ON COMPASS CORRECTION. 29 



Then again, when the ship's head is north-west the quadrantal error 

 is at the maximum in the same direction as when the ship's head was 

 south-east. Let the quadrantal error be corrected in any one of the 

 maximum points, and the compass will point accurately for any 

 position of the ship. Get, first a north or south, then an east or west, 

 and then any one of the four quadrantal points correct by the 

 Astronomer Royal's method, and the compass will be correct for all 

 positions of the ship. This exceedingly beautiful and simple method 

 might seem at once to settle the whole of the problem of ship's 

 magnetism and allow us to send ships to sea with confidence that the 

 compass would always point correctly. But it would be a most 

 dangerous confidence, because there are a number of points which in a 

 short plausible statement such as that I have brought before you are 

 altogether overlooked, and points which are of paramount importance 

 in the practical problem. 



First of all, can the correction be made perfect for any one place ot 

 the ship ? If the process I have indicated be carried out, and the 

 ship is sent to sea, will the compass be right within half-an-hour after 

 the correction is made ? It will not be right. It will be far wrong in 

 actual compasses and modes of adjustment to which the method has 

 been frequently applied, and is still very often applied. There is one 

 little assumption at the beginning of the Astronomer Royal's method in 

 his mathematical paper, which he makes in common with all mathema- 

 ticiansthat is, that the needles are infinitely small. But what are 

 they in reality ? From about 7| inches long in the Admiralty standard 

 compass, to 14 or 15 in some of the compasses to which this method 

 has been applied. The consequence is, as discovered by Captain 

 Evans, that, although the compass is quite correct in the north and 

 north-east and east positions, and so on, there is an error amounting, if I 

 remember rightly, sometimes to about 5 in the intervening points. 

 In one of the compasses of the " Great Eastern," for instance, Captain 

 Evans found in turning the ship round to all points, that, although the 

 quadrantal error had been corrected perfectly enough for practical 

 purposes, for the north-east point there were errors in the intermediate 

 points having a maximum part of the way round from the earth 

 in one direction, and another error when another $. part of the way 

 round. These errors he called octantal errors. In conjunction with 



