Upon the calculation of the refraction and reflection from a Ball of Water or Glass, we have much the same Phaenomena, namely, an obliquity of the undulation in the same manner as we have found it here.  Which, because it is very much to our present purpose, and affords such an Instancia crucis, as no one that I know has hitherto taken notice of, I shall further examine.  For it does very plainly and positively distinguish, and shew, which of the two Hypotheses, either the Cartesian or this is to be followed, by affording a generation of all the colors in the Rainbow, where according to the Cartesian Principles there should be none at all generated.  And secondly, by affording an instance that does more closely confine the cause of these Phaenomena of colours to this present Hypothesis.

And first, for the Cartesian, we have this to object against it, That whereas he says (Meteorum Cap. 8.  Sect. 5.) Sed judicabam unicam (refractione scilicet) ad minimum requiri, & quidem talem ut ejus effectus alia contraria (refractione) non destruatur:  Nam experientia docet si superficies NM_ & NP (nempe refringentes) Parallelae forent, radios tantundem per alteram iterum erectos quantum per unam frangerentur, nullos colores depicturos_; This Principle of his holds true indeed in a prisme where the refracting surfaces are plain, but is contradicted by the Ball or Cylinder, whether of Water Or Glass, where the refracting surfaces are Orbicular or Cylindrical.  For if we examine the passage of any Globule or Ray of the primary Iris, we shall find it to pass out of the Ball or Cylinder again, with the same inclination and refraction that it enter’d in withall, and that that last refraction by means of the intermediate reflection shall be the same as if without any reflection at all the Ray had been twice refracted by two Parallel surfaces.

And that this is true, not onely in one, but in every Ray that goes to the constitution of the Primary Iris; nay, in every Ray, that suffers only two refractions, and one reflection, by the surface of the round body, we shall presently see most evident, if we repeat the Cartesian Scheme, mentioned in the tenth Section of the eighth Chapter of his Meteors, where EFKNP in the third Figure[9] is one of the Rays of the Primary Iris, twice refracted at F and N, and once reflected at K by the surface of the Water-ball.  For, first it is evident, that KF and KN are equal, because KN being the reflected part of KF they have both the same inclination on the surface K that is the angles FKT, and NKV made by the two Rays and the Tangent of K are equal, which is evident by the Laws of reflection; whence it will follow also, that KN has the same inclination on the surface N, or the Tangent of it XN that the Ray KF has to the surface F, or the Tangent of it FY, whence it must necessarily follow, that the refractions at