Evidence in support of a Theory of LigJit. 281 



wliich is fairly open to the skill of philosophers to attempt, and 

 which, however difficult to imagine, it is impossible to say may 

 not be accomplished. All we contend for is, that the condition 

 just mentioned must essentially be fulfilled, in order to render 

 any solution sufficient. 



Guided by these considerations, we shall readily acknow- 

 ledo-e, that neither, on the one hand, is it necessary to ascribe a 

 demonstrative character to the evidence of etherial impulses for 

 the support of the undulations as a legitimate theory ; nor, on 

 the other, will the real or alleged exceptions in which the theory 

 does not apply, diminish the satisfactory nature of those expla- 

 nations which it does afford, or impair the certainty of the ex- 

 istence of some hind of vibratory motion for the production of 

 light. Nor, again, will such deficiencies in its completeness of 

 application even be real exceptions, until it has been demonstrat- 

 ed that the fundamental principles are absolutely incapable of 

 such modification as to include them. The course to be pur- 

 sued, in such cases, is only to investigate more closely the origi- 

 nally assumed principles, so as to ascertain to what modifica- 

 tions they may be subjected, without injuring their application 

 to the cases they already explain, and whether by such modifi- 

 cations they can be made to include the cases in question also. 



The theory of undulations is then applicable to the pheno- 

 mena of hght, simply in this way, that waves propagated under 

 certain conditions, and consistent with the principle of " the su- 

 perposition of small motions,'''' will explain with the greatest 

 exactness extensive classes of facts which optical experiment has 

 exhibited ; but of the physical character or conditions by which 

 the actual motions constituting the waves are determined, very 

 little need be assumed. A variety of such suppositions may be 

 made, by which all that is wanted for the actual representation 

 of the simpler phenomena will be equally well supplied. 



In Che earlier steps of optical inquiry, undulations produced 

 by vibrations of any kind would suffice. When the considera- 

 tion of polarised light, and the characteristic phenomena de- 

 pendent on it, were introduced, then that more restricted view 

 of undulations which involved the transverse vibration of Fres- 

 nel, was rendered necessary ; and such vibrations of equal pe- 



