481 



or to causes which affect the changes from one hour to the next in a 

 perceptible manner, but to gradual and regular daily change. If we 

 examine the daily means most free from irregular or intermittent 

 disturbance, we shall find that they vary plus or minus of the monthly 

 mean ; if the difference amounts to I in any case, then the whole 

 observations of the day may be rejected though they follow the nor- 

 mal law. By taking a proper value of I this case may not happen fre- 

 quently,but cases like the following will. At Hobarton the daily means 

 of magnetic declination differ in some months from the monthly means 

 by 2'*0 nearly ; as the limit chosen by General Sabine is 2''4, any 

 observation in such days differing by 0'*4 from the normal mean 

 would be rejected. The 25th and 26th days of March 1844 had 

 been chosen by me as days free from magnetic disturbance, and fol- 

 lowing the normal law at Makerstoun (Mak. Obs. 1844, p. 339), 

 yet the means of horizontal force for these days differed 0*00064 and 

 0*00075 from the monthly means ; had the former quantity been 

 the limit, all the observations on these days might have been rejected. 



Altogether it appears to me that the method of rejecting observa- 

 tions beyond certain limits should not be employed at all, or if 

 employed, only when interpolated observations are substituted ; and 

 that this interpolation should constitute a second part of the discus- 

 sion, the first including all the observations*. 



These considerations may appear somewhat elementary, but it is 

 essential that results which present such anomalies as the lunar 

 diurnal variation of magnetic declination should be obtained in a 

 manner the most free from objection, even though the objections 

 should touch on quantities of a second order compared with those 

 obtained. 



The discussion of which I now proceed to note the results, 

 includes all the hourly observations without exception, made in the 

 Trevandrum Observatory (within a degree and a half of the magnetic 

 equator) during the five years 1854 to 1858 ; the second part of the 

 discussion, in which days of great magnetic irregularity have been 



* I should note here my belief that a peculiarity noticed by General Sabine in 

 his discussions as requiring explanation, namely, that the excursions of the decli- 

 nation needle east and west in the lunar diurnal variation have very different 

 magnitudes, is due to the rejection of observations, while the means by which the 

 differences were obtained included the rejected quantities. 



2L2 



