CHAP. XII.] LATITUDE BY STARS 273 



There remain the other errors, some of which may be directly 

 allowed for, but only approximately ; others cannot be cor- 

 rected at all, and the latitude resulting from observations of 

 a single body — as, e.g., the sun — will be therefore always more 

 or less in error. 



The only way satisfactorily to clear these errors is to observe Pairs of 

 stars, in pairs, of as equal altitude as can be found, one. north, ^*^" 

 and one south, of the zenith. These errors will then act in 

 opposite directions, as everything tending to increase or 

 diminish the altitude on one side of the zenith will act simi- 

 larly on the other : but, in working out the latitude, tie 

 resulting error will increase the latitude in one case and 

 decrease it in the other, so that the mean of the latitudes 

 obtained by each star of such a pair \y\\\ approximate very 

 closely to the correct one. 



To eliminate the artificial horizon roof error when observing 

 pairs of stars, the roof must always be in the same position 

 with respect to the observer, and therefore must be reversed 

 when changing from face north to face south, and vice versa. 

 If observing a single object, as the sun, the roof must be 

 reversed Avhen half-way through the observation. 



The use of a sextant stand, when once the observer has got Sextant 

 thoroughly accustomed to it, is an immense assistance to good s*^^^- 

 observations, as the images of the stars, instead of quavering 

 and shaking with every slight motion of the hand of the 

 observer, remain perfectly still, and can be made to pass over 

 one another with great accuracy. 



Certain preparations are necessary for good star observa- Prepara- 

 tions, for all scurry that can be avoided should be. *^°°^ 



' "^ necessary. 



In the first place, stars must be selected and arranged for 

 observations according to their pairs. 



If stars given in the " Nautical Almanac " only are used, the 

 chances are very much against a sufficient number of pairs 

 being obtainable, as only a small proportion of observable 

 stars are there included, though the number has been lately 

 much increased. 



A surveying vessel will have the Greenwich and Cape starCata 

 Observatory Catalogues of Stars, and out of these enough pairs ^°^"®^- 

 can nearly always be picked to enable us to get a satisfactory 

 latitude in one night, including stars down to the foiu'th 



18 



