S^6 APPENDIX. 



magnetic action in consequence of a ship's head being for a consider- 

 able time towards the east, or west : yet this is but a conjecture. In 

 the measures between Bahia and Rio de Janeiro, and in those between 

 Rio de Janeiro and Cape Horn, there is no evidence of any permanent 

 cause of error ; but the greater part of those measurements were 

 made with the ship's head usually near the meridian. 



Were I to select three measurements which I thought less trust- 

 worthy than others — I should decide on that from the Galapagos to 

 Otaheite, from Otahelte to New Zealand, and from Hobart Town to 

 King George Sound ; but I do not think that either one of these can 

 be five seconds of time in error, according to regular computation, 

 without supposing some unknovsm cause of error to exist. If each 

 of the three were five seconds wrong, and each error lay in the same 

 direction, still there would only be fifteen seconds out of thirty- 

 two accounted for. Such a, supposition as this, however, that each 

 of these three measurements is five seconds, or thereabouts, in error 

 (referring only to error caused by known means) appears to be ex- 

 tremely improbable, I would almost say impossible. 



It will naturally occur to the reader, that as error, undetected as 

 to locality, exists, arbitrary correction must be made in order to 

 reduce 24h. Om. 33s. to 24h. 



Otaheite has been selected as a point at which such a correction 

 might be made vdth the least degree of inconvenience : to that place 

 the longitudes in the accompanying tables are given as measured 

 westward by Cape Horn, and eastward from Greenwich by the Cape 

 of Good Hope ; and there, as the two portions of the chain overlap, 

 a mean has been taken between the resulting longitudes. 



I will now recapitulate the principal measurements, and confront 

 them with various other determinations. Limited space prevents my 

 quoting many ; but I trust that enough will be given to show that 

 some weight may be attached to at least a proportion of the results 

 obtained by the Beagle's officers. 



