Twilight is usually reckoned to last until the Sun has sunk 18° below the horizon, but the question of its duration depends on where the observer is stationed, on the season of the year, and (in a slight degree) on the condition of the atmosphere. The general rule is that the twilight is least in the tropics and increases as the observer moves away from the equator towards either pole. Whilst in the tropics a depression of 16° or 17° is sufficient to put an end to the phenomenon, in the latitude of England a depression of from 17° to 20° is required. As implied above, it varies with the latitude; and as regards the different seasons of the year, it is least on March 1 and October 12, being three weeks before the vernal equinox and three weeks after the autumnal equinox. The duration at the equator may be about 1 hour 12 minutes; it amounts to nearly 2 hours at the latitude of Greenwich, and so on towards the pole. At each pole in turn the Sun is below the horizon for 6 months, but as it is less than 18° below the horizon for about 3½ of those 6 months it may be said that there is a continual twilight for those 3½ months. Something of the same sort of thing as this occurs in the latitude of Greenwich, for there is no true night at Greenwich from May 22 to July 21, but constant twilight from sunset to sunrise, or 2 months of twilight in all. Though twilight at the equator is commonly set down as lasting about an hour, this period is there, as elsewhere, affected by the elevation of the observer above the sea-level. Where the air is very rarified, as at places situated as Quito and Lima are, the twilight is said to last no more than 20 minutes, and this would accord with the theory that where there is no air at all (_e.g._, on the Moon) there is no twilight at all. The greater purity and clearness of mountain air, rarified as it is, is another cause which contributes to vary by reducing the duration of twilight. It is sometimes stated that a secondary twilight may be noticed, and Sir John Herschel has spoken of it as “consequent on a re-reflection of the rays dispersed through the atmosphere in the primary one. The phenomenon seen in the clear atmosphere of the Nubian Desert, described by travellers under the name of the ‘afterglow,’ would seem to arise from this cause.” I am not acquainted with any records which throw light on these remarks of Sir John Herschel. The phenomenon of twinkling is a subject which has been much neglected, possibly on account of its apparent, but only apparent, simplicity. The familiar verse of our days of childhood— “Twinkle, twinkle little star, How I wonder what you are, Up above the earth so high, Like a diamond in the sky,” contains even in this simple form a good deal of food for reflection; whilst the new version— “Twinkle, twinkle little star, Now we’ve found out what you are, When unto the midnight sky We the spectroscope apply,” does so yet more.