When we look up into the sky on a cloudless day we behold a continuous canopy, the prevailing color of which is blue. This canopy is a veil that hides the starry hosts beyond, and its presence seems, at first sight, incompatible with the fact that the air is a transparent medium. We see the stars by night through the same intervening atmosphere. Why are they cut off from our sight by day? The answer to this question can, perhaps, best be made plain by a simple experiment. Place a lighted candle behind a sheet hung across a room not otherwise illuminated. The flame of the candle will be distinctly visible through the sheet. Next, let the room be brightly lighted, say with electric light or daylight. The candle can now no longer be seen through the sheet, owing to the bright illumination of the latter as compared with the feeble light of the candle.

In the atmosphere the counterpart of our sheet is a layer, several miles in depth, of minute particles, which by day are lighted up by the sun. Some of the particles are tiny dust motes, others are fine droplets of water or bits of ice, and the rest are the molecules of the atmospheric gases themselves. It is the light that comes to us from these particles that makes our eyes insensitive to the fainter light of the stars, and makes the sky itself a visible luminous vault. Next, why is the clear sky generally blue, rather than some other color? To answer this question, we must recall the fact that sunlight is made up of ether waves of many different sizes. In combination, these waves produce upon our eyes the sensation of white light. When they are separated, as by passing through a prism, the smallest waves—or, in more technical language, the vibrations of shortest wave length—register the sensation of violet, and the largest or longest waves that of red. The whole sequence of colors runs in the order violet, indigo, blue, green, yellow, orange and red (easily fixed in the memory by means of the word VIBGYOR, formed from the initial letters of these words).

Now the passage of sunlight through the atmosphere is obstructed to a certain extent, not only by suspended dust particles, but also by the molecules of the air. Let us consider, first, the effect of air molecules and of the finest dust particles, not much above molecular size. These tiny objects have different effects on light waves of different lengths. The longest waves are little disturbed by them, just as ordinary waves in water are little affected by a floating cork, for instance. The shortest waves are so small in proportion to the size of the obstacles that they are diffused or scattered by them, as a tiny ripple in water might be broken up by a floating cork. It is this diffuse light, of short wave length, that gives the sky its color. A large part of the violet and indigo light is lost by further scattering before it reaches the earth, leaving a preponderance of blue in the sky as we see it. When the air contains a considerable amount of suspended particles larger than those above considered—whether in the form of solid dust or crystals of ice or tiny droplets of water—light of all wave lengths is reflected by them, and the sky looks white or grayish.

On account of the action of atmospheric particles in filtering out the shorter light waves, as just described, sunlight becomes relatively rich in red and orange in passing through the air. When the sun is high, the path of the sunbeams to the earth is short, and the color of their light is but little affected. Near the time of sunrise and sunset, however, sunlight comes to us through a much greater extent of air, and the filtering process is much more effective. Hence the sunshine is both enfeebled and reddened when the sun is near the horizon. The diffuse light of the sky around the sun is filtered in the same manner, and therefore is commonly red when the sun is low.

A gray sunset sky after a clear day is due to the presence of water drops in the air, and indicates conditions favorable for rain, since, unless the air were saturated to a considerable altitude, the comparatively warm sunshine of the afternoon would favor evaporation rather than condensation of moisture. A gray sunrise sky has, as a general rule, just the opposite meaning. It often indicates the presence in the air of water drops formed on dust particles during the night, after the manner of dew, because the upper air has been dry enough to permit rapid radiation from the dust. These drops will be speedily evaporated by the rising sun, and the general dryness of the atmosphere will not favor further condensation. Several familiar weather proverbs are thus justified, e.g.:

There are many other interesting optical phenomena connected with sunrise and sunset, including, first of all, the morning and evening twilight. When the sun, or any other heavenly body, is only a little below a clear horizon, it is still visible, on account of the bending of its rays by the atmosphere. This lifting effect, known as astronomical refraction, amounts to about half a degree (at the horizon), which is about equivalent to the apparent diameter of the sun or moon. As the sun sinks farther below the horizon, in the evening, the only daylight that comes to us is that reflected from the upper levels of the atmosphere, which are still illuminated. This is called twilight, and it lasts until the sun is about 18 degrees below the horizon, when total darkness sets in. The period as a whole is sometimes called astronomical twilight, in distinction from the briefer period known as civil twilight, during which there is light enough for outdoor occupations; the latter lasts from sunset until the sun is about 6 degrees below the horizon. Morning twilight is more commonly called “dawn.”

An interesting succession of light and color effects is observed before and after sunset and, in inverse position and order, about sunrise. Considering sunset only: After the sun has sunk out of sight, a broad band of golden light, called the bright segment, is seen along the western horizon. Above this, in the western sky, appears a more or less circular expanse of rosy glow, known as the purple light. In the eastern heavens, after sunset, there rises steadily from the horizon the so-called dark segment, which is the blue or ashy shadow of the earth on the sky. This is bordered above by the pink or purplish antitwilight arch. As time goes on, the purple light in the west, after increasing in brightness for a while, finally sinks behind the bright segment; while in the east the rising dark segment encroaches upon and finally obliterates the antitwilight arch. Sometimes, in clear weather, there is a fainter repetition of these lights and colors (second purple light, etc.).

Among the Alps and other snow-capped mountains, these sunset and sunrise phenomena assume a particularly beautiful form, known as the Alpenglow. In fine weather, just before sunset, the peaks to the eastward begin to show a reddish or golden hue. This fades gradually, but in a few minutes, when the sun is a little below the horizon of the observer, but the peaks themselves are still bathed in direct sunlight, an intense red glow, beginning down the slopes, moves upward to the summits. This is identical with the antitwilight arch described above. Presently this glow is succeeded by an ashy tint, as the peaks are immersed in the rising shadow of the earth (the dark segment). Their rocks and snows assume a livid appearance, aptly described by the inhabitants of Chamonix, whence the phenomena in question are well seen on the summit of Mont Blanc, as the teinte cadavÉreuse. In ordinary weather darkness succeeds without any further notable phenomena, but occasionally there occurs a remarkable renewal of rosy light upon the peaks, known as the recoloration or afterglow. At Chamonix this is termed the “resurrection of Mont Blanc.” The afterglow has been variously explained, but it is probably due, mainly at least, to the reflection of the purple light in the western sky. Sometimes it lasts until an hour after sunset, and it passes away from below upward. On very rare occasions there is a second afterglow, presumably the reflection of the second purple light mentioned above. Similar phenomena are often seen in reverse order at sunrise.

A pretty phenomenon observed chiefly in the late afternoon and early morning consists of beams of light radiating from the sun, known technically as crepuscular rays. The beams are made visible by the presence of abundant dust or water droplets in the atmosphere, and the intervening dark spaces are the shadows of clouds. When the sun is above the horizon and the beams are directed downward, the phenomenon is popularly described as “the sun drawing water” and is regarded as a sign of rain. Sailors call these beams the “backstays of the sun,” and they have several other names based upon the legendry associated with them in different parts of the world. After sunset or before sunrise a fanlike sheaf of the beams often extends upward from the western or eastern horizon, respectively. The Homeric expression “rosy-fingered dawn” probably refers to this phenomenon. In all cases the apparent divergence of these beams is an effect of perspective, as they are really parallel. A rarer phenomenon is that of anticrepuscular rays, which appear to converge to a point opposite the sun. In this case the beams and shadows are projected entirely across the sky, but their paths can very seldom be traced in the upper part of the heavens because in this direction the observer’s line of sight passes through a comparatively shallow extent of dusty atmosphere.

An analogous phenomenon is seen in the shadows which near-by isolated mountain peaks frequently cast upon the sky opposite the sun at sunrise and sunset. Travelers have described such shadows cast by Adam’s Peak in Ceylon, Pike’s Peak in the Rocky Mountains, and Fujiyama in Japan. The phenomenon is said to be especially striking in the polar regions, where the air is often heavily charged with particles of ice.

One more optical phenomenon of sunrise and sunset that requires mention here seems to be comparatively little known to the nonscientific public, notwithstanding the fact that it has supplied the subject and title of a diverting novel by Jules Verne. The conditions required for its appearance are a clear and steady atmosphere and a sharply defined horizon, such as that of the ocean. At the instant the sun is appearing or disappearing, and when only a very small segment of its disk is visible above the horizon, this portion appears to be colored a bright emerald green, sometimes blending into blue. This transient phenomenon is known as the green flash. It is best explained as due to the different degrees of refraction undergone by rays of different wave lengths coming to us from the sun. The effect of refraction in elevating the solar image as a whole when near the horizon has already been mentioned. This effect is a little greater for the green and blue rays than for the orange and red. It is still more pronounced for the violet and indigo rays, but these are mostly sifted out of the solar beams in their long passage through the atmosphere when the sun is low. Hence at the upper edge of the solar image there is a narrow green or blue fringe, which is not, however, perceptible except when a screen is interposed between the eye and the bright image of the sun. A sharp horizon furnishes such a screen. Through a telescope it is possible, in suitable weather, to see the green flash—and also a corresponding “red flash” at the lower edge of the sun—by placing an opaque diaphragm in the focal plane of the object glass. Another explanation of the green flash—which could, however, account only for its appearance at sunset and not at sunrise—is that it is a physiological effect; the eye, fatigued by the reds and yellows that predominate in the light of the setting sun, sees an “after image” of complementary hue the instant after the real image has disappeared.

We now turn to a group of phenomena, also due to atmospheric refraction, which includes some of the most bizarre of optical illusions. The simplest of these phenomena consists of a slight apparent elevation of all objects in the surrounding landscape through terrestrial refraction, which is identical in principle with astronomical refraction and depends upon the difference in density, and hence of refractive power, of the air at different levels above the earth.

Normally the air decreases in density at a nearly regular rate with increasing altitude. Sometimes, however, this change in density is greatly modified by local effects of temperature. Over a cold surface of land or water the adjacent air may be abnormally dense, resulting in an unusually rapid decrease of density with altitude. Over a hot surface, as in the case of a desert under strong sunshine, the adjacent air may become so much rarefied that, for a certain distance upward, there is actually an increase of density with ascent, instead of the reverse. The rays of light coming to us from distant objects are bent in different directions and to various degrees by virtue of these abnormalities in the density of the atmosphere. The apparent positions of such objects depend upon the angle at which the light rays, coming from them, strike the eye of the observer. Sometimes the objects appear to be lifted far above their true positions (a phenomenon known as looming) and sometimes depressed far below them; and occasionally local irregularities in air density produce curiously distorted images of these objects.

Most of these strange effects are known collectively as mirage. There are many varieties. There is the “desert mirage,” first made famous through the experience of Napoleon’s soldiers in Egypt. There are mirages that suspend the images of remote objects in the sky; sometimes inverted, sometimes right side up. There is the lateral mirage, occasionally seen when one looks along the face of a heated wall or cliff. Lastly, there are the complex displacements and distortions of objects known as the Fata Morgana—a name originally applied to a phenomenon of this kind visible, on rare occasions, at the Straits of Messina, but now used generically for similar appearances in other parts of the world. Some of the finest examples of Fata Morgana are witnessed in the polar regions.

In the desert mirage an image of the lower part of the sky is brought down to earth and simulates the appearance of water, while the images of terrestrial objects, also depressed and inverted by the mirage, look like the reflections of the same objects upon the liquid surface. Humphreys says: “This type of mirage is very common on the west coast of Great Salt Lake. Indeed, on approaching this lake from the west one can often see the railway over which he has just passed apparently disappearing beneath a shimmering surface. It is also common over smooth-paved streets, provided one’s eyes are just above the street level.” The confusing and obscuring effects of the desert mirage were illustrated during the fighting between the British and Turks in Mesopotamia, in April, 1917, when, according to the report of General Maude, a battle had to be suspended on account of one of these optical disturbances.

The strong vertical contrasts in air temperature that occur in the polar regions produce many remarkable examples of mirage. The pictures and descriptions of those observed a century ago along the coast of Greenland by Captain William Scoresby, Jr., have become classical. A recent episode connected with mirage was the expedition sent north in 1913 to explore “Crocker Land,” which Peary believed he had sighted from an elevated point in Grant Land in 1906, and which for a time figured on all maps of the Arctic. The later explorers found no land at the place indicated, but they observed the same mirage that Peary had mistaken for distant hills and mountains. Currents of air of different densities produce, through their varying effects on atmospheric refraction, the twinkling or scintillation of the stars, as well as of distant terrestrial lights. Twinkling is much more violent near the horizon than near the zenith, and more pronounced on some nights than others. The shimmering of the air over heated surfaces faces and the “boiling” of celestial objects as seen in the telescope are analogous phenomena.

In the refraction phenomena that we have thus far considered the air is the medium in which the light rays are bent and distorted. In the production of the rainbow, light undergoes refraction, dispersion (separation of the spectral colors) and reflection by passing through drops of water in the atmosphere; especially falling raindrops.

The rainbow, perhaps because it is such a common sight, is seldom observed with careful attention. Hence few people realize that there are many varieties of this beautiful meteor, and various erroneous ideas about it are prevalent. The rainbow is always seen in the part of the sky opposite the sun—or the moon, in the case of the lunar rainbow—and is high in the heavens when the luminary is low, and low when the luminary is high. Generally less than a semicircle of the bow is visible, and never more, except from an eminence. (Aeronauts occasionally see a complete circle.) The outer border of the bow is red and the inner blue or violet. Contrary to popular belief and to statements sometimes found in reference books, it is almost never possible to distinguish all seven of the spectral colors in a rainbow; four or five is the usual limit.

The ordinary or primary rainbow has a radius of about 42 degrees at its outer edge. Very commonly a secondary rainbow is seen, concentric with the primary bow, and having a radius of about 50 degrees. The secondary is fainter than the primary, and its colors are in opposite order—red inside and violet outside. Additional bands of color, chiefly red and green, may often be detected adjacent to the inner edge of the primary bow and, less frequently, along the outer edge of the secondary bow. These are known as supernumerary bows. The space between the primary and secondary bows is somewhat darker than the rest of the sky. The common rainbows differ much among themselves in the number and purity of their colors, the width of the bows, etc., these differences depending especially on the size of the raindrops. The minute drops of a fog sometimes give rise to a bow that is almost devoid of color—the “white rainbow,” or “fog bow.” The rainbows produced by the moon commonly show little color, on account of the relative faintness of the light, but the brighter lunar rainbows are often very distinctly colored.

Reflected rainbows are sometimes seen upon a sheet of water; and again the image of the sun, as reflected by such a surface, may give rise to both primary and secondary rainbows in the sky, which appear to intersect those produced by the sun directly. A horizontal layer of water drops below the level of the observer’s eye occasionally produces the so-called horizontal rainbow. This may be formed over a bedewed field or other surface (the “dew bow”); or the drops may be those of a low-lying sheet of fog, or of water deposited on a floating film of oil, or, finally, actual raindrops, seen from an elevation, such as the summit of a mountain. Horizontal rainbows formed by rain have been seen from the Eiffel Tower.

The common saying,

is, on the whole, well justified for the following reasons: We see the rainbow where rain is falling, while the sun is shining in the opposite part of the sky. Our rainstorms usually come from the west and pass away to the east. A morning rainbow can only be seen in the west, and indicates that rain is approaching us. An evening rainbow (ignoring lunar bows) is seen only in the east, and shows that the rain area is receding from us, giving place to clear skies.

Ice crystals in the atmosphere, such as those composing the higher clouds, produce a great variety of optical phenomena, known as halos. Some phenomena of this class are common, others exceedingly rare. Moreover, there are several theoretically possible forms of halo of which observations have never yet been reported, so that halo observing can be recommended to the amateur meteorologist as offering opportunities for making interesting discoveries.

Halos take the form of narrow rings of definite angular size around the sun or moon (not to be confused with the coronas, of variable dimensions, described below), rings passing through the luminary, arcs in various other positions, and roundish spots of colored or white light. They may be seen separately or in combination. In rare cases, a dozen or more different forms of halo are visible at the same time, producing a most spectacular display. One of the most remarkable displays of this kind in the history of science was seen, in different degrees of development, over the eastern United States on November 1 and 2, 1913; an event which greatly stimulated interest in the study of halos in this country. Complex halos are quite common in the polar regions; where they are seen not only in the sky, but also in the air, charged with ice particles, close to the earth.

Whenever a thin veil of cirrus or cirro-stratus clouds overspreads the sky there is a likelihood that halos will be visible. Those formed near the sun, however, frequently pass unnoticed, on account of the dazzling brightness of that luminary. Smoked or tinted glasses greatly facilitate their observation.

The commonest halo is a circle of 22 degrees radius (the 22-degree halo) about the sun or moon. When formed by the sun it generally shows a distinct reddish inner border and traces of other spectral colors. The lunar 22-degree halo usually appears colorless. This halo is visible, in whole or in part, to the attentive observer about once in three days, on an average. Less common, but by no means rare, are the parhelia or “sun dogs” of 22 degrees (called paraselenÆ or “moon dogs” when formed by the moon), the beautiful circumzenithal arc, and a few other members of the halo family. Most forms of halo are so uncommon that their appearance is an event of some scientific importance.

The accompanying diagrams, by Dr. Louis Besson of the Observatoire de Montsouris, show the positions, with respect to the sun (or moon), of the majority of known halo phenomena. The upper diagram shows the halos that occur on the same side of the sky as the sun (or moon), and the lower those that appear on the opposite side. Most of these halos, when bright, show the spectral colors. The circumzenithal arc, h (commonly described, by the uninitiated, as a “rainbow”), and the parhelia of 22 degrees, e, e', are especially brilliant in their coloration. The parhelic circle, m, which sometimes extends entirely around the sky, is white, and so are a few of the rarer forms of halo.

The upper and lower tangent arcs of the halo of 22 degrees, c and d, undergo striking alterations, with changes in the altitude of the sun. When the luminary is more than about 40 degrees above the horizon, these two arcs become joined at their tips to form the circumscribed halo, and at still greater solar altitudes this halo contracts from an elliptical to a circular form, thus blending into the 22-degree halo as shown on the next page, where the solar altitudes corresponding to the different forms of the halo are indicated. The positions of the parhelia of 22 degrees, e, e', also depend upon solar altitude. When the sun is on the horizon these “sun dogs” are 22 degrees from the luminary, and therefore lie in the 22-degree halo; at greater solar altitudes they lie outside this halo.

The reader who wishes to acquaint himself further with the different forms of halo and the methods of observing them will find a comprehensive article on the subject (devoid of mathematical discussions) in the “Monthly Weather Review” (Washington, D. C.) for July, 1914.

The ice crystals that produce halos consist of hexagonal plates or columns, occasionally including complications of structure, such as pyramidal bases, combinations of plates and columns, etc. These have the well-known effect of prisms in refracting and dispersing light that passes through them. It is evident that there are many possible paths for the light rays through the sides and bases of such crystals, resulting in different deflections and corresponding differences in the forms and positions of the halos produced. The attitudes assumed by the crystals as they slowly sink through the air, and the oscillations they undergo, are further points to be considered in working out the theory of each form of halo by the application of the laws of optics. Nearly all the known forms have been fully explained. A few species of halo—notably the parhelic circle (called paraselenic circle when formed by the moon)—are due to simple reflection from the faces of the ice crystals, and not to refraction.

The last group of optical phenomena that we shall consider consists of those due to the process called diffraction, which occurs when light is bent around objects in its path, instead of passing through them, as in refraction. The process involves separation of the prismatic colors. The diffraction phenomena of the atmosphere are produced by the water drops of clouds and fog, or sometimes by fine dust.

Everybody is familiar with the nocturnal spectacle which Tennyson describes as

This diffuse reddish or rainbow-tinted circle is called a corona. It occurs about the sun as well as the moon (though not easy to see on account of the glaring brightness of the luminary), and also about street lamps and other terrestrial lights when viewed through a misty atmosphere. Unlike the halos, it has no definite angular size. It is usually only a few degrees in radius. Small coronas are produced by large water drops and large coronas by small drops, while the largest of all coronas, known as Bishop’s ring, is due to exceedingly fine dust in the atmosphere, and has been seen after great volcanic eruptions. In its commonest form the corona consists of a brownish-red ring, which, together with the bluish-white inner field between the ring and the luminary, forms the so-called aureole. If other colors are distinguishable, they follow the brownish red of the aureole (in the direction away from the luminary) in the order from violet to red; the reverse of the order seen in halos. Sometimes the sequence of colors is repeated three or four times.

Patches and fringes of iridescence are sometimes seen in the clouds at a greater distance from the luminary than that of the ordinary corona. Probably they are fragments of coronas of unusual size produced by exceedingly fine cloud particles.

Similar in appearance to the corona is the glory; a series of concentric colored rings seen around the shadow of the observer, or of his head only, cast upon a cloud or fog bank. Such a shadow, with or without the glory, constitutes the specter of the Brocken, often seen from mountain tops and from aircraft. The colored circles are sometimes called Ulloa’s rings, from the name of a Spanish savant who observed the phenomenon among the mountains of South America in the eighteenth century and has left us a vivid description of it.

The Brocken specter, though it owes its name to legends associated with the famous German mountain where witches were once believed to assemble on Walpurgis Night, is actually less frequently witnessed there than in many other parts of the world. Whenever the sun is low on one side of a mountain and a wall of mist arises from a near-by valley on the other, the mountaineer is likely to see his shadow upon the mist. If the latter consists of fine droplets of approximately uniform size, the colored rings will probably appear, and occasionally there is also a white fogbow outside of the glory. As all shadows cast by the sun taper rapidly (on account of the angular breadth of the solar disk), a well-defined Brocken specter can never be more than a few yards away from the observer. Its distance is, however, commonly overestimated—some observers have supposed it to be miles away!—and hence the erroneous idea prevails that the specters are of enormous size.

Rarely from a favorable point of vantage on a mountain, and very frequently from aircraft, the specter, instead of being seen on a vertical wall of mist when the sun is low, appears on a horizontal sheet of cloud below the observer when the sun is high. The aeronaut may thus observe the complete outline of his balloon or aeroplane, encircled with the rainbow tints of the glory. During the World War the appearance of the luminous rings was likened to the emblem painted on the wings of the Allied aeroplanes and was regarded by superstitious aviators as an omen favorable to their cause.

The glory is due to the light that is reflected back to the observer after penetrating the cloud or fog a little way and is diffracted by the superficial layer of drops in emerging.

The Brocken specter and the glory have occasionally been photographed.