Reductions to Standard Instruments 63 



Kelvin compass and azimuth instrument {K). — Declination results by the Kelvin dry 

 compass and azimuth instrument were for the most part experimental and were used only 

 as checks upon determinations by the other compasses and azimuth circles. 



Horizontal-Intensity Observations. 



Sea deflector for horizontal-intensity observations. — As shown in specimen Form 25, page 

 38, the horizontal intensity is computed from sea-deflector observations by the formula 



H. ^^ 



sm u 



in which m is the magnetic moment of the deflecting magnet, C is a constant involving the 

 deflection distance (r), the distribution coefficients (P and Q), and the induction factor 

 fjL, and u is the observed angular deflection produced by the deflecting magnet when its axis 

 is perpendicular to that of the compass card. The sea deflector is a relative instrument 

 and values of the so-called constant, mC = H sin u, must be determined from comparison 

 horizontal-intensity observations made at shore stations with standardized absolute 

 instruments. The constant, mC, is subject to changes arising from (1) decrease in m with 

 time, (2) effects of temperature variations on m and r, and (3) effects of changes in vertical 

 intensity, Z. In the Galilee work, except as noted below under constants for Cruise I, all 

 available data for log mC were subjected to a least-square adjustment based on the general 

 formula 



log mC = log mCio at To-l-xAr-f-i/(Ar)Hg(20''-<) 



in which r is the epoch of observation expressed in years, tq is the selected reference epoch, 

 At is (r — To),q is the factor representing the combined effect of a change in temperature of 

 1° centigrade on m and C (on the latter because of change in r), and t is the temperature 

 of observation ; the standard temperature of reference is 20° centigrade. Instead of deriving 

 all the unknowns in above equation sinuiltaneously, it was found better to make a separate 

 determination of the temperature factor, q, selecting the observations best suited for this 

 purpose. The final results were arrived at by a process of successive approximations, in 

 the last steps of which q was treated as a constant. 



As will be noted, the general form contains no term to correct for effects of changes in 

 vertical intensity (Z) . In the Galilee work it is reasonable to assume that such effects were 

 at no time in excess of those determined for the improved and more accurate revolving- 

 compass pattern of sea deflector used on board the Carnegie. In the case of that instrument 

 the maximum effect (see pp. 238-239) on log mC, corrected for time and temperature, was, 

 for the extreme range in Z, of the order 0.0020, which is equivalent to less than 0.005//^; in 

 general, the correction for the AZ-effect would be less than half of that amount. More- 

 over, the AZ-effect was partly eUminated by the introduction of the (Ar)'-term in the adjust- 

 ment of the constant observations made at widely distributed ports. It is a fair assumption 

 therefore, that the introduction of an additional term in log mC, to correct for outstanding 

 effects of changes in Z, would probably not change the values computed by the adopted 

 formulae more than 0.001 5/f in the extreme case. With the compass-azimuth-circle pat- 

 tern of sea deflector used during the Galilee cruises, the error of observation at sea is about 

 of the order 0.002i7; a recomputation is, therefore, not warranted — especially since, in 

 general, the values given in the present volume are the means of both deflector and dip- 

 circle determinations. 



Sea deflector (Dl). — For Cruise I, the adopted values of log mC are the means of all 

 shore observations, the available data being insufficient for a reliable determination of a 

 time-change coefficient for the short period of this cruise. The constants adopted from 

 August 2 to December 13, 1905, for the 2 positions of thedeflecting magnets in their housings, 



