642 



SCIENCE. 



[N. S. Vol. XX. No. olS. 



of the former, the north pole (N^) lying 

 now west of the south pole {82)- The 

 poles of the two systems are situated, ap- 

 proximately, near the 40° parallels— this 

 is even true of the tertiary system N^S^. 



The secondary magnetic equators (the 

 lines along which the residual vertical force 

 is zero) occupy practically the same posi- 

 tions for the three epochs. It is as yet too 

 early to decide as to any probable secular 

 shifting of the positions of the secondary 

 poles. The interval is too short, in view 

 of the meagerness of the data on which the 

 charts depend, to make certain any deduc- 

 tions. 



What has thus far been gained by the 

 decomposition of the earth's total magnetic 

 field into a primary and into a secondary 

 one F 



In the first place, the residual field clear- 

 ly exliibits the fact that it is not a hetero- 

 geneous one, but, in general, remarkably 

 systematic in its structure. There is, there- 

 fore, a very strong indication that it is 

 produced by some distinct physical cause 

 operating in the same general manner over 

 the entire earth. The hope is thus clearly 

 held out that we may still further resolve 

 the residual field, starting with funda- 

 mental, physical causes. 



My present belief is that the chief phys- 

 ical cause of the residual field is to be re- 

 ferred to the distribution of temperature 

 within the stratum of the earth 's crust here 

 concerned. 



There is a very remarkable correspond- 

 ence between the principal features of the 

 residual magnetic field and those exhibited 

 on a chart of isabnormal temperatures. 

 It is found that the earth as a magnet acts 

 like any other magnet as regards applica- 

 tion of heat. Thus, wherever the earth's 

 surface is relatively warm, on the average 

 for the year, there the magnetization of 

 the earth shows a decrease, and where, on 

 the other hand, it is relatively cold, there 



it suffers an increase. The comparison 

 held so far, that it is possible to reproduce 

 the residual magnetic field, in its general 

 characteristics, with the aid of temperature 

 charts. 



The criticism has been made that this 

 relation between residual magnetism and 

 temperature distribution may only be an 

 apparent one, since the latter referred to 

 surface conditions, whereas the former per- 

 tained to strata at considerable depths be- 

 low the surface. However, the isabnormal 

 temperatures plotted were based on annual 

 means; hence the effects due to annual 

 variation and diurnal variation were elim- 

 inated. I am not aware that any one has 

 given a physical explanation of the situa- 

 tions of the maxima and minima shown on 

 an annual isanomalous temperature chart. 

 Their annual positions are probably largely 

 dependent on the radiation of the internal 

 heat of the earth. We can not say, as yet, 

 at what depth the principal thermal fea- 

 tures shown at the surface are eliminated; 

 it is known that the isothermal surfaces in 

 the interior conform with those of the sur- 

 face to a considerable depth. In any case, 

 there is no question that as land areas are 

 pierced, a steady increase of temperature 

 is encountered. Over oceanic areas, on 

 the other hand, there is at first a decrease 

 until nearly a zero temperature is reached 

 at the ocean beds, and then, presumably, 

 an increase as the penetration continiies. 

 So that we shall have temperature gradi- 

 ents along parallels of latitude down to a 

 considerable depth. 



I shall not discuss this matter further 

 now, as it is being made the sub.ject of a 

 special examination. Many have surmised 

 that the distortion of the earth's magnetic 

 field is to be attributed to the distribution 

 of land and water ; but the problem is to 

 show in what manner the distribution 

 causes the observed effects. The first at- 

 tempt, as stated, will be to ascertain whether 



