REPORT OF THE SECRETARY. 19 



deduce, if possible, the law by which they are governed. This is 

 the most intricate part of the whole series, and is the final object to 

 which the preceding investigations were preparatory. From a care- 

 ful study of all the observations on the dip, it was found that 1,446 

 might be considered as giving abnormal values, and of the total force 

 of the earth's magnetism 1,470 indicated abnormal changes, which 

 amounted to about one-fifteenth part of the whole number of obser- 

 vations. These abnormal disturbances were analyzed in relation to 

 their frequency during the hours of the day, the month of the year, 

 and successive years ; they were also studied as to their tendency to 

 exhibit an increase or diminution in their variations and the times of 

 the greatest and least action in relation to the periods above men- 

 tioned. This part ends with a table of the relative magnitude of the 

 disturbances and a comparison of those of Toronto and Philadelphia, 

 from which it appears that in some cases there is an agreement in 

 the character of the simultaneous changes in the two places, and at 

 others not. 



The second part of the fourth section treats of the solar diurnal 

 and annual inequality of the dip and total force, that is, of the changes 

 due to the sun which take place in the dip and total force from 

 hour to hour and from month to month. In this investigation all the 

 greater perturbations are omitted and the laws of the simpler 

 or more normal changes are sought. The diurnal changes in the dip 

 are shown analytically and graphically for each month and for the 

 whole year. The general character of the curve exhibits a maxi- 

 mum at about 11 a. m. and a minimum at about 5 a. m., with a range 

 of one minute and two-tenths — a quantity too minute to be recognized 

 by the ordinary dip circle, and which can only be observed by the 

 differential reflecting instrument. In summer the epochs occur ear- 

 lier, with a range of a minute and a half, while in winter they occur 

 later, with a range of only one minute. There is also a secondary 

 fluctuation of small magnitude. The diurnal deviation of the dip is 

 greatest about the time of the equinoxes, and of these maxima that 

 of the winter is the least. The diurnal changes in the total force as 

 deduced from the average of the year are represented by a single 

 crested curve, but from the average of the observations in winter 

 alone this assumes the form of a double curve. The principal maxi- 

 mum as deduced from the whole year coincides with the hour of 

 2 p. m., and in winter occurs about an hour and a half earlier. The 

 principal minimum coincides with 10 p. m., and occurs in winter about 

 two hours earlier. This part ends with an attempt to deduce from 

 the data the annual changes in dip and intensity. The result, how- 



