Royal Institution. 317 



them as the results should seem to guide him. Too great im- 

 portance cannot be attached to experiments as distinguished from 

 observations. Thus, it may truly be said that the discovery of gra- 

 vitation is founded rather upon the experiments of Galileo and the 

 purely meclianical deductions from them, than u)>on the observations 

 of the planets, which alone would never have led to it. Still, how- 

 ever, our knowledge derived from experiment must be combined 

 with the observations : and now the question is, with what class of 

 observations shall we begin? Shall we attempt to explain the mean 

 state of magnetism, — or the diurnal inequalities, — or the capricious 

 inequalities? In the lecturer's opinion, this choice would depend 

 very much upon our judgement whether all these were to be ascribed 

 to causes of the same class or of different classes. He believed 

 himself that the causes of all were the same, varied in their effects 

 only by the specialities of the circumstances under which they acted. 

 They do not differ more than a broken sea among rocks differs from 

 a smooth swell in the open ocean ; — they do not differ more than 

 the trade wind, the monsoon, the land and sea breezes, the variable 

 winds, and the tornadoes difft-'r : yet no one doubts that these are 

 due to general causes of the same kind, modified by assignable 

 specialities of circumstance. If, then, it were supposed that the 

 causes of these classes of magnetic plisenomena were the same, the 

 class upon which he would propose to fix is the capricious disturb- 

 ances. In the other classes we probably see, at any one time, 

 only the result of an infinity of combined operations ; in this, we 

 see nature acting under the simplest circumstances. The compara- 

 tive minuteness and obscurity of these pheenomena is no argument 

 against their efficiency for the*^cientific inquiry. The general laws 

 of ordinary light had been known for centuries to many persons : 

 the facts of depolarization are not known to one in a million, and 

 have not been computed by one in a hundred millions ; yet it is 

 upon these that the undulatory theory of light in general is founded. 

 It would probably be found then, that the successful course would 

 be to confine the examination to a single momentary disturbance 

 which could be traced well through all the magnetic observatories 

 in the world (the instance to which the attention of the audience 

 had been previously directed, showed the absolute necessity of 

 limiting the disturbance to the shortest time possible), — and with 

 such lights as experiment could give to determine by some mathe- 

 matical process the locality of the disturbing cause. But no rule 

 could be given for the process to be used. This only we might 

 predict with certainty, that the investigation would be long and 

 troublesome ; but such investigations are not without their redeeming 

 pleasure, and no title in philosophy could be too high for him who 

 should bring the investigation to a successful end. 



