Astrunofiiicul and Nautical Cv//eclio/is. 357 



antibus 86°, tabulce mece, ut docet (dlata comparatio, prorsus 

 cum observationibus Bradleiaais congruunt." And, in fact, the 

 mean errors, as deduced from his own computations, p. 53, are 

 these : 



We may also consider Mr. Delambre's autliority as amply 

 sufficient for the refraction at the horizon, which he makes 

 33' 46". 3, from several hundred observations, made at Bourges, 

 from 70° to 90° 20' zenith distance. Now Bessel's table gives, 

 for the horizon, about 36' 30" ; that is, 2' 44" too much. 



It may be said that these errors afford no practical objection 

 to the table, because observations are very rarely made at such 

 altitudes : but surely they are objectionable in a theoretical point 

 of view, since it is only the extreme cases that afford any test 

 of the truth of the theory : for in common cases, all theories 

 agree sufficiently well ; and in fact, Mr. Bessel's supposition, 

 that the density is so related to the height s as to be expressed 

 by e"""*, is contrary to all experience with respect to the distri- 

 bution of temperature in the atmosphere. 



I have to thank you for your Auxiliary Tables for 1823 ; but 

 I must enter my formal protest against the decided manner in 

 which you mention the differences between the declinations of 



Greenwich and of Konigsberg. 



* * * 



iii. Specimen of^h. Stock ler's Inverse Method of Limits. In 



a Letter to Charles BAEBAGE,Esq., F.R.S. 

 Dear Siu, 



I have received from Mr. Stockier the manuscript of a work 

 in Portuguese, dedicated to the Royal Society, and entitled 

 Methodo Invtrso dos Limites. I do not know that there is any 

 immediate probability of its being made public : but 1 wish to 

 ask your opinion of the degree of utility that is to be expected 



