Lastly, a large quartz plate is exceedingly good as a superposition" plate, as the colour may be changed by rotation of the analyser till the most pleasing effect is produced. It is somewhat strange that either quartz plates, or substitutes for them constructed of two mica films as hereafter described, are not more used for microscopic superposition work, instead of a complicated set of films in expensive mountings.

149. Other Rotatory Crystals.-Some other crystals besides quartz have the power of rotating a polarised beam. Amongst these may be mentioned cinnabar, periodate of soda, sulphate of strychnia; and there are a few others. Chloride of sodium (common salt) has sometimes the property in all directions, therein resembling fluids.

150. Electro-magnetic Rotation.-Faraday discovered that a cylinder of glass placed with its ends axially between two poles of a powerful magnet, or surrounded by a helix traversed by an electric current, similarly rotated the plane of polarisation, differentiating the two colours of a bi-quartz. His "heavy" glass (borate of lead), and some peculiar forms of flint-glass, give the most effect; good common flint having a rotatory power about half that of the "heavy" glass. Further experiments have shown that the electromagnetic force rotates the beam in some degree through nearly all, if not all, liquids, and even gases, thus opening up many most interesting questions of molecular physics. Dr. Kerr further discovered that a beam of plane-polarised light reflected from the polished pole of an electro-magnet, also showed effects of rotation whenever the iron was magnetised by the passage of a current.

But there is this remarkable difference between electromagnetic rotation and other forms of the same phenomenon. If we send the ray through a quartz which rotates it to the left of the observer, and then reverse the quartz, the ray is still rotated in the same direction. Consequently, if we

reflect the ray back again through the same quartz, the rotation conferred by the first transmission is exactly reversed by the second: the circular waves go back upon their former paths. In glass or fluid whose rotary property depends upon electro-magnetic action, on the contrary, the rotation is repeated in a reflected ray; and whichever way the ray is transmitted, and as often as it is transmitted, follows the direction of the current.

151. Rotation and Molecular Constitution.-We have seen that a limited number of crystals possess rotatory power. Nearly all of them are found both right-handed and left-handed. They are all either uni-axial or have no double-refraction at all. And, as a rule, if not universally,' those of them which are soluble in any fluid, lose the property of rotation when so dissolved. We are driven to the conclusion, that in the case of crystals the property of rotation is dependent upon crystalline structure; or, in other words, upon the arrangement of groups of molecules. This supposition is confirmed by the fact that in most of such substances the ray of light must pass through them in one particular direction-that of the single optic axis-to be rotated. And it is still more strongly confirmed by the successful production of rotatory mica-combinations as hereafter described (§ 157).

We might suppose the same of fluids-that is, that the property depended upon the arrangement of molecular groups were rotation confined to the fluid state. But experiment proves that when rotary fluids are volatilised, the vapour exerts sensibly the same rotatory power, for equal weights, as the fluid. We are, therefore, forced to the conclusion that in this case the property depends upon the single molecules of which the gaseous matter is composed. As analogy compels us still to regard it as dependent upon

1 I am not sure that the rule is quite universal.

some "structure," we are driven to the further deduction that here rotation depends upon the grouping of atoms into the molecule. It has, indeed, been disputed whether in fluids there really is any real circular double-refraction, as in quartz, whose two circular waves were so triumphantly separated by Fresnel; and Dove's experiments failed to decide the matter.' But the primary phenomena which led Fresnel to suppose it in quartz exist in the case of fluids; he himself, in a paper presented to the French Academy on March 30, 1818, reports experiments which he held to bear out this view; and my own experiments in the production of spiral figures (see Chapter XVI.), in which quartz is successfully replaced by either rotatory fluids or mica combinations, are very strong evidence in favour of the existence of similar circular waves.

2

Regarding the molecule, then, as the source of the property, it is surely strangely significant to find that all, or nearly all, the substances possessing rotatory power in solution or vapour are complex carbon compounds. Many beautiful and refined researches by Pasteur, Van't Hoff, and others, have moreover shown that in all such rotatory substances there are either one or more asymmetrical carbon-atoms ;2 but on the other hand, there are many substances which contain such asymmetrical atoms in which no rotation has yet been observed. Against this, however, is to be set the remarkable fact that a tartaric acid with no rotation has been separated into two varieties of opposite rotations, whose combination makes the neutral form. These two forms of rotatory tartaric acid, though in most reactions and in chemical constitution identical, differ in some reactions, 1 Poggendorr's Annalen, cx. 290.

2 I believe this statement correctly represents our knowledge of the matter at the time it is written. Should any exceptions have been discovered, or in future be so, these will not much affect the physical aspect of the fact as a general rule.

and crystallise in right-handed and left-handed forms (§ 144). There are a few other substances in which the inactive and one rotational form, but not the other, are known; and some -especially several ethereal oils-in which opposite rotations, but not the neutral form, are known. Some of these forms have been obtained with great difficulty; and hence the theory has been advanced, and is highly probable, that all carbon compounds containing asymmetrical atoms really are composed of molecules whose atoms are arranged helically, but that in the inactive ones two such molecules of opposite rotations are combined in a kind of sexual combination, and have not yet been separated.

It is further to be observed, that substances which possess rotatory power in fluid or solution, may often be varied or reversed in power by the addition of other fluids or substances. And it is still more remarkable, that when capable of crystallisation, they crystallise almost, if not quite, invariably as bi-axials (see Chapter XV.), which itself is a strong evidence (see § 170) of unsymmetrical atomic structure. In crystallising they appear to lose their rotatory power; but as this would be, with our present means of observation, overpowered in any direction of the ray by the far stronger ordinary double refraction, this can hardly be held to be proved. If the crystals are melted, however, so as to destroy the crystalline structure and restore the amorphous condition of a fluid-as, if a crystal of sugar be fused-the rotatory property remains or is restored. Heat will also convert some rotatory substances into inactive forms; as if there were a tendency under the influence of heat vibrations for a number of molecules to pass into the form of the opposite sex, as it were; the two kinds afterwards uniting in pairs. But enough has been said to show the exceeding interest of the subject, and the light it is likely to throw upon the nature of atomic grouping and the structure of molecules.

CHAPTER XIV.

CIRCULAR AND ELLIPTIC POLARISATION.

Fresnel's Rhomb-Composition of two Rectilinear Vibrations into a Circular one- -Quarter-Wave Plates-Other Methods of Producing Circular Polarisation-Rotational Colours of Circularly-Polarised Light-Reusch's Artificial Quartzes-Behaviour of Quartz in Circular Light-Phenomena of Thin Films when Analysed as well as Polarised Circularly.

152. Fresnel's Rhomb.-We have thus examined the effects of contrary circular orbits as exhibited in quartz; but circular orbits, or circular polarisation, may be produced in other ways, which furnish a most conclusive and elegant proof of the truth of the Undulatory Theory, showing as they do how the motions we are dealing with answer to every process to which they can be subjected. Fresnel not only tested his theory of the phenomena by separating the two actual rays in quartz; but he further calculated from his mathematical conceptions, that if he constructed a rhomb of glass with parallel faces so disposed that a ray (A B, Fig. 170), was "totally reflected" twice within it at an angle of 54° 37′ as shown; if that ray was planepolarised in a plane inclined at 45° to the reflecting surfaces, it would be as it were so divided and spun round by the relation of the reflecting surfaces to its vibrations, as to