ster was led by theoretical considerations to the conviction that a neutral point would be found between the sun and the horizon. Observations made with the polariscope at places around its probable locality appeared to confirm this belief; but it was not till the 28th February 1842, when the sun was in the meridian with an altitude of 22°, that he had the high gratification of actually observing it. The spot beneath the sun was fortunately visible from the end of a long dark passage running north and south; and having concealed the sun himself, and every part of the sky around him except the probable position of the spot, he obtained a distinct view of the new neutral point, which was about 15° or 16° below the sun. Babinet, after several unsuccessful efforts, obtained a distinct view of this point on the 23d July 1846. This spot has been called Brewster's neutral point, and it has hitherto been seen only by these two observers.

638. With regard to the causes which disturb the polarisation of the atmosphere, Sir David Brewster and Dr Rubenson have found that clouds, fogs, and smoke were the causes of the greatest perturbations, and that the intensity of the polarisation was reduced by those crystals of ice floating in the atmosphere which form the halo of 23°. Dr Rubenson has been led by his observations to the conclusion that the polarisation of the atmosphere is subject to a diminution during the morning, and to an increase during the evening, the precise hour of least polarisation being uncertain. These changes are frequently influenced by brief perturbations which occur at all hours of the day; they arise from smoke, clouds, and probably often from cirrus too faint to be seen. As regards the daily period of atmospheric polarisation, it will be observed that the time of diminution occurs in the earlier part of the day, when vapour is carried by the ascending currents into the higher regions of the atmosphere by which the cumulus cloud is formed, and, no doubt, higher up invisible crystals are deposited similar to those of the cirrus; and that the period of increase occurs towards evening, when the temperature of the lower stratum of the air is

falling, and the ascending currents having ceased, the cumulus clouds are dissolving, and the quantity of invisible crystals likewise diminishing. Thus the maximum polarisation occurs at that time of the day when the air is least encumbered with the invisible crystals of ice, and the minimum when these crystals are present in greatest abundance.

639. M. Liais, in 1859, made observations on the polarisation of the atmosphere, for the purpose of determining its height, during a voyage from France to Brazil, and in the Bay of Rio Janeiro. From the observations, he obtained 212 miles as the height of the atmosphere. From the above results-especially those showing the influence of the different states of the aqueous vapour of the air on its polarisation-we see the importance of this property of the atmosphere as a branch of meteorology. When the subject has been more widely observed, and the law of its modification by the presence of aqueous vapour is more accurately known, it is probable that the knowledge may be turned to excellent practical account in giving the earliest indication of the commencement of the saturation of the air in the upper regions; and, inferentially, of the approach of storms and of the rainy season within the tropics.

640. Theories of Atmospheric Polarisation.—In addition to the vertical polarisation (that is, polarisation in the vertical plane passing through the sun and the observer) produced by the illumination of the particles of the air, there must be an opposite polarisation by which the neutral points are produced. Arago, Babinet, and other writers, have supposed that the opposite polarisation results from a secondary illumination which the aerial particles receive from the rest of the atmosphere which sends light to them polarised horizontally, or oppositely to the light polarised vertically. Sir David Brewster did not consider this explanation to be satisfactory, on the ground that it omitted all consideration of the light polarised by refraction, and that there is no evidence of such a secondary reflection even in a perfectly cloudless sky; and still less evidence that, if it did exist, it could neutralise the

light polarised by reflection at considerable distances from the antisolar point. The theory of atmospheric polarisation proposed by this distinguished philosopher is, that the three neutral points in the atmosphere, and the partial polarisation of the light which it reflects, are produced by the opposite action of lights polarised by reflection and refraction which have nearly the same relative intensity.*

METEORS.

641. Meteors and Shooting Stars can scarcely with propriety be included among meteorological phenomena, since the whole of the facts known regarding them lead to their being classed with astronomical bodies. Of late years they have been carefully observed and their movements made the subject of scientific investigation; and attention has been forcibly drawn to their consideration by the magnificent star-shower of the night of the 13th and 14th November 1866, which had been predicted beforehand. During a shower of meteors, if the lines of the tracks of the different meteors be projected on the sky, it is found that they meet nearly all in one point. This is the vanishing point in the perspective on which their tracks are projected, and points to that region in space over which the meteors are distributed. A number of these vanishing points have been determined. Olmstead first drew attention to this inquiry regarding the distribution in space of the meteoric matter, and showed that the great star-shower of November 1833 had its vanishing point in the constellation Leo. Professor Newton, America, has collected the records of the great November showers from the tenth century, from which there is evidence of eleven instances of their occurrence; and these records prove that they recur at regu

* For an exposition of this branch of meteorology, see Sir David Brewster's papers in Dr Keith Johnston's 'Physical Atlas,' in the 'Transac tions of the Royal Society of Edinburgh,' and in the Philosophical Magazine,' from which the above notice of the subject has been prepared.

lar intervals of thirty-three years and part of a day, the date of their occurrence having been from the 9th to the 14th November. In addition to the annual period in November, meteors also are frequently observed about the 11th of August; and, though not so general, at other times of the year. These showers indicate the regions of space through which immense numbers of meteors are moving with planetary velocity; and when the earth in its motion round the sun travels through these spaces, it attracts to it many of these cosmical bodies, which, as they enter the atmosphere with a speed of about thirty miles a second, are set on fire by the rapidity of their flight, like the arrow shot by Acestes

"Chafed by the speed, it fired; and as it flew,

A trail of following flames ascending drew."

Meteors move in orbits variously inclined to that of the earth. Alexander Herschel has particularly studied this class of bodies, and, from twenty well-observed cases, has determined the weight of the meteors as varying from 30 grains to upwards of 7 lb.

642. Since they become luminous on entering the atmosphere, it is evident that careful observations made at different places would give an approximate solution to the problem of the height of the atmosphere. From such observations it is concluded that the earth's atmosphere is at least one hundred miles high, and that probably, though in an extremely attenuated form, it reaches to about two hundred miles. The extraordinary quantity of latent heat in air so attenuated, becomes sensible as the air is compressed before the meteor, and thus we have a satisfactory explanation of the speedy ignition of all meteors as they traverse the atmosphere, and the rapid conversion of the smaller ones into thin mist, which in some cases remains floating in the sky for half an hour. Meteors enter the atmosphere bringing with them the temperature of the stellar spaces, which Sir John Herschel supposes to be about -239°; and since their surfaces quickly become highly heated, an explosion soon takes place from

the difference of the temperatures. Such explosions, however, do not always take place; thus the meteoric mass which fell at Dhurmsala, in India, on the 14th of July 1861, was first very hot, but as the great cold of the interior of the meteor quickly counteracted the heat produced on its surface by its rapid flight through the air, it soon became so intensely cold that it could not be touched.

643. Speculation has gravely attempted to bring this interesting class of bodies within the legitimate domain of meteorology, by supposing that a shadow from the annular ring of meteors rotating round the sun falls on the earth at certain seasons-viz., in February, May, August, and November-in consequence of which part of the sun's heat is cut off, and the temperature of the earth therefore falls. It is conclusive proof against this theory, that when these interruptions of temperature are particularly investigated, it is found that the date of their occurrence varies backward and forward to some extent from year to year; that the duration of their occurrence varies from three to six days, or sometimes longer; that they are accompanied with a prevalence of the polar current, as indicated by northerly winds and increasing pressure in advancing northward; and that instead of appearing simultaneously at different places, they are propa-. gated from place to place like other changes of temperature, thus pointing to a terrestrial instead of an astronomical origin.