On Variations in the Magnetic Declination at Bloemfontein. 359 
tions were taken as frequently as possible between the hours of 6 a.m. 
and midnight, and at intervals of about two weeks a complete day's record 
of hourly observations was secured. 
Table I. shows the time of the two maxima and minima for each day 
in September and October, whilst Table II. gives the complete daily 
records from midnight to midnight in divisions of the scale for selected 
No. 5. September 25th, No. 7. October 20th. 
No. 6. October 7th. No. 8. October 31st. 
days. The readings were taken with the magnet swinging slightly, and 
the scale in the magnet could easily be read to y^^th of a division. The 
value of each division was stated on the Kew certificate to be 1*8'. 
The extreme range of the variation on any day did not exceed 10*8' for 
the maximum and 2-6' for the minimum. The first diurnal minimum to the 
east occurred at about 7 a.m. and the second somewhere between 2 and 
and 3.30 p.m., but many divergencies may be noted. The curves drawn 
from the twelve complete days' work are seen to be fairly regular and 
