448 Astronomical and Nautical Collections. 



right line, we shall find .0599 for the ordinate of that line 

 corresponding to the intermediate point ; or, in other 

 words, if the dark stripe of the third order described a 

 right line, its distance from the geometrical shadow, at the 

 intermediate point, would be .0599, instead of .0866, as it 

 was found in the experiment : and the difference, .0267, is 

 about half as great again as the interval between the middle 

 points of the second and third bands at the same distance : 

 for this interval was only .01654 ; and it is evident that the 

 difference .0267 cannot be attributed to any difficulty in as- 

 certaining the exact situation of the point in question, since 

 an error of so great an extent would have implied the pass- 

 ing over not only a bright stripe, but another dark stripe, 

 and half of a second bright one. 



I have made many other observations of the same kind, 

 all of which tend to confirm the same singular result. But 

 the experiment now related is sufficient to place beyond 

 doubt the existence of a sensible curvature in the tra- 

 jectories which represent the passage of the external 

 fringes. [Newton himself had indeed noticed the same cur- 

 vature in the 9th observation of his third book, and had 

 inferred from it, *' that the light which makes the fringes 

 upon the paper is not the same light at all distances of the 

 paper from the knives."] 



It appears [however] that this remarkable result is very 

 difficult to reconcile with the system of emission : for the 

 most natural manner of explaining the external fringes, in 

 this system, would be to suppose, that the pencil of light 

 which nearly touches the margin of the opaque body under- 

 goes in its neighbourhood alternate dilatations and conden- 

 sations, which give rise to dark and bright stripes: but 

 then these different bundles of pencils, so condensed and 

 dilated, ought to proceed in a right line, after having passed 

 the body : for if we admit, in the Newtonian theory, that 

 bodies may exert very powerful attractions and repulsions 

 on the molecules of light, it has never been imagined that 

 these forces can extend their action to distances so consi- 

 derable as the dimensions of these trajectories would require, 

 since they exhibit a sensible curvature for the length of se- 



