Invariable Pendulum^ during a Russian Scientific Voyage. 421 



gression : but I shall content myself, for the present, by taking 

 the mean of the two results. 



Tile other stations were Valparaiso, Sitka, Petropaulouski ; 

 the islands of Ualan, Guam, Bonine and St. Helena; and 

 lastly, the Observatory of St. Petersburgh. Here the experi- 

 ments had for their object, as much the length of the pendu- 

 lum, as the change of length from temperature. The two 

 series (one of which was made at the mean temperature of 

 31°-5 Fahr., and the other at 82°"5) showed a difference of 

 O'^SS vibration in a mean solar day, for each degree of the 

 scale. This result is 0*0"3 greater than that found by Capt. 

 Sabine, from a similar process; although the two pendulums 

 were made of the same kind of metal (bell-metal*), and were 

 nearly of the same dimensions. But, I do not find, in our ex- 

 periments, any thing that should cause this difference, unless 

 it be the smaller density of the metal of which our pendulum 

 is composed. 



In voyages of this kind, the stay at each port is generally 

 very short; and the time which we can give to each species 

 of observation naturally very limited. It is evident therefore 

 that we depend much on accidental circumstances, which may 

 influence the success of our labours very considerably. Hence 

 it happens that, although we have always paid the same at- 

 tention to every thing that could contribute to the accuracy 

 of the experiments, yet they are not all of the same value. 

 Those which deserve the greatest confidence are those that 

 were made at Greenwich, St. Petersburgh, Petropaulouski, 

 Valparaiso, and the Bonine isles. At these five stations, I 

 think I can answer for y^th of a vibration. Then come Sitka, 

 and the island of Ualan ; where the mean result may be un- 

 certain to jth of a vibration. The experiments, which are the 

 least to be depended upon, are those made at the islands of 

 Guam and St. Helena ; where I do not pretend to a precision 

 greater than i a vibration. 



The latitudes were determined by circum-meridian alti- 

 tudes oi the sun and stars, observed with a sextant and a cir- 

 cle, each by Troughton, and with a reflecting repeating circle 

 by Dollond. We had not an astronomical rejieating circle : 

 but, we endeavoured, by multiplying the observations, and 

 varying the circumstances, to make the above-mentioned in- 

 struments serve the same purpose. 



I now conic to the results; which are contained in the fol- 

 lowing Table: where the "ird colunni contains the number of 



* We suspect that ("apt. Liictkc is wrong in designating the pendulum 

 as made of /jcli-7)uifi/; as \vc believe they are all iu:\dco( /jiass, — Edit. 



