718 THE POPULAR SCIENCE MONTHLY. 



and depends on the skill of the individual steersman. Looking at 

 the compass-card, in steady weather, a course may seem very closely 

 followed ; perhaps the needle's end may not be a hundredth part of 

 an inch (on the average) from the position it should have. But a hun- 

 dredth part of an inch on the circumference of the compass-card 

 would correspond to a considerable deviation in the course of a run 

 of twenty or thirty knots ; and there is nothing to prevent the errors 

 so arising from accumulating in a long journey until a ship might be 

 thirty or forty miles from her estimated place. To this may be added 

 the circumstance that the direction of the needle is different in differ- 

 ent parts of the earth. In some places it points to the east of the 

 north, in others to the west. And, although the actual " variation of 

 the compass," as this peculiarity is called, is known in a general way 

 for all parts of the earth, yet such knowledge has no claim to actual 

 exactness. There is also an important danger, as recent instances 

 have shown, in the possible change of the position of the ship's com- 

 pass, on account of iron in her cargo. 



But a far more important cause of error, in determinations merely 

 depending on the log-book, is that arising from uncertainty as to the 

 ship's rate of progress. The log-line gives only a rough idea of the 

 ship's rate at the time when the log is cast ; ' and, of course, a ship's 

 rate does not remain constant, even when she is under steam alone. 

 Then, again, currents carry the ship along sometimes with consider- 

 able rapidity ; and the log-line affords no indication of their action : 

 while no reliance can be placed on the estimated rates, even of known 

 currents. Thus the distance made on any course may differ consider- 

 ably from the estimated distance ; and, when several days' sailing are 

 dealt with, an error of large amount may readily accumulate. 



For these and other reasons, a ship's captain places little reliance 

 on what is called " the day's work " that is, the change in the ship's 

 position from noon to noon as estimated from the compass-courses en- 

 tered in the log-book, and the distances supposed to be run on these 

 courses. It is absolutely essential that such estimates should be careful- 

 ly made, because, under favorable conditions of weather, there may be 

 no other means of guessing at the ship's position. But the only really 

 reliable way of determining a ship's place is by astronomical observa- 

 tions. It is on this account that the almanac published by the Ad- 

 miralty, in which the position and apparent motions of the celestial 

 bodies are indicated, four or five years in advance, is called, par excel- 



1 The log is a flat piece of wood of quadrantal shape, so loaded at the rim as to float 

 with the point (that is, the centre of the quadrant) uppermost. To this a line about 

 300 yards long is fastened. The log is thrown overboard, and comes almost immediately 

 to rest on the surface of the sea, the line being suffered to run freely out. By marks on 

 the log-line divided into equal spaces, called knots, of known length, and by observing 

 how many of these run out, while the sand in a half-minute hour-glass is running, the 

 ship's rate of motion is roughly inferred. The whole process is necessarily rough, since 

 the line cannot even be straightened. 



