820 j\Ir. W . S. Jevuiis un the Semidiurnal 



amount of the elastic force of vapour, as indicated by a hygro- 

 meter at the surface of the earth, from the total pressure of the 

 atmosphere, and considering the remainder as the gaseous pres- 

 sure. I have attempted to prove that this practice is erroneous, 

 and can have no jjossible use or meanin(j ; so that the theory ia 

 question will thus be without foundation. But even if we adopt 

 the true mode of estimating the barometric effects of the varia- 

 tion of humidity, I have brought some results to show that no 

 exjilanation is obtained of the oscillations of the barometer. 



12. It will not have escaped my readers that, as a further 

 consequence of my arguments, the weight of the atmosphere 

 must be diminished by the weight of water falling from it as 

 rain, and that the barometer must be thereby affected in some 

 place or other. Tlie precipitation of a foot depth of rain over 

 the whole surface of the earth must cause a uniform fall of the 

 barometer to the extent of '088 inch. But the varied conditions 

 under which evaporation and the fall of rain take place, as well 

 as the many great and complex efl'ects which the removal of in- 

 comprehensibly vast bodies of water from place to place must 

 occasion, form part of the general mechanical problem oj the atmo- 

 fjjhere in whicli meteorolo:;y really consists. I will only notice 

 that, in a previous paper " On the Forms of Clouds " (I'hil. Mag. 

 April 1858), 1 have shown this same motive Jurce, viz. the dimi- 

 nution of weight and pressure produced by the separation of 

 cloud-particles and rain, to be the probable cause of those mo- 

 tions between masses of air which constitute the cirrus, the 

 cumulostratus, and even the thunder-cloud. 



13. Probable explanation of the semidiurnal oscillation of the 

 barometer^. — The doubly-pointed curve which the barometer de- 

 scribes during the day, simply indicates a double undulation occa- 

 sioned in the atmos])here by the disturbance of the sun^s rays. 

 This may be explained as follows : — Let fig. 1 represent an atmo- 

 sphere reposing on a horizontal 

 surface, divided into imaginary 

 columns. If the centre column 

 A be subjected to the action of 

 the sun's rays for any given 

 period, say six hours, the air 

 Miil increase in volume and will 

 overflow upon the adjoining co- 

 lumns B and B'. The baro- 

 metric pressure will therefore 

 diminish at the base of A and increase at the bases of B and 

 B', so that undulations will exist as shown by the line 1°; and 



* Being without means of reference, I am unaware whether this theory 

 is entirely new; I have "ot -'vs<*.lf seen i* -^'n where stated. 



