MAGNETIC DECLINATION AT TREVANDEUM. 757 



for sunrise, but not so well marked. The curves, Plate XXVII., for the moon's 

 phase seem to show that the turning point occurs generally within two hours of 

 sunset, but these are derived from the superposition of successive days, and are, 

 in consequence, less fitted to show the facts exactly than the curves from the 

 single days under consideration. 



It is instructive to remark the differences of the results from the various cal- 

 culations. In the last few single days we see how the needle behaves under 

 the influence of the moon on different meridians, in different circumstances of 

 day and night, of sunrise and sunset. In the former, we have results which 

 always diverge more from the true movements, the greater the number of 

 successive days or months which are included in the calculations. 





