106 



THE ROUYAL SOCIETY OF CANADA 



gression so obtained in the hourly means from a large number of days 

 is seen in but very few if any of the individual days. 



In Table I is given the diurnal inequality of Declination at 

 Agincourt for each month of the year derived from all days of the eleven 

 years 1902 to 1912. Distinct maxima and minima are in heavy type 

 and the last line gives the ranges as derived from the hourly values. 



Table I. — Diurnal Inequality of Declination at Agincourt — All days, 1902-1912. 



Jan. Feb. Mar. Apr. May J 



July Aug. Sept. Oct. Nov. Dec. 



-0-17 

 + 05 



-0-25 

 -0-31 

 -0-30 

 -0-48 

 -0-63 

 -1-44 

 -2-58 

 -2-31 

 -011 

 + 1-79 

 j + 2-98 

 + 3-27 

 + 2-81 



16 +2-05 



17 1+0-94 



9 



10... 

 11 



Noon 

 13. .. 



14 



15 



+ 0-35 

 -0-36 

 -0-82 

 -1-23 

 -1-58 

 -101 

 -0-66 



-0-95 

 -0-82 

 -0 59 



-0-83 

 -100 

 -1-23 

 -2-29 

 -3-60 

 -3 73 

 -2-611 

 000 

 + 3-04J 

 + 4-63 

 + 5 03 

 + 4-36 

 + 3-11 

 + 1-66 

 + 0-85 

 + 0-13 

 -0-47 

 -1-14 

 -1 36 

 -1141 

 -1-05 



-0-71 

 -0-71 

 -0 57 

 -0-94 

 -2-33 

 -3-88 

 -4-59 

 -4-84 

 -3-69 

 -111 

 05 +2-31 

 + 4-77 

 + 5 60 

 + 5-48 

 + 4-52 

 + 2-91 

 + 1-36 

 + 0-15 

 -0-24 

 -0-42 

 -0-72 

 -0-66 

 -1 01 

 -0-70 



29-0-74 



01 +016 



361+0-07 

 70J-0-41 

 99—1-74 

 3-71 

 5-32 

 5-78 

 4-82 

 2-21 

 21 



251+4-21 

 44+5-62 

 36J+5-73 

 601+4-82 



23 +3-51 

 61 +1-74 

 53 +0-60 



02 -0-21 



24 -0-20 

 71 -0-45 

 55-0-68 



83 

 0-67 



78 

 21 

 49 

 30 



81 

 471 + 1 



-0-52 

 -0-35 

 -1-18 

 -1-95 

 -1-73 

 -2-92 

 -4-52 

 -4 63 

 -3-17 

 -0-33 

 + 3-08 

 + 5-76 

 + 6-22 

 + 5-43 

 + 3-47 

 + 1-56 

 + 0-21 

 -0-35 

 -0-49 

 -0-66 

 -1-29 

 -0-94 

 -0-48 

 -0 07 



-0-46 

 -0 26 

 -0-63 

 -0-85 

 -0-90 

 -0-87 

 -1-96 

 -2-75 

 -3 02 

 -1-36 

 + 0-95 

 + 3-49 

 + 4-49 

 + 4-20 

 + 3-26 

 + 1-99 

 + 0-96 

 + 0-48 

 -0-45 

 -1-28 

 -1 70 

 -1-28 

 -111 

 -0-88 



-0-12 

 + 37 

 -000 

 -0-47 

 -0-91 

 -1-34 

 -1-37 

 -2-02 

 -2 37 

 -112 

 + 0-96 

 + 3-04 

 + 3-79 

 + 3-51 

 + 2-42 

 + 1-47 

 + 0-70 

 + 005 

 -0-72 

 -1-13 

 -1-50 

 -1-61 

 -109 

 -0-64 



-0-32 

 + 0-19 

 + 019 

 -0-25 

 -0-64 

 -0-76 

 -0-89 

 -1-35 

 -1-97 

 -1-26 

 + 0-46 

 + 2-33 

 + 3-28 

 + 309 

 + 2-31 

 + 1-62 

 + 0-71 

 + 0-11 

 -0-72 

 -1-21 

 -1-72 

 -1-21 

 -115 

 -0-56 



Range j 5-85 



1-51 13-13 



lo ss 



7-51 6-16 



5-25 



The monthly diurnal curves are exhibited graphically in Plate 1, 

 where the distinctive characteristics of the different seasons of the 

 year can be readily seen. The extent to which the type of the curves 

 vary with the season can better be grasped by taking the inequality 

 for March and drawing it to scale, and then represent December and 

 June on scales so related that the mean length of the 24 hour ordinates 

 is the same as for March. This is shewn in Plate II. It will be seen 

 that the difference in type is mainly confined to the hours between 

 6 p.m. and 9 a.m. when the differences from the mean are the least 

 and the changes less rapid. 



If now we analyze these diurnal inequalities as given in Table I 

 by means of Fourier series it is found that a close degree of accuracy 

 is attained from the first four terms having periods of 24, 12,8 and 6 

 hours respectively, an accuracy which is less than the probable error 

 of observation of the magnetic elements. It must not be assumed, 

 however, that distinct natural forces are in operation, which may 

 be assigned to these different Fourier waves, although if such forces 

 did exist it would add greatly to the usefulness of this method of 



