133 



North comp. East comp. 



Ampl. Phase Ampl. Phase 



5 cm. 20^ 23 cm. 15'^ 



whilst above from the balloon observatioDwS have been dednced 28 ctn. 

 and 20° as being the mosl probable valnes for the East component; 

 a small value being found for amplitude of the North component. 



Here, therefore, is a close ae;reement. 



This agreement holds up to a height of 4 km., because the phase of 

 the semi-diurnal variation of the East component follows the one of 

 the air-pressure, and higher up the amplitude of the North component 

 also remains considerably smaller than the one of the East component. 

 E.g. at an average from 4- 5 km. we deduce from our observations: 



North component 11^ cm. East component 51 cm. 



51 



proportion = 4.5, wherens Gold finds on the surface of the earth 



11. 5 



5 and 23 cm., i.e. a proportion of 4.6 ; but as to a reverse in the 



change of phase there is nothing to be found in Gold's treatise. 



However, are there any indications to explain a similar reverse? 

 • In the first place we looked for them in the vertical motion of 

 the air, as found by Gold's theoretical investigation and have 

 considered whether, as a consequence of the rising or descending of 

 air, a returning flow in the upper layers may take place. 



According to Gold the maximum vertical velocity of the semi- 

 diurnal motion occurs about the time of the greatest horizontal 

 velocity viz. an upward motion in the case of Western motion ; e.g. 

 close to the equator at 5 km. he finds an amplitude of the vertical 

 oscillation of 2 mm. pro sec. If in consequence of this we assume 

 an average rise of 1 mm. pro sec. occurring over ^|^ of the 

 earth's surface or 10.000 km., then we may also surmise that 

 this air would again have to flow off through a layer of a thickness 

 of say 10 km. This would make an average velocity of 1 m. pro sec, 

 i.e. actually a velocity of the order required and it may be accepted 

 as probable that this velocity is in inverse direction to that near 

 the earth's surface. 



Also, that on account of the ever increasing outflow and inflow 

 with increasing height, a gradual increase (up to inversion) of phase 

 should occur. 



Apparently a similar explanation would not seem illogical, but 

 often such arguments are misleading when we are solving suchlike 

 problems, so that we will only consider it as a hint in that direction. 



Finally we examined the possibility that the inversion of phase 



