the end of the wire aiul artovwiuds lironirlit in conl:u-t with tlic li;ill whih' tlic 

 latter was heateil. A numlier of dirt'oiciit siispeiitliivj; wires weio tried and thai 

 at last made use of was of [datinuin .30 mm. in diameter whicli was of sufficient 

 strength to insure, after a few (hiy"s sns|)eusi()n, a [lenduhun i)f invariahle length 

 during the time of the experiment, suliject, of course, to slight changes due to 

 variation in temperature and the low co-efBeient of expansion of platinum reduces 

 these clu'.nges to a minimum. The knife-edge of the pendulum, which was of 

 steel, rested on a pair uf agate plates which were iirndy secured in a |)latc of 

 bmss and which were accurately leveled hy means of four leveling screws, after 

 wliich the plate was firmly clamped to the ii^n liar upon which it rested S" as to 

 jirevent lateral motion and render thc'supiiort as rigid as po.s.'-iiile. 



The measuring a[>paratus was also liy Salleron and was of the form used hy 

 Borda. The rod was of iron atid its length was read hy means of a xcrnier and 

 microscope to hundredtlis of a millimetre. There was also a nielallie tlnrnid- 

 meter attached for giving the temiierature of the bar. 



A strong plat-form was firmly secured to the stone pier immediately helow 

 the lower end of the pendulum and uimu this was placed the small circular plane 

 talde which, in the process of making a measurement, was elevated hy means of 

 a screw until it was tangent to the lower surface of the i>endiduni hall. When 

 this was done the pendulum was removed from its place and the nicasin'ing rod 

 substituted. The deflection of the pendulum su|ipiirt, due to the excess of the 

 weight of the measming rod over that of the pendulum itself was measured and 

 found to he .02 mm. and this correction was ajiplied to the indicated length 

 as was, also, the projwr correction for temperature. For the purpose of verify- 

 ing the length of the rod we were enabled, through the kindness of the officials 

 of tlie Im[ierial Treasury, to compare it with a standard metre hy l>eleuil in the 

 jiosscssion of that Department. The result of this c<imparison was that a correc- 

 tion of .04 mm. was made ui^on the length of the measuring rod at 0'\ — Very 

 recently another standard has been received l)y the same Department, wliich is 

 certified to Ijc a cojyy of the standard at the (Jonser\afoire des Arts et Metiers 

 and the nece.s.sary corrections f(n' it have been furnished. The previous standard 

 hiis been carefully cojnpared with this and the agreement is so clo.se as not (o 

 di^niand any further correction for the pendulum metre. In all the oliservations 

 the teni|)("ratnre was n-conled as read from a thermometer hanging very near to 

 the middle point ol'the [lendulum and IIk^ mefallie I herniometor connected with 

 tho mt^asnring rod was also read and i-ecorded at each measurement of length. 



As the apjiaratiiH was arranged it was very eawy to make a measurement of 

 the length of the |H^ndulum and this was done at very short intervals, always 

 both Ixjfore and after a s<!ries of vibratioTis. Ft was found, howe\er, that w lien 

 the temperature was constant the lengfli riniaini<l sensibly the same. 



The an: <if vibration was mi-asured Ity means of a scale |ilaced immediately 

 Ix-hind the 8Us|K'nding wire ami a telesco|M' jilaced aliout live metres awav. Tiie 

 nvjan arcs of vibration varied in the dilferent experiments from 40 to 70 . 



