74 THE TRANSATLANTIC LONGITUDE. 



instrument of the Coast Survey on four nights immediately after the close of the 

 comparisons just described — using delicately adjusted keys, to which both of us were 

 accustomed — gave as the difference between Mr. Chandler and myself 



Gould— Chandler = —0^021, 

 instead of + 0.216 as above; while the difference between Mr. Boutelle and myself, 

 as measured in past years, has rarely attained the limit of 0'.2. 



The comparisons between Messrs. Dean and Mosman seem to have been similarly, 

 although not equally affected by the same cause ; and I have thus been led to the 

 conviction that but little, if any, weight ought to be assigned to these determina- 

 tions of personal equation, as regards their application to the clock-errors, from 

 which the longitude must be deduced. If farther argument were needed, it would 

 only be necessary to apply to the series of preliminary results already deduced in 

 Chapters VII. and VIII., the values of personal difference here obtained. The 

 accordance, now so satisfactory, would be entirely destroyed; and the probable 

 error of the result increased more than tenfold, for each of the two longitudes. 



The difference here found between Messrs. Dean and GoodfeUow is the only 

 satisfactory one. These gentlemen have been accustomed to observe in connection 

 with one another for ten or twelve years ; and a very extensive series of measure- 

 ments, both by observations specially made for the purpose, and by the comparison 

 of longitude-results deduced from their observations before and after exchanging 

 stations, shoAvs that their personal difference has usually scarcely exceeded the 

 limits of probable error, while it has varied in sign, as already stated. 



A satisfactory explanation of the phenomenon is, I think, to be found in the 

 break-circuit keys employed, of which the springs were so strong as to prompt a 

 memorandum on each date when I observed, "to the effect that my observations 

 were embarrassed by the strong tension of the keys, which were those used at New- 

 foundland. Many of my observations were lost in this way at the commencement 

 of the work, and my first night's comparisons proved futile for this reason ; inas- 

 much as the greater proportion of my signal-taps were found not to have been re- 

 corded at all upon the chronograph, which was in another building, some twenty- 

 five rods distant. My pressure upon the button had not been forcible enough to break 

 the contact. Mr. Boutelle also complained of the stiffness of the observing key, 

 and caused a note to this effect to be entered upon the journal of the observations 

 for personal equation. 



Under these embarrassing circumstances only two courses seem to be available. 

 A repetition of the comparisons, using more delicate signal-keys, would have been 

 highly desirable, and was earnestly hoped for; but, apart from the other serious 

 obstacles, the assignment of the various observers to other duties, some of them at 

 very remote stations, precluded all possibility of this solution of the difficulty. 

 We may however totally discard all consideration of the personal equation, except 

 the value between Dean and GoodfeUow, which latter may be regarded as so small 

 and well established as to reduce nearly to a minimum the effects of the misappre- 

 hension by which the time-determinations, at Calais, for the two steps in the longi- 

 tude, were made by different persons ; or, on the other hand, we may fix upon ap- 

 proximate values, by considering the tolerably accordant determinations made at 



