II. _Halos._ Produced by the light of a heavenly body striking the surrounding air and forming a circle as a stone does when thrown into a pond. Formed far away from the heavenly body and comparatively near the earth in the region of the wind. Require a particular state of the atmosphere neither too dense nor too thin. More frequent at night than day for this reason: by day the sun rarefies the air too much by its heat. Method of dissipation gives indication of wind or rain. Calmness a condition of formation, as in the analogous case of water 12 III. _Rainbows._ Generally by day, produced by inequalities of surface and density in clouds. Another species seen in a burst pipe or a fuller at work. Various explanations. Light and shade will not explain the varied colours. Some explain the rainbow as a confused reflection of the sun from individual drops of rain: every bounded surface, large or small, thus reflects--fish-pond and dew-drop equally. Aristotle attributes the confusion of colours to weakness of human sight; parallels may be found in persons whose sight is abnormally weak. As the innumerable drops, apparently without intervals, fall, human vision fails to distinguish severally the reflections of the sun, which thus become blended and confused. Vision is similarly deceived in the case of an oar in water, apples in a glass globe, etc., even in the size and movements of the sun himself. At any rate the rainbow requires both sun and cloud, and these opposite to each other. These two in operation produce the varieties of colour 16 IV. That the rainbow is an image is shown by the relation of sun to cloud in position, by the rapidity of formation and dispersion. Artemidorus’ explanation of the shape of the cloud (concave), and the consequent position of the red in the rainbow 22 V. Arguments to show that the cloud is coloured by the sun, like a dove’s neck or a peacock’s tail, and that the rainbow is not a _reflection_ of the sun. The position (opposite) would be equally necessary in this case. Answer to this contention by Posidonius. The colour effects. Author agrees with Posidonius’ position but not his arguments. The only _proof_ is the geometrical one 23