CHAP. XII.] LATITUDE BY STARS 275 



than 2° or 3° if possible. Generally pairs within this limit can 

 be found. 



If one star of a pair be lost, it is useless taking the other, 

 unless a substitute for the one lost can be had. It is well, 

 therefore, to be provided with spare stars for pairs, as tliis 

 may often happen from clouds intervening, etc. 



Care must be taken in choosing pairs to leave sufficient 

 time between each meridian passage for the due observation 

 of each star before and after culmination. 



This will vary with the latitude and declination, as a star 

 should not be observed so far from the meridian as to bring 

 the mean of the altitudes observed less than a minute or so 

 under the meridian altitude. Fifteen minutes elapsing between 

 each passage will give plenty of time under most circum- 

 stances. 



Time must also be allowed for changing the position from 

 north to south, and vice versa, but all this will vary with the 

 quickness and experience of the observer. Beginners must be 

 satisfied with a few stars, and must allow more time. 



In preparing the ground, we must look out for a spot whence Preparing 

 we can see clear in the line of the meridian north and south, Q^ound 

 and one far enough from the beach to be beyond the distance 

 where surf will shake our quicksilver. The latter point is 

 sometimes — as, for instance, where jungle comes down to the 

 very beach — difficult to find, but it is well worth looking for, 

 and going inland a bit to get it, as otherwise good observations 

 may be rendered impossible from the vibration set up. The 

 more solid the ground the better, as it is astonishing what 

 slight causes will suffice to set the surface of the mercury in 

 motion. The use of the amalgamated trough mentioned on 

 p. 15 will, however, enable observations to be obtained when 

 impossible with the older form of horizon. 



Wind is a frequent source of quaking mercury, and care 

 should be taken to have the horizon trough firmly placed, and 

 the roof so fitted that the wind cannot get under its lower 

 edge.* 



A screen of canvas to windward is sometimes a good thing, 

 but on some ground this causes such vibration of the earth as 

 to be worse than the free blast of the wind. 



* Vide Artificial Horizon, p. 14. 



18—2 



