72 THE TRANSATLANTIC LONGITUDE. 



of the observers at Newfoundland failed of accomplishment is attended by a mini- 

 mum of embarrassment, since the equation between Messrs. Dean and Goodfellow 

 has varied between very narrow limits, on the two sides of nothing, for a number 

 of years. 



It was found impracticable to make the arrangements for the series of personal 

 comparisons, without fitting up a small building specially for the purpose, which 

 could not be accomplished till the middle of April, on account of the snow and 

 various delays. On the 9th of April the comparisons commenced, and were con- 

 tinued on every occasion that the extremely unfavorable weather permitted, until 

 sixteen comparisons had been made between eight pairs of observers ; four of the 

 six observers comparing each with three others, and two of them each with two 

 others. It was provided that a single comparison should depend upon not less than 

 ten pairs of stars, ten transits over twenty-five threads being thus observed by each 

 person, and that no person should take part in more than one comparison on the 

 same night, lest the results be affected by his fatigue. 



The results of these comparisons, together with their mean errors (stars between 

 25° and 50° being almost exclusively used), are as follows: — 



Gould — Dean 



= 



4-0^427 



± 0^034 



April 13 







-f- 0.380 



± 



0-026 



18 



Gould — Mosman 



= 



-f- 0.472 



± 



0.028 



May 23 







+ 0.459 



± 



0.070 



28 



Gould — Chandler 



= 



-1- 0.190 





0.037 



June 1 







+ 0.202 





0.033 



19 



Dean — Goodfellow 



= 



— 0.013 





0.023 



April 9 







— 0.008 





0.024 



11 



Dean — Mosman 



= 



+ 0,109 



± 



0.014 



19 







+ 0.094 





0.024 



23 



Boutelle — Goodfellow 



= 



— 0.134 





0.029 



19 







— 0.146 





0.029 



23 



Boutelle — Chandler 



= 



— 0.147 





0.028 



11 



Goodfellow — Chandler 



= 



— 0.021 





0.032 



13 







— 0.072 



± 



0.026 



April 18 



Farther comparisons between Messrs. Boutelle, Mosman and Chandler, were con- 

 templated, but were prevented by duties which called two of these gentlemen 

 away, before farther observations could be obtained. One comparison between 

 Mr. Chandler and myself was rejected for manifest error, on a night when the stars 

 were only visible between rapidly flying clouds. 



Assigning to these several determinations their appropriate weights and equating, 

 we arrive at the following values — 



Goukl- 



—Dean = 



4- 



0'.303 



Gould- 



—Mosman = 



+ 



0-454 



Gould- 



—Chandler = 



+ 



0.216 



Dean— 



-Goodfellow = 



— 



0.029 



Dean— 



-Mosman = 



+ 



0.121 



