ABSOLUTE RANGES. 



135 



39. Table XLIX gives an analysis of the results obtained by grouping the ranges of Declination 

 according to their amplitude. The first group gives the number of days in which the range did not 

 exceed 30', the second the number of days in which the r;inj,'o, while in excess of 30', did not exceed 1, 

 and so on, the seventh and last group giving the number of days in which the range exceeded 3. Days 

 were arranged in two principal classes, the first including all days when the record was complete, or when 

 loss of trace was due solely to one or both limits of registration being exceeded, the second including 

 all days when there was imperfection of record through photographic failure, absence of sheet, or similar 

 cause. Each of these two classes was sub-divided into two sub-classes, according as the trace did or did 

 not keep within the limits of registration. The days are arranged under three seasons, Midwinter (May 

 to July), Midsummer (November to January), Equinox (March, April, September, and October). The 

 results for the remaining months, August, 1902 and 1903, and February, 1903, are combined. 



TABLE XLIX. Declination Ranges. 



It will be noticed that out of a total of 111 clays in Midsummer there was only one in which the range 

 did not exceed 1, while at least 44 had a range in excess of 3. Taking the whole period, we find that 

 out of 461 days for which the record was complete, except for the limits of registration being exceeded, 

 329, or 71 per cent., had a range over 1; 250, or 54 per cent., had a range over 1|; 179, or 39 per cent., 

 a range over 2, and 75, or 16 per cent., a range over 3. Of the 581 days, complete and incomplete, 

 included in the table, 407, or 70 per cent., had a range over 1; 218, or 38 per cent., a range over 2; 

 90, or 15 per cent., a range over 3; while 24, or fully 4 per cent., had a range over 4. On seven days 

 the trace exceeded the limits of registration on both sides of the sheet, whose complete width represented 

 from 4 50' to 4 55'. 



40. Results for H corresponding to those for D, just discussed, appear in Table L. It contains two 

 principal classes, each with two sub-classes analogous to those in Table XLIX. The ranges are again 

 dealt with in seven groups, the first containing days in which the range did not exceed 25y, the second 

 days in which the range exceeded 25y but did not exceed 50y, and so on, the last group containing days 

 when the range exceeded 150y. Owing to the sensitiveness of the Horizontal-Force magnetograph, the 

 limits of registration were exceeded in about one day out of two. In Midsummer the limits were exceeded 



