490 EDITOR'S NOTES. 



place on opposite sides of the mean or normal values 

 had the same or different periodic laws. For this inves- 

 tigation, the hourly observations, which had been main- 

 tained with steady perseverance for several years at the 

 British Colonial Observatories, appeared likely to furnish 

 a suitable provision, if a selection were made from them 

 of a sufficient body of the particular class of phenomena 

 to enable their laws to be studied. As the magnitude 

 of deflection was the only known criterion of a dis 

 turbance of the class which had been termed "irre- 

 gular," a number of the largest deflections (reckoned 

 from the mean or normal values in the same month 

 and at the same hour) were separated from the whole 

 body of the observations for the proposed examination. 

 Care was taken that the amount of deflection, which 

 should cause an observation to be regarded as "dis- 

 turbed," should be on the one hand so low as to sepa- 

 rate a sufficient number of disturbed observations for the 

 deduction of mean effects, whilst, on the other hand, it 

 should be so high as not to include irregularities of the 

 ordinary regular and well-known variations : and that 

 the same disturbance-value should be taken throughout, 

 in order to insure a just comparison between different 

 months and different years. The observations so sepa- 

 rated were formed into tables both of the numbers and 

 of the aggregate values of the deflections (reckoned from 

 the normals) according to the years, months, and hours, 

 of their occurrence ; and the mean annual, monthly, 

 and daily numbers, and aggregate values, being severally 

 taken as units, the relative proportions of frequency 

 and of amount of disturbance, in the several years, 

 months, and hours, were obtained. The aggregate 



