PRESIDENTIAL ADDRESS. 



587 



Office ill 1»79, but subsequently iiumeiical calculatiou wa« used iustoad.^ A cou- 

 siderable amount of labour bas been spent over tbe computation of Fourier 

 coefficients. Not many great generalisations bave flowed from this metbod up 

 to tbe present time. I bave no doubt that tbere is mucb to be done in the way 

 of classifying temperature conditions, for climatic purposes, by tbe analysis of 

 the seasonal variations. A beginning was made in a paper which was brought to 

 the notice of the Association at Glasgow. Tbe most striking result of tbe 

 IFourier analysis we owe to Hann, who bas shown that, if we confine our atten- 

 tion to the second Fourier coefficient of the diurnal variation of pressure— that is, 

 to the component of twelve-hour period— we get a variation very marked in inter- 

 tropical regions, and gradually diminishing poleward in both hemispheres, but 

 synchronous in phase throughout the 360 degrees of a meridian; The maximum 

 occurs along all meridians in turn about 10 a.m. and 10 p.m. local time. This 

 semi-diurnal variation with its regular recurrence is well known to mariners, and 

 we have recently detected it, true to its proper phase, in the observations at tbe 

 winter quarters of tbe ' Discovery ' ; small in amplitude indeed — about a thousandth 

 of an inch of mercury — but certainly identifiable. 



The reality of this variation of pressure, common to the whole earth, cannot 

 be doubted, and, so far as it goes, we may represent it (if indeed we may 

 represent pressure differences as differences in vertical heights of atmosphere) as 

 tbe deformation of a spherical atmosphere into an ellipsoid, with its longest axis 

 in the Equator pointing permanently 30° to tbe west of the sun. Its shortest 

 axis would also be in the Equator, and its middle axis would be along the polar 

 axis of the earth. Somehov? or other this protuberance remains fixed in direction 

 with regard to the sun, while the solid earth revolves beneath it. Whatever may 

 be tbe cause of this effect, obviously cosmical, and attributable to the sun, at 

 which it indirectly points, its existence has long been recognised, and further 

 investigation only confirms the generalisation. It is now accepted as one of the 

 fundamental general facts of meteorology. 



Professor Schuster, for whose absence from this meeting I may venture to 

 express a regret which will be imanimous, has already contributed a paper to the 

 Royal Society pointing out the possible relations between tbe diurnal variations 

 of pressure and those of terrestrial magnetic force. Going back again to the 

 ubiquity of tbe application of the relation of pressure and wind, in accordance 

 with the dynamical explanation of Buys Ballot's law, we should expect the effect 

 of a pressure variation that has its counterpart in that of terrestrial magnetism to 

 be traceable also in wind observations. 



Mr. J. S. Dines has just given me particulars of the discovery of that effect in 

 the great air-current, the variations of which I have called the pulse of the atmo- 

 spheric circulation— I mean the south-east Trade Wind, the most persistent atmo- 

 spheric current in the world. It is difficult as a rule to get observers to pay much 

 attention to that current, because it is so steady ; but in 1801 the Meteorological 

 Council set up an anemometer at St. Helena, in the very heart of tbe current, and 

 we bave just got out the results of the hourly tabulations. When the observa- 

 tions for the hours 1 to 24 are grouped separately for months, so as to give the 

 ■vector resultants for each hour and for each month, it appears that tbere is 

 a conspicuous semi-diurnal variation in the current, which shows itself as a closed 

 polygon of vector variations from the mean of tbe day. 



The month of April gives the most striking diagram of the twelve. It 

 displays the superposition of two practically complete dodekagons, one a large 

 one, completing its cycle from 5 a.m. to 5 p.m., the other a small one, from 5 p.m. 

 to 5 A.M. The resultant wind for the whole day is very nearly south-east, and 

 practically remains so for all the months of the year, the monthly variation of 

 resultant wind being confined to a change of velocity from about thirteen miles 

 per hour in May to about twenty miles per hour in September. 



If, instead of combining the south and east components to form a vector 

 diagram, we plot their variations separately, the semi-diurnal variation in each is 

 plainly marked ; and the calculation of its constants shows that its amplitude is 

 about three-quarters of a mile per hour in the south, and rather less in the east 



