1876.] 



Presidents Address, 



349 



For the purpose of bringing one class of the observations on which a 

 scientific meteorology must be based into a form suited for the application 

 of theoretical tests such as these, the harmonic analysis seems to supply 

 the necessary means. This method may be familiarly explained as having 

 for its object to break up any observed series of quantities representing a 

 recurring phenomenon, such as the diurnal or annual variations of tem- 

 perature or atmospheric pressure, into other series so arranged that each 

 observed quantity shall be conceived to be an aggregate consequence of a 

 number of different series of variations from the mean value — the first of 

 such series being completed but once in the whole epoch under considera- 

 tion, the second recurring twice, or being completed in half the epoch, the 

 third recurring thrice, or in one third of the epoch, and so forth. 



The arithmetical computations requisite for thus transforming perio- 

 dical observations being very laborious, Sir W. Thomson (adopting an idea 

 of his brother's. Prof. J. Thomson) has proposed to construct a machine 

 that shall perform the calculations with the aid only of the graphical pro- 

 jection of the curve resulting from the recorded observations — an 

 illustrative model of which he exhibited to the Society in the course of 

 last Session, and an account of which has been published in oui' ' Pro- 

 ceedings.' 



It is well to remark that this treatment of meteorological or other 

 observations gives no direct aid in referring the phenomena to physical 

 causes, and is only to be regarded as a means of bringing them into a 

 shape in which they can be compared with theoretical formulae or dyna- 

 mical or other hypotheses. It has too long been thought that the arith- 

 metical manipulation of the results of meteorological observation was a 

 sufficient end to be attained ; and too often the necessity for seeking for 

 the efficient causes of the phenomena has been lost sight of. An alto- 

 gether useless refinement has also frequently been insisted upon in re- 

 cording observations of what, in a scientific sense, may be termed insig- 

 nificant details ; and a wholly illusory appearance of acciu'acy has been 

 aimed at, far beyond what can in fact be attained. The true conception 

 of averages is, in meteorological calculations as in many others, very often 

 missed, and mean results are exhibited which have no real signification. 



The relation of meteorology to physics may be compared to that of the 

 natural-history sciences to physiology. Physics include the study of the 

 forces the operation of which on the atmosphere gives rise to meteorolo- 

 gical phenomena. The intelligent application of physical research is un- 

 questionably one of the most necessary elements in the satisfactory pro- 

 gress of this science ; and Professor Tyndall's study of the action of the 

 air and other gases in relation to radiant heat affords an excellent illus- 

 tration of the manner in which experimental investigation contributes to 

 the knowledge required to explain atmospheric phenomena. 



The prodigality of nature in supplying the germs from which life on 

 the earth is sustained, and the comparatively extremely small proportion 



