One of the most important results of Newton's First Paper on Light and Colours (1672) and Opticks (1704) is that 'the light of the Sun consists of Rays differently Refrangible', or that sunlight is a 'heterogeneous mixture' of 'Rays differently Refrangible'. In agreement with his 'hypotheses non fingo' assertion, Newton claimed to prove this from'phenomena' in his one and two prism experiments.
In the first experiment a sun light ray was passed through a prism. On the other side of the prism the colour spectrum is displayed and the different degree of refrangibility (or wavelength in modern terms) for each colour can be observed. In the second experiment Newton used two prisms. As before, the spectrum is produced by the first prism but by the use of a screen with a hole at the exit of the first prism, all colours are blocked but one which then passes through the second prism. No new colours were obtained, only the original one.
From the first experiment Newton claimed that sun light is made up of a mixture of differently refrangible rays, and from the second one that the degree of refrangibility is an intrinsic property of each ray and can not be modified. For Newton this result is a theorem (Theorem II, Book, I, Part I of the Opticks) and he states that "the prooffollows from experiments". The second experiment was crucial to counter objections according to which the colour spectrum from the first prism could have been created within the prism. Here there are two interesting points1:
i) The observation that red rays are less refracted than
blue ones was obtained in prisms of different material e.g. flint glass,
water, crown glass. Does it follow that this holds for any transparent
medium? Is this just enumerative induction? And,
i) What if there were a transparent medium - call it
"magic glass"- that Newton had not investigated and which reversed the order
of refraction? That is, a medium in which red rays are more
refracted than blue rays. As a result ofsome experiments using
water optical elements that we know Newton did, the last question may have
indeed been investigated by him and is what here we call "Newton's
missing experiment"2.
As we know3 in any substance the index of refraction n, is a function of angular frequency ω, and the change of refractive index with frequency dn/dω is called dispersion. In 'normal dispersion' the index of refraction n(ω) increases with ω (or diminishes with wavelength λ, since 2πv = λω, where v is the speed of light in the substance). In normal dispersion if white light passes through a glass prism the blue constituent will have a higher index than the red and will therefore be deviated through a larger angle.
However, due to their internal structure, all materials exhibit absorption at certain resonant frequencies. For glasses these resonant frequencies typically occur at wavelengths of about 100nm (well in the ultraviolet and outside our eye detection capability) and this is the reason why we are used to dealing mostly with normal dispersion. In the regions immediately surrounding the resonant frequencies, called absorption bands, the dispersion dn/dω is negative and the process is spoken of as anomalous (i.e. abnormal) dispersion. That is, in normal dispersion (within a region of normal dispersion) smaller wavelengths (higher frequencies) have larger indices of refraction whereas in anomalous dispersion (within a region of anomalous dispersion) larger wavelengths (lower frequencies) have larger indices of refraction. Since all substances possess absorption bands somewhere within the electromagnetic frequency spectrum the term anomalous dispersion, is certainly a misnomer. As already said, for glasses and many other substances the absorption bands lie outside the visible region, some exceptions are iodine vapour and fuchsine dye. It is known4 that anomalous dispersion was first observed in about 1840 by Fox Talbot and the effect was christened in 1862 by Le Roux, however his work was forgotten and eight years later rediscovered by C. Christiansen.
It is interesting to note that in order to observe in a prism 'something somehow looking like anomalous dispersion' (i.e. that the red rays will be deviated through a larger angle than the blue Rays) it is not necessary to have a prism made out of a fancy anomalous absorption material. This can easily be done for example with an air prism immersed in water or (more difficult to build) an air prism inside a glass medium. What is important for the sake of the effect we wish to observe is not only the kind of absorption we have (normal or anomalous) but the quotient of the refractive index nmedium/nprism. In most circumstances we have air as medium (n = 1) and a glass prism (n > 1), however in order to observe apparent anomalous dispersion we need only invert the situation, having for example a water or glass medium (n > 1) and an air prism (n = 1).
It is known that Newton used to keep to himself many results of his research, so even if he did not know about materials presenting anomalous dispersion, he may or may not have done the sort of experiments just mentioned with air prisms in water. What would have been the difference for Newton's conclusions if he had also done an experiment in air with a fuchsine filled prism or in water with an air prism? We can only speculate about this question and about "Newton's missing experiment".
I believe that in any case Newton conclusion 'from phenomena' would have been the one previously stated by Worrall1, i.e. that "the degree of refrangibility would instead be a relational affair between a type of ray and a type of transparent material", which is consistent with today's scientific knowledge. On the other hand, the implications of these experiments for Newton scientific methodology are very important1,2.
To carry out Newton's experiments with water prisms, as well as "Newton missing experiment" (air prisms in a water media), can be a very instructive and interesting experience for any student both from the scientific and the historic point of view.