ON COMPARING AND REDUCING MAGNETIC OBSERVATIONS. 81 



extent and producing no real inconvenience in diurnal inequalities de- 

 pending on all days of the month, in which also days of different character 

 become combined, became very pronounced in the five day results, because 

 of the small number of days employed and their restriction to days of one 

 kind, those of quiet magnetism. The effect in question is the non-cyclic 

 character of the resulting diurnal inequality, which in horizontal force he 

 found to be such as to cause the value at the terminating midnight 

 to be persistently greater than at the commencing midnight. And 

 in the Report for 1896 Dr. Chree discussed the matter at greater 

 length. 



My present object may be said to be to discuss to some extent the five 

 day results for Greenwich, in order to make comparison with results that 

 Dr. Chree has found for Kew, and also to compare for Greenwich alone 

 the diurnal inequalities and diurnal ranges as found from the five day 

 tabulation with those given by the ordinary Greenwich tabulation, in 

 which only days of extreme disturbance are omitted. When the all day 

 tabulation is spoken of, it will be understood as implying the omission of 

 days of extreme disturbance. The five day tabulation will be indiffer- 

 ently distinguished thus, or as the quiet day tabulation. I propose to 

 consider first the comparison of Greenwich results. 



To avoid lengthened titles to the various appended tables, I would 

 first remark that the Greenwich results are all founded on mean results 

 for the individual months of each year, these being combined in the differ- 

 ent tables by months (that is, grouping together the same month in 

 different years), or in quarters, or years, as necessary. Thus, in Table I. 

 the results for each month depend on twenty-five days (five in each of the 

 five years combined), and the quarters on seventy-five days, and so on. In 

 the tabulation by quarters, first quarter is to be understood as comprising 

 the months of January, February and March, and second quarter those of 

 April, May, and June, and so on. In the seasonal tabulation, Spring is 

 to be understood as including the months of February, March, and April, 

 and Summer those of May, June, and July, and so on. 



The values for declination are given in minutes of arc ; those for 

 horizontal and vertical force are throughout in c.g.s. measure x 10^, 

 excepting in Table V., in which they are in c.g.s. measure x 10\ The 

 approximate absolute value of horizontal force at Greenwich for the period 

 treated is 0-183 c.g.s. and of vertical force 0-437 cg.s., and nearly the 

 same at Kew. To convert declination values in arc into westerly force 

 in c.g.s. measure, to make comparison in the various tables with the 

 numbers for horizontal and vertical force, multiply the minutes of arc 

 by 53, excepting in Table V., for which the factor is 5-3. This it is 

 convenient to remember for the ready estimation of the comparative mag- 

 nitude of declination and horizontal and vertical force variations.' 



In Tables I., II., and III., the Greenwich results for the years 1890 to 

 1894 have been combined to form mean monthly diurnal inequalities of 

 declination and horizontal and vertical force, both for the all day tabula- 

 tion and the quiet day tabulation, for comparison of the diurnal inequali- 

 ties resulting from the two different methods of tabulation. The days 

 omitted in the all day tabulation on account of extreme disturbance, 



' When horizontal and vertical force values are hereafter quoted in the text with 

 the letters c.g.s. attached, they are to be understood, excepting for Table V., as 

 mdicating c.g.s. measure x. 10". 



1898. G 



