380 Astronomical and Nautical Collections. 



it is scarcely probable that this law would agree well enough 

 with the results of observation, to make it necessary to in- 

 vestigate it here in a general manner. It is, however, obvious, 

 that when the compass is in the plane of the magnetic equator 

 of the sphere, the direction of the needle, as influenced by it, 

 will be parallel to the magnetic axis of the sphere, and conse- 

 quently in the magnetic meridian of the earth, so that the distur- 

 bance will disappear, as it did in Mr. Barlow's experiments ; and 

 this circumstance, if it were ascertained by observation, would 

 assist us in determining the place of the supposed sphere in the 

 ship. But in the case here stated as an example of the situation 

 of the sphere, the disturbance would never vanish, unless the dip 

 were less than 54° 44': the cosine of the angle formed by the 

 ship with the meridian, when the force vanishes, being V | the 

 tangent of the dip : and this would happen first, in the northern 

 hemisphere, when the ship's head pointed nearly south, while 

 in the situation diametrically opposite, the disturbance would 

 by no means vanish : so that the supposition of an induced 

 magnetism, like that of a sphere, does not appear to be con- 

 sistent with actual observation. Nor is it possible that a sphere 

 should be so placed as to cause no disturbance whatever at the 

 magnetic equator : and if the disturbance really vanishes at 

 the equator, as has been asserted, it can only arise from an 

 effect resembHng that of the induced magnetism of a vertical bar. 



iv. " Errors of the Nautical Almanac" for 1822. 



Mr. Schumacher, Astronomer Royal of Copenhagen, having 

 had occasion to calculate the moon's place throughout the year 

 1822 for Greenwich time, in a cursory manner, he has com- 

 pared his results with the Nautical Almanac for the year. Out 

 of more than 1400 results, 11 only exhibit a difference of 2"; 

 but on the 2d Nov. N. and M. the moon's longitude appears to 

 be put down 4" too little in the Nautical Almanac, a difference 

 equivalent to an error of 2', or 2 miles of longitude. The errors 

 of 2" may be attributed to either series of calculations with 

 equal probability. 



