Page 241 hydrography 3384 



A series of 18 sets of star sights by 3 observers was analysed in a different way. 

 There were four or more star sights in each set. All were observed from the vessel at 

 anchor. The probable error of the resulting position was 0.07 nautical mile and the 

 probable error of the position determined by any one set was 0.29 nautical mile. It is 

 to be noted that the systematic error in longitude discussed in 3384B is not accounted 

 for in this determination of probable errors. 



A. ERRORS OF OBSERVATIONS 



There are certain errors in astronomic sights that are constant or symmetrical, or 

 nearly so, and which will affect all observations of a series alike. These are partly 

 eliminated by the method prescribed for a series of morning or evening star sights (see 

 3387A). Among these are index, refraction, and personal equation errors. A small 

 error in the index correction or an incorrect determination of the height of eye affects 

 each line of position an ec^ual amount and only alters the size of the figure of error formed 

 by the plotted lines of position, provided a proper selection of celestial bodies has been 

 made. The same is partly true of refraction if it is of abnormal amount in all direc- 

 tions. Systematic personal equation errors likewise are more or less canceled out. A 

 good sextant with an accurately graduated arc should always be used for astronomic 

 observations, but systematic errors of graduation or errors caused by sextant mirrors 

 will be largely compensatory when celestial bodies of nearly the same altitude are used. 

 Errors due to nonverticality of the sextant, likewise, will more or less cancel out when 

 bodies of approximately the same altitude are observed. 



Accidental errors of observation, however, are not eliminated by observing a series 

 of stars symmetrically arranged, and only observations made by experienced officers 

 who are able to obtain rapid and consistently good results should be used. Inexperi- 

 enced officers should be given every opportunity to take and compute sights for practice, 

 but their results should not be used officially until they are known to be trustworthy. 



Regardless of the fact that an error in the index correction or height of eye is 

 eliminated by observations on stars symmetrically arranged, both should always be 

 carefully measured because the observation of stars in all directions may be prevented. 

 In addition to the errors which may exist in an individual star sight, another kind of 

 error may exist in a sim sight. This is the error caused by imperfect shade glasses, 

 described in 4515(f) to which reference should be made. 



It must be recognized then that a position determined from a series of stellar 

 observations taken under the best conditions is likely to be far more accurate than a 

 position on a line adjusted to individual sun sights observed under the best conditions 

 during the day. 



B. ERROR IN LONGITUDE 



All other conditions being equal, there is likely to be a greater error in an astro- 

 nomic sight on a celestial body in an east or west direction than on one in the north or 

 south. The reason for this is twofold: An error in time will affect the line of position 

 proportionately more, the more nearly the body bears east or west; and celestial bodies 

 change proportionately faster in altitude, the more nearly they bear east and west, 

 making accurate observations more difficult. 



The effect of an error in time on a line of position is discussed in 3384; there may 

 be an error in the chronometer time, in the watch time, or the recorder may have a 

 personal error'. 



4-'o:lS2— 4-1 IT 



