Reduction Formula and Instrumental Constants 239 



The first effect (1) may be expressed by the introduction of a temperature factor q in 

 a term of the form qAt. The factor q may be determined from shore observations by 

 selecting those best suited, i. e., those made at stations where the range in temperature 

 during observations has been large, and for which magnetic conditions were normal. The 

 separate determinations of q from various shore observations must be weighted according 

 to the mean range of temperature from which each is deduced; it may also be necessary 

 to weight values somewhat according to the magnetic conditions at the station concerned. 

 Every precaution must be taken, of course, to guard against sudden or irregular tempera- 

 ture changes during observations to determine constants at shore stations. 



The second effect (2) may be expressed satisfactorily, at least for well-seasoned mag- 

 nets, by a term involving the first power of the difference in time from some selected epoch. 

 To, and an aging coefficient x, thus: x {r — ro), or xAr. 



The third effect (3) is more complex. There is a value of the vertical intensity, desig- 

 nated 2, for which the magnetic field of the compass system of magnets, tilted about the 

 pivot support by the action of Z, is symmetrically disposed with reference to the field of 

 the deflecting magnet. Any departure of the magnet system from this balanced position, 

 such as occurs when the instrument is at a station for which the value of Z is different 

 from z, always increases the constant mC. This is shown by examination of the deflector 

 intensity-constants on Cruises I to IV, for each magnet and for each distance used. The 

 third effect (3) may, accordingly, be expressed by a squared term, involving the unknown 

 2, as defined above, and multiplied by a vertical-intensity factor y, thus: y {z — Zy. 



It appears, therefore, that the variable intensity-constants for the deflector may be 

 expressed by the general formula 



log mC = w + xAt + y{z-Zy + qAt 



in which w represents the value of log mC at a standard temperature, t„ for an epoch, to, 

 at a place where the vertical intensity is z; and in which At is (t, — t),t being the tempera- 

 ture of observation. 



To obtain data for the determination of the intensity constants and their changes, 

 observations of intensity with the sea deflector are made at every port for each deflecting 

 magnet and each deflection distance. Simultaneous determinations of intensity are always 

 obtained with a standardized magnetometer at an auxiliary station; the correction for 

 the difference in intensity between the two stations is determined by simultaneous mag- 

 netometer observations, involving exchange of station in accordance with the usual practice 

 (see Vol. I, p. 219). The deflector observations are always made for four different orien- 

 tations of the bearing ring, to eliminate and determine possible periodic effects. The scheme 

 of observation followed is similar to that used on board ship (see p. 194). 



On page 240 are given specimen horizontal-intensity observations with sea deflector 4 

 at the shore station Suva Vou, A, Fiji, for the deflecting magnet 45 at deflection distances 

 1 and 3 and for the orientation 0°, together with an abstract of all the results and compu- 

 tations of log mC, at the same station, for both deflectmg magnets 45 and 2L at deflection 

 distances 1 and 3. Observations to determine log 7nC for magnet 2L at distances 1 and 3 

 for orientation 0°, corresponding to those given in the specimen for magnet 45, were made 

 between chronometer times 3^ 22" and 4*^ 02° and 3'' 31° and 3'' 53°, respectively; those 

 determinations thus apply to practically the same mean time as the determinations with 

 magnet 45. The same order of observation is followed for each orientation. 



