358 
Transactions of the Boyal Society of South Africa. 
pointed steel rods at the top were then fixed one behind the other, by 
bringing the images in coincidence with the cross-wires. The position 
of the magnetometer was then so adjusted that the straight line joining 
these two rods when produced passed through the centre of suspension 
of the magnet. 
The magnet's suspension consisted of several strands of unspun silk 
fibre, and the magnet was not removed during the period over which 
No. 1. August 15th. No. 3. September 2nd. 
No. 2. August 22nd. No. 4. September 13th. 
the observations were taken. The effect of any change in the torsional 
zero was probably negligible and could not have affected the value 
of the declination in the most extreme cases by more than a few 
seconds. 
As the observations were "eye" ones the exact time, and also the 
magnitude of the maximum or minimum values, can only be deduced 
by interpolation, so that those given in the table are approximate. The 
declination was found to be very near to 24° W., but of course it varied 
during the day, the variation amounting to a few minutes. The observa- 
