622 PHILOSOPHICAL TRANSACTIONS. [aNNO 1795. 



To compare this length of the base with that assigned by General Roy, it be- 

 comes necessary to rectify a small oversight in the 4th step of the process pub- 

 lished in the Phil. Trans, for 1785. The equation for 6° difference of tempe- 

 rature there specified, should consist of the difference of the numbers for brass 

 and glass, and not of that for brass alone, viz.— X (3.38938 — 1.41638) = 

 0.9864 feet, instead of 1 .6946, which makes the base 0.7082 feet too long. 

 Therefore the length of the base, as measured by the glass rods, is 27404.0843 

 feet, being only about 24- inches less than by the above reduction ; consequently 

 27404.2, the mean of the 2 results, may be taken as the true length of the base. 



Mr. Ramsden's method of ascertaining the actual lengths of the chains a and 

 B, was thus. These chains were originally compared with the brass points in- 

 serted in the stone coping of the wall of St. James's church-yard; but the tem- 

 perature at the time of that comparison was afterwards forgotten by Mr. Rams- 

 den. After the mensuration on Hounslow-heath was finished, the chains were 

 again compared with those points; but the result did not prove to be satisfactory, 

 as there were reasons for supposing that some alteration had taken place in the 

 length of the coping; but, independent of this, the great irregularities between 

 the joints of the stones, some of which projected half an inch above others, 

 rendered it at best a very rude and inaccurate operation. Mr. Ramsden had 

 points remaining on his great plank, which had been transferred from the brass 

 standard ; but as the plank itself was found to be subject to a daily expansion and 

 contraction, he turned his thoughts to the invention of some other method of 

 measuring the lengths of the chains, in a more unexceptionable manner. 



On considering that the expansion of cast-iron is nearly the same as that of 

 the steel chain, he procured a prismatic bar of that metal, of 21 feet long, 

 judging it to be the most proper material for the present occasion, as well as for 

 establishing a permanent standard for future comparisons of the same kind. The 

 great plank was cut to the length of about 22 feet, and on one of its narrow 

 edges 21 brackets were fixed; each of which had a triangular notch to receive 

 and support the bar, with one of its angles downwards, so that the upper sur- 

 face became one of the faces of the prism. Before the brass points were inserted 

 in this bar, Mr. Ramsden compared his brass standard with that belonging to 

 the R. s. for which purpose, on Nov. 22, 1791j it was sent to their apartments 

 in Somerset-[)lace, where, after the 2 standards had remained together about 24 

 hours, they were found to be precisely of the same length. Brass points were 

 then inserted in the upper surface of the bar, from Mr. Ramsden's standard, at 

 the distance of 40 inches from each other, the whole length of 20 feet being laid 

 off on those points in the temperature of 54°. 



The chains were measured in the Duke of Marlborough's riding-house, where 

 the light was very convenient for the purpose, and the whole apparatus wasshel- 



