[Illustration: FIG. 15]
But that the meaning of this Experiment may more clearly appear, it is
to be considered that the Rays which are equally refrangible do fall
upon a Circle answering to the Sun's Disque. For this was proved in the
third Experiment. By a Circle I understand not here a perfect
geometrical Circle, but any orbicular Figure whose length is equal to
its breadth, and which, as to Sense, may seem circular. Let therefore AG
[in _Fig._ 15.] represent the Circle which all the most refrangible Rays
propagated from the whole Disque of the Sun, would illuminate and paint
upon the opposite Wall if they were alone; EL the Circle which all the
least refrangible Rays would in like manner illuminate and paint if they
were alone; BH, CJ, DK, the Circles which so many intermediate sorts of
Rays would successively paint upon the Wall, if they were singly
propagated from the Sun in successive order, the rest being always
intercepted; and conceive that there are other intermediate Circles
without Number, which innumerable other intermediate sorts of Rays would
successively paint upon the Wall if the Sun should successively emit
every sort apart. And seeing the Sun emits all these sorts at once, they
must all together illuminate and paint innumerable equal Circles, of all
which, being according to their degrees of Refrangibility placed in
order in a continual Series, that oblong Spectrum PT is composed which I
described in the third Experiment. Now if the Sun's circular Image Y [in
_Fig._ 15.] which is made by an unrefracted beam of Light was by any
Dilation of the single Rays, or by any other irregularity in the
Refraction of the first Prism, converted into the oblong Spectrum, PT:
then ought every Circle AG, BH, CJ, &c. in that Spectrum, by the cross
Refraction of the second Prism again dilating or otherwise scattering
the Rays as before, to be in like manner drawn out and transformed into
an oblong Figure, and thereby the breadth of the Image PT would be now
as much augmented as the length of the Image Y was before by the
Refraction of the first Prism; and thus by the Refractions of both
Prisms together would be formed a four square Figure _p[Greek:
p]t[Greek: t]_, as I described above. Wherefore since the breadth of the
Spectrum PT is not increased by the Refraction sideways, it is certain
that the Rays are not split or dilated, or otherways irregularly
scatter'd by that Refraction, but that every Circle is by a regular and
uniform Refraction translated entire into another Place, as the Circle
AG by the greatest Refraction into the place _ag_, the Circle BH by a
less Refraction into the place _bh_, the Circle CJ by a Refraction still
24