392 Special Reports 



20. In the instrument described here, n/\ with a certain suspension of phosphor- 

 bronze strip, is only about 0.001, and the mirror can easily be hung on the suspension 

 in such a way that the fiber is not twisted by more than a few degrees. Thus a need not 

 be greater than a fraction of 1'. The angle $ for the Bureau of Standards mirror is less 

 than 2', and for the mirror constructed here is about 6'. The angle 70, as stated above, 

 need not exceed a few minutes at most. In using the instrument 7 and 7' should also be 

 kept small, though cos(/3 7), multiplying GI in (1), and neglected in (17), does not 

 differ from unity by more than 0.0001 until the angle reaches nearly 50', corresponding 

 to about 30 divisions on the scale when the scale distance is 35 cm. In a similar way it 

 is seen that no appreciable error in H will be made through even a considerable error 

 in leveling the instrument. 



The angle 70 is made zero, or nearly zero, by the following process: First, the axis 

 of the cylindrical chamber of the magnetometer box within which the magnet hangs 

 is made parallel with the axis of the coils. This is done with the aid of the brass cylinder 

 H shown in Figures A, B, and C of Plate 9 and the bent pointer shown in Figures A and B. 

 The cylinder, which fits the chamber precisely, is pushed in until the axial motion is 

 stopped by a terminal shoulder, and the magnetometer box is then moved in azimuth 

 until the pointer, properly adjusted, just touches, or comes within the same minute distance 

 of touching, the surface of the marble cylinder at both ends of a horizontal diameter. The 

 screws clamping the two flanges C and D, Figure C, together are then tightened. 1 



Second, the axis of the cylindrical chamber is brought into the vertical plane con- 

 taining the axis of collimation of the telescope and the central division of the scale 

 (which can be shifted laterally if necessary) immediately beneath it. To accomplish 

 this, the brass tube, Figures A and B, carrying a plane mirror at one end, is shoved into the 

 cylindrical chamber of the magnetometer box, which it fits precisely, until the motion is 

 stopped by a square terminal shoulder. Then the telescope, adjusted in azimuth until 

 its axis is normal to the scale as judged by the eye, and previously focused upon the 

 scale as seen with the magnet-mirror, is turned, together with the arm which supports 

 it, about the axis of the magnetometer until the center of the scale is at the intersection 

 of the cross-hairs, and clamped. As the surface of the plane mirror is only a short dis- 

 tance in front of the magnet-mirror, the focus is still sufficiently good. If the plane 

 mirror is not strictly normal to the axis of the tube, the adjustment can nevertheless 

 be made exact by using two settings of the tube 180 apart. 



21. The only satisfactory method of measuring the current J through the coils of 

 the instrument is by means of a standard resistance coil and a Weston standard cell. 

 This can be done in three different ways. 



(1) The potential difference between the voltage terminals of a standard resistance 

 coil, with resistance R, in series with the instrument, may be made equal to the open 

 circuit electromotive force e of a Weston standard cell by adjusting the current in the 

 circuit. The current is then determined from the relation J ejR. 



(2) The potential difference E between the terminals of a standard resistance coil, 

 with resistance R, in series with the instrument may be determined by comparison with 

 the open circuit e. m. f. c of a Weston cell, through the agency of a standardized poten- 

 tiometer or suitable arrangement of resistance coils, and J found from the relation 

 J = E/R. 



(3) With the arrangement of (2) otherwise unchanged, a standard cell, with open 

 circuit e. m. f. e, may be inserted between the voltage terminals of the standard resist- 

 ance coil and the potentiometer, its direction being such as to oppose the potential 



1 The cylinder H was made for use in turning the spool support. For use in making the adjustment of 70. a very short 

 cylinder with as broad a shoulder as practicable, making close contact with the plane face "f the magnetometer box, would 

 be preferable, as the adjustment could then be made without removing the magnet-mirror. 



