On the prefent Stale of Aerojlation^ 51 



POSTSCRIPT. 



So long as the vibration of pendulums is performed in a 

 medium of varying denfity, we niuft not look for an accurate 

 time-piece for afcertaining the longitude, &c. ; unlefs a felt- 

 corrci5lino- mercurial pendulum could be contrived, adapted 

 to counteraiSl the fmalleft variations effe£ted by the ambient 

 air. The errors of a time-piece are but half correded by 

 the fabrication of pendulums adapted to obviate the expanfioa 

 of metals by increafe of temperature, if the works themfelvcs 

 Itill remain conftrucled of fuch expanfible materials. A cor- 

 real time-piece, therefore, will be that of which not only the 

 works and pendulum are conllrufted of the leaft e.xpanfible 

 materials, but the pendulum itfelf (liall vibrate in a medium 

 of unalterable denfity ; a delideratum only to be obtained by 

 caufing the vibrations to be performed in vacuo, or by a fell- 

 correflino- pendulum, as above alluded to. I know of no 

 fubftance fo well adapted to compofe the works of fuch a 

 machine as ivorv, or the horn of the narwhal or fea unicorn 

 (nearly entirely compofed of enamel, and ufed in the fabrica- 

 tion of artificial teeth), efpecially if we had a means of In- 

 creafino- their hardnefs fo as to vie with the metals (a fubjeil 

 on which I ])ropofe to make fome experiments) : but for the 

 pendulum itfelf, the common gridiron pendulum is allowed 

 to be the moil accurate in ufe, provided its vibrations were 

 not obftruftcd by unequal refiftance from the air: but for a 

 mercurial pendulum to move in open fpace, glafs will be in- 

 dicated as the leaft expanfible and moll proper fubftance for 

 this purpofe. I add, that the moft perfe6t time-piece hither- 

 to conftrucled, has been fo only in as far as the works have 

 been moft exaftly accurate, and the expanfion of the pen- 

 dulum been counteracted (as in the gridiron pendulum) by 

 oppofmg expanfion to expanfion. The exaft concordance of 

 fuch a machine, on its return to the firft placeof o' ''ervation, 

 will have depended on the nature of the voyage performed in 

 the interim. As, if fuch voyage has been performed acrofs 

 the equator and back (as to the Eaft Indies), the aggregate 

 number of anomalies in the pendulum in the voyage out, 

 will in the return be nearly balanced by the number of ano- 

 malies in the voyage home, io that the whole difference (hall 

 at laft, perhajis,'not amount to 100". But if (fome weeks 

 after the failing of the velfei) we had it in our power to know 

 at a certain inftant the precife moment indicated by an equally 

 accurate chronometer left at the place of departure, we (liould 

 find a greater variation i)ct\veen the two, owing to the difier- 

 cnce in the denfities of the media in which the pendulum at 



home 



