OF ARTS AND SCIENCES. 299 



change of length since their original construction, hut in no case does 

 the evidence seem to me to he conclusive. It is understood that Colonel 

 Clarke finds a well-defined change in some of the standards originally 

 measured in 1842-55. The platinum meters of the Royal Society 

 present some evidences of a change of length inter se. 



According to the Fifth Report of the Standards Commission (Ap- 

 pendix), we have the Royal Society end-meter equals: — 



Royal Society line-meter —(—0.01759 mm. (Arago, 1818.) 

 " " -(-0.01881 (Kater, 1818.) 



" « -1-0.00940 (Baily, 1835.) 



" " -|-0.00837 (Standard's Office, 1869.) 



"We have here an appearance of a change. In deciding whether it is 

 a real or an apparent change, it should be remembered that in 1818 

 there was no defining cross-line on the line-meter, and that there is no 

 existing data by which the accuracy of the constants of the contact 

 pieces used with the end-meter can be estimated. 



The Russian standard of length used in the geodetic surveys pre- 

 vious to the work done by Prazmowski and Wagner presents the most 

 authentic instance of a well-defined change of length which has come 

 under my notice. This bar is made of iron, and has a length of seven 

 feet. I am not certain whether it was forged or drawn through dies. 

 Conical end-pieces of tempered steel were inserted in each end. In 

 the course of two or three years, this bar was transported a distance 

 exceeding 8,000 miles, being supported in a packing of feathers. At 

 the end of this time it was found by Prazmowski to be one thirteenth 

 of a line, or about .006 inch shorter than at the commencement of the 

 expedition, the two sets of comparisons having been made at the same 

 temperature. 



If it can be established that no permanent flexure of the bar took 

 place, we have here an authenticated instance of an actual change of 

 length. Upon the discovery of this change, from whatever cause pro- 

 duced, a new standard was constructed, also made of iron. It was 

 allowed to anneal for eight days after being forged. It is understood 

 that no change of length has ever been detected in this bar. Against 

 this somewhat doubtful evidence we have the positive evidence by 

 Chisholin, that Bronze No. 6 showed no evidence of a change in length 

 in 14 years, and of Baeyer that the precise mean length of Bessel's 

 standard bars at 13° Reaumur had not altered in the 20 years from 

 1834 to 1854. The change in the mean length is to be distinguished 

 from a change in the coeffi ient of expansion. The evidence of Baeyer 



