The _dispersion_ of light is estimated by the variable angle formed by the red and violet rays which bound the solar spectrum;--or rather, it is the excess of the refraction of the most refrangible ray above that of the least refrangible ray. The dispersion is not proportional to the refraction--that is, the substances which have an equal mean refraction, do not _disperse_ light in the same ratio. For example, if we make a prism with plates of glass, and fill it with oil of Cassia, and adjust its refracting angle ACB, (fig. 31, p. 127,) so that the middle of the spectrum which it forms falls exactly at the same place where the green rays of a spectrum formed by a glass prism would fall--then we shall find that the spectrum formed by the _oil of Cassia_ prism will be two or three times _longer_ than that of the _glass_ prism. The oil of Cassia, therefore, is said to _disperse_ the rays of light more than the glass, that is, to separate the extreme red and violet rays at O and P more than the mean ray at _green_, and to have a greater _dispersive power_. Sir I. Newton appears to have made use of prisms composed of different substances, yet, strange to tell, he never observed that they formed spectrums, whose lengths were different, when the refraction of the green ray was the same; but thought that the dispersion was proportional to the refraction. This error continued to be overlooked by philosophers for a considerable time, and was the cause of retarding the invention of the achromatic telescope for more than 50 years.