SECT. 1] THE MAGNETIC FIELD OVER THE OCEANS 185 



N, is about 100,000. A change of unity in the count then corresponds to a 

 change of | y in F. If the possible accuracy is to be achieved, the gate that 

 starts and stops the 100-kc/s count may be opened and closed consistently to 

 10"5 sec ; that is, to 1/50 of a cycle of the proton signal. As the signal decays to 

 about 30% during the count, the gate must be designed so that its times of 

 opening and closing are not affected by the amplitude of the signal. 



The last three digits of the count are recorded on a potentiometer recorder 

 and all six digits on punched tape. A description of the punching arrangements 

 has been given by Bullard (1961). The whole equipment is cycled once every 

 minute, or once every half minute. The cycle is initiated by a clock driven by a 

 precision oscillator. 



3. General Features of the Field 



Over most of the earth's surface the field is within 20% of that which would 

 be produced by a dipole at the centre of the Earth. No analysis has been made 

 for the land and the oceans separately, but there is no reason to suppose that 

 there is any significant difference in the moment or direction of the dipole 

 between land and sea. If the field due to the best-fitting dipole is removed, the 

 residue is known as the non-dipole field. The three components of this for the 

 year 1945 are shown in Fig. 5. The rate of change of the vertical component in 

 1942 is shown in Fig. 6. 



Such diagrams use data collected over a long period and are the result of 

 extensive smoothing and extrapolation both in time and space ; they do not 

 represent the immediate results of observation in the way that the magnetic 

 survey of a single country does. The extent of their deficiencies is illustrated by 

 measurements made by recent Japanese Antarctic expeditions (Nagata, Oguti 

 and Kakinuma, 1958; Oguti and Kakinuma, 1959). They found that in the 

 region between South Africa and the Antarctic continent the field is decreasing 

 by 200 y/year, which is about three times the rate of decrease given by current 

 magnetic charts. 



In spite of these uncertainties, Figs, 5 and 6 do demonstrate that the great 

 geographical and geological differences between the continents and the oceans 

 are not reflected in the world-wide pattern of the magnetic field. This is further 

 demonstrated by Fig. 7, which shows a section from a chart of the total field in 

 one of the areas where it is best known. The lines representing the smoothed 

 field run from continent to ocean without any conspicuous disturbance. This 

 broad similarity of the field over oceans and continents is to be expected if the 

 main features of the field have their origin in the core of the earth, the outer 

 surface of which is at a depth of 2900 km. 



A possible exception to this general similarity has been suggested by Run- 

 corn, who has pointed out that the centres of rapid change, shown in Fig. 6, tend 

 to avoid the Pacific Ocean. It is possible that this is due to the temperatures 

 beneath the Pacific being higher than elsewhere and producing a higher electri- 

 cal conductivity in the mantle and a greater screening of the secular variation 

 (Bullard, Maxwell and Revelle, 1956). 



