Evidence in support of a Theory of Light. 279 



made nearly to coincide with each other in direction, should 

 produce absolute darkness, is a fact utterly contradictory to any 

 conception of light as caused by an impact of material particles. 

 If, indeed, the two interfering rays were always of different na- 

 tures, as e. g. a red and a violet ray, it might be an admissible 

 supposition that they might be capable of neutralizing each 

 other's effects; the production of light being supposed in some 

 sort analogous to electric or chemical action, which might be in 

 different states in different rays. But when we recollect that 

 the interfering rays are absolutely of the same nature, portions 

 indifferently taken from the same original pencil, all such sup- 

 positions are at once overthrown The only kind of mechanical 

 actions which it is possible to imagine capable of mutually de- 

 stroying each other, are such as belong to motions existing and 

 propagated in some medium, and whose result or effect in some 

 way constitutes light. That motion of translation, then, which 

 is necessary to be supposed in the radiation of light, must now 

 be understood as a motion of translation of an effect or residt of 

 other motions, and not as a simple translation of matter. And 

 the component or constituent motions, from which these result, 

 must of necessity be such as are capable of composition, and of 

 destroying each other. Thus we are compelled by the facts to 

 recognise the idea of some reciprocating kind of motions among 

 points distributed in space, whose result or effect propagated in 

 space, constitutes light. By whatever name we call them, and 

 by whatever cause, or in whatever manner, or in whatever sort 

 of medium we suppose them produced, these reciprocating mo- 

 tions are a neces.sary consequence from facts. They must be of 

 such a kind as shall be consistent with the established mathe- 

 matical doctrine, and which is termed the superpositions of small 

 motions, and capable of being expressed by the formulas which 

 that doctrine suppHes. And further, it is a matter of measure- 

 ment and calculation to determine the periods or intervals (or 

 by whatever name we designate them) by which these motions 

 are limited. Other classes of facts oblige us to superadd some 

 other characteristics of the sort of motion necessary to be sup- 

 posed. We frame such systems of these motions as can become 

 subject of calculation. The results are found to accord with a 

 vast number of experimental results. The argument from ac- 



