulums, amounting to upwards of 40 

 in number. The results of these experiments give in each case a 

 different value for the factor n ; and which appears to depend on tlie 

 extent of surface, in proportion to the bulk of the body exposed to 

 tlie direct action of the .lir when in motion : further experiments, 

 however, are requisite to establish this point in a satisfactory manner*. 



* Since this paper was read, tlie author has made a number of addi- 

 tional experimunts on various otlicr jiciululunis, which, by permission of 

 the Council, will loan part of the original paper ; and frcm which he is 



