ZODIACAL LIGHT 97 o clock for several nights in succession." On reading this state ment, Baron Humboldt communicated to the Monatslerichte d. kon. prcuss. Akad. d. Wiss. some unpublished observations of his own on a similar phenomenon. These were to the effect that on the 17th and 18th of March the light was very bright in the west, and "we constantly perceived in the east (and this is beyond doubt a very striking phenomenon) a whitish light which was also of a pyra midal form. The latter augmented the brightness of the sky in a very striking manner." The light in the west was so conspicuous that "even the sailors were delighted with this double light." He noted, too, what is a very important fact, that the two lights set at the same time ; and there is much to be said in defence of his view that the eastern light was the reflexion of the true zodiacal light in the west, just as the eastern sunset glows are the reflexion of those in the west. The chief obstacle in the way of accepting this view is that it is difficult to believe that a light so faint as the zodiacal light could have a reflexion bright enough to be seen, and even to be seen distinctly. On the other side must be set the cir cumstance that the simultaneous glow at both horizons seems never to have been observed from great altitudes. This is easily under stood if the second glow is a mere reflexion ; whereas, if it comes from another source of light, it ought to be more conspicuous at high than at low levels. Position. The exact position of the axis of the zodiacal light relatively to the ecliptic has not yet been satisfactorily fixed. The extreme haziness of outline and the excessive fain tn ess of the light near the vertex make it quite impossible to use a telescope for measuring its limits. Most observers have tried to fix its position by tracing its outline on a star chart, while Prof. C. Piazzi Smyth employed two sights mounted equatorially. But even by these means no great accuracy can be attained, for the limits of the light can be traced only when the eye is quite unfatigued and when the light is looked at with averted vision. The difficulty experienced is well illustrated by the wide divergencies between the results of different observers divergencies not only in the extent of the light, which would be quite natural, but even in the direction of the axis of the cone. Thus, Captain Jacob, who observed in Madras in 1856- 57-58, found that his evening observations placed the vertex of the western cone in from 2 S. lat. to 6 N. lat., with an average of about 3 N. lat. His morning observations were less numerous and gave positions for the vertex of the eastern cone varying from J N. to 8 S. , with an average of about 2 S. Mr Backhouse from 418 observations found a mean deviation from the ecliptic of 2 06, while Captain Tupman, observing in the Mediterranean in August and September, found an inclination of no less than 20. He also found that the plane of the light did not pass through the sun. His observations, however, differ somewhat widely from those of most other observers. Mr A. Searle has made a very careful study of all the best published records. In his first paper more than 650 observations by Jones, Heis, Schmidt, and others are discussed, and he concludes that apparent changes in lati tude are mainly, if not entirely, produced by the effect of atmo spheric absorption, which affects the lower boundary more than the upper and to an extent depending on the inclination of the axis of the light to the horizon. In a second paper he discusses Jones s observations alone and comes to the conclusion that, after making allowance for the effects of absorption, there is some evidence that the zodiacal light, as seen during the second half of the 19th century, has had a more northern latitude near 180 long, than near long. It seems somewhat doubtful, however, whether the observations hitherto made are sufficiently accurate to justify this conclusion, and all that can be confidently asserted is that the observations made so far point to such a difference as being probable. Similar variations in the greatest elongation, the breadth, and the bright ness have been asserted to exist. Thus Mr J. F. Julius Schmidt, as the result of several years observations, finds a variation in the inclination of the axis of the cone of light to the ecliptic from 4 18 at the end of December to towards the end of March, a similar variation in the greatest elongation from the sun from 120 in January to 70 in April, and a further similar variation in breadth at corresponding elongations, with a minimum for all three about 30th March. It may bo pointed out that observations regarding the extent of the visible light are absolutely valueless unless accom panied by careful records of the clearness of the sky as evidenced by the visibility of faint objects, for all observers of the zodiacal light must have noticed that a thin film of mist, which hardly affects the colour of the sky or the visibility of the brighter stars, may almost completely obliterate the zodiacal light. No accurate photo metric observations seem to have been made on the strength of the light and observers have differed widely in their estimates. Some assert that it is much brighter than the Milky Way, while others and especially those who have observed in the tropics say that it is seldom so bright as the brighter parts of the Milky AVay. It is certainly often brighter than most of the Milky Way. Mr Searle deduces from Sir AV. Herschel s and Celoria s observa tions that the Milky AVay is about 2 magnitudes brighter than the mean brightness of the sky, and says that on this estimate the brighter parts of the zodiacal light would be commonly 3 or 4 magnitudes brighter than the surrounding sky. This is almost certainly an over-estimate. There seems to be very little doubt that the brightness undergoes periodic fluctuations, but no estimate can as yet be made of the length of the period. It is probable that the brightness has been below the average for the last few years, but is now increasing. Most observers have also reported rapid changes of brightness, or undulations, such as are seen in the aurora and in the tails of some comets ; but, as was pointed out by Olbers in 1833, these undulations must be produced in the earth s atmo sphere. In this connexion it may be mentioned that in observing the spectrum rapid flickerings, like waves moving along the spectrum, are often very marked. The Moon s Zodiacal Light. Several observers have recorded observations which appeared to show that the moon produced an appearance very similar to that of the zodiacal light. Piazzi Smyth, however, when observing on the Peak of Tenerilfe, saw this appear ance and showed by actual measurement that the glow seen before moonrise does not lie near the ecliptic, but is nearly vertical, and is due simply to refraction in the earth s atmosphere. This ex planation will hardly account for an interesting observation made by Mr L. Trouvelot, which if repeated would require to be very carefully investigated. On a night when the zodiacal light was very bright and there were magnetic disturbances followed by an auroral display, but when no aurora was actually visible, he saw a conical light rising obliquely from the top of the roof of a building behind which the moon, then about 15 or 20 above the horizon, was concealed. The axis of the light coincided nearly with the ecliptic and the light could be traced on both sides of the moon, when the moon itself was concealed. The whole of the circum stances led Mr Trouvelot to conclude that this light and the zodiacal light were phenomena of the same order, while this and other observations, he considered, rendered it probable that there was some connexion between the zodiacal light and auroras. Physical Constitution. As has already been pointed out, it is impossible to see the zodiacal light through a telescope, and this, taken along with the extreme faintness of the light, renders it exceedingly difficult to examine it satisfactorily with either the polariscope or the spectroscope. Many attempts have been made to determine whether or not the light was to any extent polarized, but with questionable results until Prof. A. W. Wright attacked the problem, using a polariscope specially designed for studying very faint lights. With this he was enabled to determine with certainty that the light was partially polarized in a plane passing through the sun, and that the amount of polar ization was most probably as much as 15 per cent., but less than 20 per cent. Many attempts have been made to observe the spectrum. In 1867 Angstrom, observing at Upsala in March, obtained the bright aurora line (W. L. 5567), and concluded that in the zodiacal light there was the same material as is found in the aurora and in the solar corona, and probably through all space. Upsala, however, is a place where the aurora spectrum can often be observed in the sky even when no aurora is visible, and it has generally been believed that what Angstrom really saw was an auroral and not a zodiacal light spectrum. Seiior A. T. Acrimiz, observing at Cadiz, obtained a con tinuous but faint spectrum with two bright lines a yellowish line, probably an aurora line, and a line in the blue, more refrangible than F, which he could not identify, but which in all probability was another auroral line. The fact that he saw this spectrum with a five-prism spectroscope attached to an equatorial seems conclusive evidence that it was not the zodiacal light spectrum that he was observing. The most satisfactory observations hitherto published seem to be those of Prof: Piazzi Smyth and Prof. A. W. Wright. Both used spectroscopes specially designed for the examination of faint lights, and their results agree completely with each other. Prof. Smyth made his observations at Palermo and found a faint con tinuous spectrum extending from about W. L. 5550 to W. L. 5000 (British inches scale), with a maximum brightness at about W. L. 5350. In fact, the light was almost exactly similar to that of faint diffused sunlight, such as is got in
the last traces of twilight. Prof. Wright s conclusion was