TERRESTRIAL MAGNETISM, ETC. 301 



frictional error of the new compass when tested 

 in the most severe manner that is to say, by 

 experiments on shore with the bowl resting on a 

 perfectly steady support, first bringing a magnet 

 near it so as to deflect the card several degrees, 

 and then withdrawing the magnet so as to allow 

 it to come back very slowly towards its true 

 position of magnetic equilibrium is not more 

 than a quarter of a degree. The whole magnetic 

 moment of the eight needles of the new card is 

 only about one-thirteenth of that of the two 

 needles of an ordinary lo-inch card, and is so 

 small that the error due to its inductive influence 

 on the iron globes used for correcting the quad- 

 rantal error is practically insensible, even in such 

 extreme cases as when the quadrantal error 

 corrected amounts to 10 or 15. The theoretical 

 anticipation of advantage from the long vibra 

 tional period in giving steadiness at sea, has 

 been fully confirmed by three years' experience 

 in iron sailing ships and steamers, some 

 crossing the Atlantic, and others making voyages 

 through the Mediterranean and round the Cape 



