Reduction Formula and Instrumental Constants 233 



more, susceptible of accurate adjustment by the method of least squares (see pp. 185-189). 

 For specimen determinations of constants, see page 186. 



Revolving-compass pattern of sea deflector. — The improved sea-deflector, designed and 

 constructed by the Department of Terrestrial Magnetism, has been used as an auxiliary 

 declination-instrument on board the Carnegie. For descriptions of the various forms of the 

 instrument and explanation of its use, see pages 190-195. Specimen observations and com- 

 putations for a sea station are given on pages 215-218. The purely instrumental corrections 

 arise from (1) card-graduation errors and eccentricity of card mounting; (2) magnetic-axis 

 and index error; (3) lack of correct adjustment of the sighting vanes attached to the bowl. 

 Card-graduation errors, e, were determined at shore stations by observing the magnetic 

 azimuths of series of marks in the horizon, i. e., at an altitude practically 0°, the marks 

 being selected to give as nearly equal angular distribution as possible. The magnetic 

 azimuths were controlled by simultaneous decUnation observations with a standardized 

 magnetometer at a second station. For deflector 3 the card-graduation and eccentricity 

 errors are small (see p. 235), while for deflector 4 they are entirely negligible. The correc- 

 tions. Ate for the "bright-Une" method and A,c for the "shadow" method, both include 

 corrections for the second and third classes of errors above ; they may vary with the altitude 

 of the object sighted upon. The data for the establishment, by graphical means, of a 

 curve representing the variation with change in altitude, were secured at shore stations 

 from series of declination observations on the Sun. The absolute values of the declination 

 were determined from simultaneous observations with the standardized magnetometer. 

 The total correction to the card reading, depending upon the sighting method used, is 

 e-Hiljc or e-f-vl„. 



Each of the terms making up the total correction to the observed card reading, viz, e, 

 and A,x or A,c, is given separately in this volume for each deflector. The signs attached 

 are in the sense of continuous graduation from the south point as 0° through 360° in a 

 clockwise direction. Accordingly, all card-readings in the southwest and northeast quad- 

 rants, that is, all readings from S to S 90° W (or W) and from N to N 90° E (or E), must 

 be numerically increased when the sign given for e, Au, or A„ is plus {+) and vice versa; 

 all card-readings in the other two quadrants must be numerically decreased when the sign 

 given for e, A„c, or Asc is plus (-|-), and vice versa. 



Specimen observations and computations to determine the change in A^c with altitude 

 are given on page 234. 



Marine collimating-compass 1 (Cl). — Marine coUimating-compass 1, designed and 

 constructed by the Department of Terrestrial Magnetism, was used on the Carnegie from 

 September 1909 to October 1916. During April to May 1914, before Cruise III, the 

 instrument was thoroughly overhauled and repaired. 



The adopted constants, resulting from least-square adjustments of all available data, 

 are given in Table 52. Specimen observations for the determination of constants are given 

 on page 186. 



Sea deflector 3 (DS). — The adopted periodic corrections to observed card-readings of 

 the compass of D3, used on Cruises I and II (as far as Cape Town, March 1911), are given 

 in Table 53. 



The adopted correction, A,c, to observed card-readings by the "shadow" method is 

 for all altitudes: 



A„ = -f0?09 from August 1909 to September 1910 

 A„ = -0.06 from October 1910 to March 1911 



The adopted corrections, Atx, to observed card-readings by the "bright-line" method, 

 deduced from a graphical adjustment of all available data, vary with the Sun's altitude, 

 and are given in Table 54. 



