Dreyer — 0)1 Astronomical Transit Observations. 499 



■and in this respect the new method is not much superior to the old one, 

 except that very large errors, as Eessel's, never have been found by 

 the chronographic method. However, the personal equation or error 

 by the new method is most commonly very different from that 

 of the old one, as might be expected from the great diif erence between 

 the two methods. The following examples show this : — 



. Eye-and-Ear Method, i Chronograph Method. 



Dimkin — Hiigli Ereen, . . 



Dunkin— Henry, 



Dunkin - Todd;* 



Criswick — Lynn,t 



Weiss — Forsler, t 



Absolute personal error : Kam, 

 „ ,, ,, Hennekeler, 



„ „ ,, F. Kaiser,§ 



- 0^-14 



+ 0^-09 



-0 -17 



+ -01 



-^ -01 



+ -05 



-0 -37 



- -14 



-0 -27 



-00 



+ -16 



-0 -03 



+ -12 



- -07 



-0 -14 



- -07 



It is an advantage in the determination of absolute errors that 

 one is able to see which of the observers changes his way of estimating 

 transits, while a determination of personal differences only shows that 

 one of the two observers (or both of them) has a diiferent way of 

 observing by th^e two methods. 



Considering the importance of the question about the constancy of 

 personal eri'ors, especially in determinations of lougitude, we shall here 

 examine a case, in which this question was investigated in a very nice 

 and remarkable way, which has hitherto not, I think, been noticed suffi- 

 ciently. For the determination of the ditference of longitude between 

 Gotha and Leipzig the eye-and-ear method, as well as the chrono- 

 graphic method, was used by the two observers, Auwers andBruhas.|| 

 On eight evenings the culminations of a certain number of standard- 

 stars, observed with eye and ear, were by both observers compared 

 with the registered culminations of another group of stars, taken from 

 the same catalogue as the first stars, and the two groups were so ar- 

 ranged, that the uncertainty in the rate of the clock could have no 

 influence. The clock-corrections, found by the two groups, gave 

 a mean difference, which contained the ditference in the errors of 

 the right ascensions, the difference between the perc onal errors for 

 one observer in using the two methods, and the interval between the 

 moments in which the contact-apparatus of the pendulum in each 



* Greenmch Observations, 1854, Introd. 

 t Greenwich Observations, 1859. 



X Bestimmung der Meridiandiiferenzen, Berlin- Wien-Leipzig. Vienna, 1872. 

 § Verslagen, e. c. 2nd seiies, vol. ii., pp. 229-231. (Series C-G). 

 II P. A. Hansen: Bestimmung der Langendiflerenz zwiscben den Stern wartea 

 zu Leipzig und Gotha, ausgefiihrt von C. Brubns und A. Auwers. Leipzig, 1836. 



