550 MAGNETICAL OBSERVATIONS. [aPP. N<> IX. 



Captain Sabine as follows : —The error produced in 

 any direction of the ship's head, will be to the error at 

 the point of greatest irregularity , [instead of east and 

 west], as the sine of the angle between the ship's head 

 and the points of no error, [instead of the magnetic 

 meridian], to the sine of eight points or radius, 

 (c.) To find the greatest anomaly and the point of change, 

 or no anomaly, the following rule appears one of the 

 most simple. Take the bearing of the sun, or any 

 other distant object whose true bearing from the ship 

 is known, with the binnacle compass when the ship's 

 head is put upon each rhomb; if they are all the same, 

 there is no anomaly ; but their differences from the 

 true bearing, after the application of the variation, 

 (which should be determined out of the vessel), 

 Avill give the anomalies on the respective courses ; 

 and the points on which there are no differences, 

 ^vill be the points of change. When the sun is 

 the object observed, it is only necessary to take an al- 

 titude, along with two or three of the bearings, and 

 the intervals of time, by an easy approximation, will 

 give the true bearings for the intermediate observa- 

 tions. When the sun is near the meridian, and par- 

 ticularly in high latitudes in summer, an allowance of 

 a degree of azimuth for 4 minutes of time, will be 

 sufficiently accurate for intervals of an hour. "The 

 true bearing of a distant fixed object may be found, in 

 a perfectly smooth sea, by sending a copper-fastened 

 boat in the direction of the object, for, when exactly in 

 the required line, the true bearing of the object from 

 the boat will give the true bearing from the ship. 

 When among ice, or near shore, the same result may 

 be obtained more accurately, by sending an observer 

 to some distance in the required line, for taking the 

 bearing of the object. The transit bearings of two 

 objects, of which the position with regard to one an- 

 other is known, is preferable perhaps to any thing else. 



