It is probable that the conditions of snow formation are more complex than might be imagined, familiar as we are with the conditions relating to the crystallisation of water on the earth’s surface. A great variety of animal, vegetable, and other substances possess a doubly refracting or depolarising structure, as: a quill cut and laid out flat on glass; the cornea of a sheep’s eye; skin, hair, a thin section of a finger-nail; sections of bone, teeth, horn, silk, cotton, whalebone; stems of plants containing silica or flint; barley, wheat, &c. The larger-grained starches form splendid objects; _tous-les-mois_, the largest, may be taken as a type of all others. This presents a black cross, the arms of which meet at the hilum (Fig. 190). On rotating the analyser, the black cross disappears, and at 90° is replaced by a white cross; another, but much fainter, black cross is seen between the arms of the white cross, no colour being perceptible. But if a thin plate of selenite be interposed between the starch-grains and the polariser, a series of delicate colours appear, all of which change on revolving the analyser, becoming complementary at every quadrant of the circle. West and East India arrow-root, sago, tapioca, and many other starch-grains, present a similar appearance; but in proportion as the grains are smaller, so are their markings and colourings less distinct. [Illustration: Fig. 190.--Potato Starch, under Polarised Light.] Molecular Rotation. For the purpose of studying the various interesting phenomena of molecular rotation, a few necessary pieces of apparatus must be added to the microscope. First, an ordinary iron three-armed retort stand, to the lower arm of which must be attached either a polarising prism or a bundle of glass plates inclined at the polarising angle; in the upper an analysing prism. The fluid to be examined should be contained in a narrow glass tube about eight inches in height, and this must be attached to the middle arm. If the prisms be crossed before inserting a fluid possessing rotatory power, the light passing through the analyser will be coloured. If a solution of sugar be employed, and the light which passes through the second prism is seen to be red, but on rotating the analyser towards the right the colour changes to yellow, and passes through green to violet, it may be concluded that the rotation is right-handed. If, on the contrary, the analyser requires to be turned towards the left hand, we conclude that the polarisation is left-handed. These phenomena are wholly distinct from those accompanying the action of doubly refracting substances upon plane polarised light. It is not easy to explain in a limited space the course to be followed in ascertaining the amount of rotation produced by different substances. Monochromatic light should be used. If we are about to examine a sugar solution with the prisms crossed, the index attached to the analyser must first be made to point to zero. The sugar