THE EAETH’S MAGNETIC FIELD IN INTEENATIONAL UNITS, 
435 
of comparisons between the different magnetic observatories carried out by Professor 
Rucker and the author. The accuracy of the circle was tested, during a long series 
of comjjarisons made between this instrument and the Kew standards, by measuring 
the angle subtended by the two fixed marks used for the declination observation at 
Kew with different parts of the circle. In this way no error amounting to 10 seconds 
of arc, the smallest angle which can be read off on the circle, was detected in any 
part of the circle. Hence, as the average value of the angle used in the following 
measurements was 46°, the accuracy attainable was amply sufficient. 
A general view of the coils, as they appeared when in use, is shown in fig. 1. The 
coils were carried by two cross bars. A, the points of four screws, C, attached to these 
cross bars resting on the cylindrical parts, B B, of the magnetometer. These 
screws, instead of being fixed directly to the cross bars, were attached to two metal 
plates which could slide along the bars. Thus, by moving these sliding pieces, as 
well as by the screws C, the relative positions of the coils and the magnetometer 
could be adjusted. The coils were prevented from turning about a horizontal axis by 
two screws, H, resting on the top of the deflection bar E. The position of the 
magnet, M, was observed by means of a plane mirror attached to the magnet and the 
telescope, T, and scale, S. The magnetometer rested on a stand, F, which passed 
through a hole in the cylindrical side of the coils. The space between the spokes of 
the flanges was such that when the telescope and the glass suspension tube were 
removed the magnetometer could be inserted. In order to 
relieve the magnetometer from the full weight of the coils, part 
of the weight was supported by strings attached to the top of 
the flanges, which, after passing over a pulley, had a counter 
weight attached. 
As the construction of such large coils with the degree of 
accuracy aimed at in these experiments is a problem of 
considerable difficulty, and one which has frequently arise]i in 
experimental researches, a short description of the method 
employed may be of some interest. The flanges of the coils in 
which the grooves to receive the wire are turned out are 
formed of two gun-metal castings. The section of a portion of 
one of them is shown in fig. 2. The spokes were strengthened 
by a rib 1‘5 centim. deep. In order to turn up these castings 
a large cast-iron bed was made with planed ways on the top to 
take the fixed and movable headstocks of one of the lathes in 
the laboratory. This bed is shown in fig. 3, and was of such a 
size that the finished coil could be swung between centres. A 
gun-metal face-plate, slightly larger in diameter than the coils, 
was also cast and turned up on the bed itself Teeth were cut 
round the circumference of this face-plate, and it was driven by a spur-wheel geaibig 
3 K 2 
0 Xcm. 
of Cell 
