VOL. LXXXV.] PHILOSOPHICAL TRANSACTIONS. 6lQ 



In order to accomplish these experiments in the most unexceptionable manner, 

 after the chain was properly stretched in the coffers, and the thermometers placed 

 by it, the whole remained till all the thermometers stood steadily at the same 

 height. The ends of the chain being then in perfect coincidence with particular 

 divisions on the brass register-heads, the chain was quickly taken out and re- 

 placed by the other, which being properly stretched in a right line, and a coin- 

 cidence made at the drawing-post end of the chain, the variation of the other 

 end from the division on its register-head showed the difFerence of the lengths 

 of the chains, which was measured by the micrometer. As it required weather 

 particularly steady to succeed in these experiments, we were obliged to catch the 

 most favourable opportunities that presented themselves, which happened on the 

 29th and 30th of July; on those days the chains were compared with each 

 other, and the results were as follow. July 29th, thermometers remaining 

 steadily at 75° during and after the operation: The chain b was found to be 6-I- 

 divisions of the micrometer-head shorter than the ghain a; and on being shifted, 

 A was found to exceed b 64- divisions. Same day, thermometers steady at 67-^°: 

 The chain b 6 divisions shorter than a; and being shifted, the chain a was 6 di- 

 visions longer than b. The mean from these experiments is, a 6^ divisions 

 longer than b. 



In the table containing the particulars of the operation it will be found, that 

 the chain b was laid aside after measuring 38 chains, on account of one of the 

 links appearing to be a little bent. Before it was sent to Mr. Ramsden it was 

 compared with the chain A, at first intended to be kept as the standard chain, 

 when it was found to be only 4J- divisions longer; which being I^ divisions less 

 than the mean 64- as found above it, shows that the chain b had lengthened I4- 

 divisions in measuring 38 chains; for when Mr. Ramsden afterwards straightened 

 the link, he could not perceive any difFerence in its length. The remainder of 

 the base was measured with the chain a, the chain b being kept as a standard, 

 and when that was completed, a comparison was again made between a and b, 

 when it appeared that a exceeded b by 14-,% divisions of the micrometer-head; 

 therefore the wear of a, by lengthening of the joints, in measuring 236 chains, 

 was 14.2—4.5 or 9.7 divisions of the micrometer. For finding the rate of ex- 

 pansion, the chain being placed in a right line, along the horizontal bottoms of 

 the coffers, and kept in a state of tension by a weight of 56 lbs., 5 thermome- 

 ters were placed close by the chain; one in the middle of each coffer; and the 

 whole was covered with a white linen cloth, when the sun shone out. After 

 remaining a few minutes, till the thermometers were nearly of the same tem- 

 perature, a perfect coincidence was made on the register-heads, at each end of 

 the chain, and the thermometers noted. Every thing remained in this state till 

 the coincidence at the weight end of the chain was observed to be altered, and 



4 K 2 



