between the Observatories of Armagh and Dublin. 141 



1 



w = * 



1+1 + 12 



s s r 



To apply this, the Dublin correction on the 14th Is derived from one star, 

 and the mean of three on the preceding, and two on the following day. I assume 

 * z= 3. 



At Armagh s' = 3. 



On the 16th, two stars, and the mean of three preceding and one following 

 give * = 3 ; *' = 5. 



On the 20th, s = s' = 7. 



On the 21st, two and the mean of seven and seven give s := 9 ; at Armagh, 

 four and the mean of seven and nine give s' = 11. 



Hence, calling the decimals of the second of a result l, we have 



May 14, w z= 0.46154 . . . wl = 0.20769 r 

 0.27273 0.08182 e 



May 16, . . 0.34091 0.10568 r 



Same 0.21818 e 



0.22059 0.17426 F 



May 20, . . 0.875 ^ 0.39375\ r 



0.77778/ 0.31040J r' 



0.82727 0.33091 E 



0.61765 0.07412 F 



* This expression of w shows, that with us the flash can be observed with about the same pre- 

 cision as the appulse of a star to a wire ; but a more important deduction may be made respecting 



the method by successive signals. As each of these adds to the denominator of w a term _ 



r 

 their number diminishes it rapidly. Thus on the 20th, if, as in the Paris and Greenwich arc, we had 



employed two intermediate stations, it would have been but 0.37 of its actual value, even supposing 



the transmission perfect. I am therefore decidedly of opinion, that stations of transmission should 



be made absolute stations, when it is possible, by furnishing them with transit instruments : this 



guards against failure, and scarcely lessens the value of the result. Thus in the case supposed, w is 



0.33, but it will be obvious that in Sir J. Herschel's operation, had this been done, instead of the 



ten complete results which he obtained, he would have got at least ninety. 



