131 



When calculating the drift components, it was assumed that the 

 height of the clouds was constant: IJ.O lun. for Ci and 10.5 km. 

 for Ci Cu. ') 



For the East component has been found : 



and consequently for the semi-diurnal term 124 cm. ^m (2 ^-[-266°). ") 



No doubt this result is pretty uncertain, but all the same it rather 

 distinctly points to a high phase (266°), as did the results of the 

 balloon observations (10 km. 317°), and it adds weight to the un- 

 expected result that the phase after originally decreasing up to 

 4 km., would increase higher up, at first slowly and afterwards 

 more quickly. 



The question now arises, how do the results obtained agree with 

 the theoretical ones? 



Margules was the first to proffer a reasonable explanation of the 

 phenomenon of the semi-diurnal barometrical oscillation, by demon- 

 strating that the period of oscillation of an infinitely thin atmo- 

 spheric shell comes ' near 12 hours, and consequently will be 

 excited by the daily disturbance of the temperature. He therefore 

 proceeds from the oscillation of the temperature and deduces baro- 

 meter and wind oscillations, but by this process he finds the phases 

 vastly different to those observed. 



This is attributed by Goi.d^) to the neglecting of the vertical 

 velocities, as a consequence of dealing with an infinitely thin atmo- 

 spheric shell, assumed in order to overcome the otherwise unsurmount- 

 able mathematical difficulties. 



Hence Gold proceeds on the contrary from the barometer varia- 

 tion and approximately solves the variations of temperature and 

 wind (horizontal as well as vertical.) 



^) FiGEE, Observations Batavia Observatory. Vol. XXX. App. II. 



-) By applying the "difference method" afterwards has been found 42 Sin{2^ 263°) 



3) Phil. Mag. Vol. 19. 



