462 



inconsiderable in magnitude corresponded at great distances. 

 If such should prove to be the fact, it would be only neces- 

 sary to project on a large scale the results of the observations 

 made about the same absolute time at the two stations, and 

 to compare the times of the corresponding maxima and 

 minima. In the observations already referred to, the maxima 

 and minima succeeded each other (as has been said) at inter- 

 vals of about forty seconds, and the epoch of their occurrence 

 was probably known to six or seven seconds. By shorten- 

 ing still further the interval of observation, it is manifest 

 that this error may be much diminished. The corresponding 

 error of the difference of longitude resulting from a single 

 comparison (supposing the probable error of epoch to be the 

 same at the two places,) will be greater in the ratio of \/2 

 to 1 ; but this error, owing to the multitude of the maxima 

 and minima compared, must necessarily be greatly reduced 

 in the final mean. 



In order to put this question to the test on the largest 

 scale, it was agreed between Mr. Bache and Mr. Lloyd, to 

 make a series of corresponding observations in Philadelphia 

 and in Dublin. Some difficulties occurred in concerting a 

 plan, and Prof. Bache underwent, in one instance, the labour 

 of an extensive series of observations, without any counter- 

 part in Dublin. At length, however, it was agreed to ob- 

 serve during the week commencing the 11th of November, 

 1839; the observations being taken during two hours on 

 each day — namely, from 12 to 1 p.m., and from 8 to 9 p.m., 

 Greenwich mean time. 



Prof. Bache's account of his observations is contained in 

 the following extract of a letter which accompanied them, 

 dated November 29, 1839. 



" The place of observation is a room in one of the 

 out buildings for the dwellings of the Professors of the 

 Girard College. As the materials used in the construc- 

 tion of the house must produce considerable local attrac- 



