020 VHILOSOPHICAL TKANSACTIONS. TaNNO IJQd. 



the thtTinometers nearly the same; at which instant they were again read off 

 and the alteration of coincidence measured by the micrometer. After the ob- 

 servation of g experiments in this way, on several different days, the mean result 

 of these Q experiments was 0.007402, or 0.0075 inch to 1° of Fahrenheit, on 

 100 feet of blistered steel; which differs only , „^g„„ th parts of an inch from 

 General Roy's conclusion with the pyrometer; but the number .0075 is preferred in 

 these measurements, as being deduced from experiments made with the chain itself. 

 After the chains were compared, and the rate of expansion determined, as re- 

 lated in tlie preceding article, several trials were made of arranging the pickets 

 and coffers in such a manner as might be supposed proper for the reception of 

 the chain. It was soon found however, that this method of measuring would 

 be neither so expeditious nor accurate, as if the coffers were placed on tressels 

 such as were used by General Roy in his measurement witli the glass rods. An 

 application was therefore made to Sir Joseph Banks, who very obligingly com- 

 plied with the request, and lent the tressels belonging to the r. s.; a description of 

 which may be seen in the 75th vol. of the Philos. Trans. As the upper part of 

 the pipe at the north-west end of the base was found to be exceedingly rotten, 

 it became necessary to saw off J 3 inches of it, which left enough of the cylinder 

 remaining to fix the brass cup in, as it had been originally bored to the depth of 

 2 feet. This cup, which was also lent by the k. s, being inserted in the pipe, 

 fitted it exactly. 



On the 15th of August, having previously traced out the line of the base, by 

 means of the transit-instrument, the operation commenced, in the presence of 

 Sir Joseph Banks, Dr. Maskelyne, and several other members of the r. s. A 

 table is then inserted, which contains the particulars of it, and will explain the order 

 of time in which the different parts of the measurement was performed. As it 

 would swell this table to a great extent, were the degrees shown by the ther- 

 mometers inserted in it, it has been considered as proper to give only their sum, 

 which is sufficient for finding the correction to be applied in the reduction of the 

 base, on account of the lengthening or contracting of the chain by variation of 

 temperature. It may however be remarked, that the 5 thermometers were laid 

 close by the chain, and suffered to remain till they had nearly the same tempera- 

 ture, when they were read off, and registered in a field-book, while an observer 

 at each end of the chain preserved a perfect coincidence between the arrow and 

 a particular division on the brass scale. When the sun shone out, the chain was 

 covered with a white linen cloth, the ends of which were put over the openings 

 of the first and last coffers, to exclude the circulation of air. The thermome- 

 ters usually remained in the coffers from 7 to 15 minutes, according to circum- 

 stances; when the sky was much overcast, a shorter time generally was found to 

 be sufficient. 



