1876.] 



Variations of the Barometer in India. 



27 



The discussion of four years' observations gave for the length of the 

 period 25'86 days, and mean amplitudes of oscillation (as repre- 

 sented by the term a sin 6) between 0*01 and 0*032 inch in different 

 years. 



These movements could not be traced in the observations made at 

 Hobarton or Makerstoun, where the daily changes of pressure are in- 

 creased tenfold by other causes, and probably by other modes of action 

 of the same cause ; but some of them were still found to exist at St. 

 Helena*. It is not intended at present to enter into the consideration 

 of the 26-day period (which will show, however, the existence of the 

 other cause of var\dng pressure alluded to), but to examine all the 

 various changes of barometric pressure within a given area and during 

 a limited period of time. When stations are chosen which lie nearly in 

 the same parallel of latitude or great circle of the sphere, it might be 

 supposed, by the purely gravitational theorist, that, however great the 

 distance, simultaneous variations of pressure are due to the rhythmical 

 expansion and overflow of the atmosphere ; in the following discussion 

 it will be seen that this suggestion cannot be made use of. 



Hourly observations of the barometer made at the three following 



i 



Simla 



\ ' 





72GO 



\7 





Madras 















stations during the months of January, February, and March, 1845, have 

 been employed in this investigation t : — 



* See 'Comptes Rendus de I'Academie des Sciences,' t, 75, pp. 16, 121. 



t In the note already cited, which appeared in the ' Comptes Eendus,' the obserra- 

 tions for the three months January to April were employed because they showed in 

 the most marked manner the large oscillations of the 26-day period ; the daily means 

 then used were those for the Gottingen astronomical day, given with the printed obser- 

 vations to correspond with the daily means of the magnetic variations. These means, 

 however, have the disadvantage that one day in every week is made up of the part of 

 the astronomical day occurring on Saturday and the part occurring on Monday. In 

 the present note the means for the civil days have been computed for Singapore and 

 Madras. General Boileau has given the means for the civil days in the work cited 

 below. It is due to the recent publication of his valuable series of observations 

 that I have been able to include so large an area in my investigation. 



