The Aurora Borealis or Polar Light.

Harper's New Monthly Magazine 229, JUN 1869

By Elias Loomis, Professor in Yale College.

The polar light is a light which is frequently seen near the horizon, bearing some resemblance to the morning whence it has received the name of aurora. In the northern hemisphere it is usually termed "aurora borealis," because it is chiefly seen in the north. A similar phenomenon is also seen in the southern hemisphere, where it is called "Aurora Australis." Each of them may, with greater propriety, be called "Aurora Polaris," or Polar Light.

Auroras exhibit an endless variety of appearances; but they may generally he referred to one of the following classes:

First. — A light near the horizon, resembling the morning aurora or break of day. The polar light may generally he distinguished from the true dawn by its position in the heavens, especially in the United States, where it always appears in the northern quarter. This is the most common form of aurora, but it is not an essentially distinct variety, being due to a blending of the other varieties in the distance. The upper boundary of this light is an arc of a small circle, which shades off gradually into the darkness of the nocturnal sky, but is much better defined than the twilight.

Second — An arch of light somewhat in the form of a rainbow. This arch frequently extends entirely across the heavens, from east to west, and cuts the magnetic meridian nearly at right angles. This arch does not remain fur a long time stationary, but changes its elevation above the horizon; and when the aurora exhibits great splendor, several parallel arches are often seen at the same time, appearing as broad belts of light stretching from the eastern to the western horizon. In high northern latitudes five or six such arches have frequently been ,eon at once; and on two occasions have been seen nine parallel arches, separated by distinct intervals. Figure 2 represents auroral arches seen a few years since in Canada.

Third — Slender, luminous beams or columns, well defined, and frequently very brilliant. These beams rise to various heights in the heavens, frequently 20 or 30 degrees, and sometimes ascending as high as the zenith. Their breadth varies from a quarter of a degree up to two or three degrees. Frequently they last but a few minutes; sometimes they continue a quarter of an hour, a half hour, or even a whole hour. Sometimes they remain for several minntes at rest, and sometimes they have a quick lateral motion. Their light is commonly of a pale yellow, sometimes it is reddish, while occasionally it is crimson or even of blood-color. Sometimes the luminous beams are interspersed with dark rays resembling dense smoke. Sometimes the tops of the beams are pointed, and having a waving motion they resemble the lambent flames of half extinguished alcohol burning upon a broad, flat surface. [See Figures 3, 4, and 5.] Faint stars are visible through the substance of the auroral beams.

Fourth. — The corona or crown. Luminous beams sometimes shoot up simultaneously from nearly every pan of the horizon, and converge to a point a little south of the zenith, forming a quivering canopy of flame. This is called a corona or crown. The sky now resembles a fiery dome. and the crown appears to rest upon variegated fiery pillars, which are frequently traversed by waves or flashes of light. This may he called a complete aurora, and comprehends most of the peculiarities of the other varieties. [See Figure 13.] The corona seldom remains complete longer than one hour. The streamers then become fewer and less intensely colored, the luminous arches break up, while a dark segment is still visible near the northern horizon, and at last nothing remains but masses of delicate clouds. During the exhibition of brilliant auroras, delicate fibrous clouds are commonly seen floating in the upper regions of the atmosphere; and on the morning after a great nocturnal display we sometimes recognize streaks of cloud similar to those which had been luminous during the preceding night. Sometimes when the sun is above the horizon these clouds arrange themselves in forms similar to the beams of the aurora, constituting what has been called a "day aurora."

Fifth — Waves or flashes of light. The luminous beams sometimes appear to shake with at tremulous motion; while flashes like waves of light roll up toward the zenith, and sometimes travel along the line of an auroral arch. Sometimes the beams have a slow, lateral motion from east to west, and sometimes from west to east. These sudden flashes of auroral light are known by the name of "Merry Dancers," and form an important feature of nearly every splendid aurora.

The duration of auroras is very variable. Some last only an hour or two; others last all night; and occasionally they appear on two successive nights under circumstances which lead us to believe that, were it not for the light of the sun, an aurora might be seen uninterruptedly for thirty-six or forty-eight hours. For more than a week, commencing August 28, 1859, in the northern part of the United States, the aurora was seen almost uninterruptedly every clear night. In the neighborhood of Hudson Bay the aurora is seen for months almost without intermission.

Auroras are characterized by recurring tits of brilliancy. After a brilliant aurora has faded away and almost entirely disappeared, it is common for it to revive, so as to rival and often to surpass its first magnificence. Two such alternations are common features of brilliant auroras, and sometimes three or four occur on the same night.

The color of the aurora is very variable. When the aurora is faint its light is usually white or a pale yellow. When the aurora is brilliant, the sky exhibits at the same time to great variety of tints; some portions of the sky are nearly white, but with a tinge of emeraldgreen; other portions arc of a pale yellow or strawcolor; others are tinged with a rosy hue; while others may have a crimson hue, which sometimes deepens to a blood-red. These colors arc ever varying in their position and in the intensity of their light. Auroras are sometimes observed simultaneously over large portions of the globe. The aurora of August 28, 1859, was seen throughout more than 140 degrees of longitude, from Eastern Europe to California; and from Jamaica on the south to an unknown distance in British America on the north. The aurora of September 2, 1859, was seen at the Sandwich Islands; it was seen throughout the whole of North America and Europe; and the disturbance of the magnetic needle indicated its presence throughout all Northern Asia, although the sky was overcast, so that at many places it could not be seen. An aurora was twen at the same time in South America and New Holland. The auroras of September 25, 1841, and November 17, 1848, were almost equally extensive.

In the United States an aurora is uniformly preceded by a hazy or slaty appearance of the sky, particulady In the neighborhood of the northern horizon. When the auroral display confluences, this hazy portion of the sky assumes the form of a dark hank or segment of a circle in the north, rising ordinarily to the height of from five to ten degrees. [See Figure 6]. This dark segment is not a cloud, for the stars are seen through it as through a smoky atmosphere, with little diminution of brilliancy.

This dark bank is simply a dense haze, and it appears darker from the contrast with the luminous are which rests upon it. In high northern latitudes, when the aurora covers the entire heavens, the whole sky seems filled with a dense haze; and in still higher latitudes, where the aurora is sometimes seen in the south, this dark segment is observed resting on the southern horizon and bordered by the auroral light. This phenomenon was noticed in the United States in the aurora of August, 1859. The highest point of this dark segment generally coincides with the magnetic meridian. Exceptions to this rule do, however, frequently occur, and in some places there is a constant deviation of ten degrees or more.

The dark segment just described is bounded by a luminous arc, whose breadth varies from half a degree to one or two degrees. The lower edge of the arc is well defined; but unless the breadth be very small the upper edge is ill defined, and blends with a general brightness of the sky. If the aurora becomes brilliant, other arcs usually form at greater elevations, sometimes pausing through the zenith. The summit of these arcs is situated nearly in the tnagne:ie meridian, and the arc sometimes extends symmetrically on each side toward the horizon. Frequently, however, the summit of the arc deviates ten degrees or more from the magnetic meridian, and in some places this deviation appears to be tolerably constant.

An auroral arch is frequently incomplete, and extends only a portion of the distance front one horizon to the other. The apparent breadth of auroral arches varies with their elevation above the horizon. The result of a large number of observations gives eight degrees as the average breadth of arches seen at altitudes less than sixty degrees; while fot arches whose altitude is greater than sixty degrees the average breadth is twenty-five degrees.

When an arch appears to move across the sky from north to south or the reverse, its angular breadth exhibits corresponding changes. If the distance of an arch from the earth remained constant during its movement of translation, and the arch was of the term of a ring whose section was a circle, its breadth when in the zenith should be double what it was when its elevation was 30 degrees. But its observed breadth in the former case is three or four times as great as in the latter, showing that a section of the ring is of an oval form with its greatest diameter parallel to the earth's surface.

Auroral arches do not meet the horizon at points distant 180 degrees from each other. Careful measurements have shown that, except near the horizon, they may be regarded us portions of small circles parallel to the earth's surface. Such a circle seen obliquely would have the appearance of an ellipse. Near the horizon the elliptic form of the auroral arch has sometimes been quite noticeable, the extremities of the arch being bent inward as shown in Figure 7. Occasionally an ellipse has been seen almost entire, and in one instance the ellip,e has been seen complete. the axes of the ellipse being in the ratio of two to one.

Sometimes an auroral arch consists of rays arranged in irregular and sinuous bands of tanous and variable curvatures, like the undulations of a streamer or flag waving in the breeze. Sometimes the appearance is that of a brilliant curtain whose folds are agitated by the wind. [See Figures 10 and 11.] These folds sometimes become very numerous and complex, and the arch assumes the form of a long sheet of rays returning into itself, the folds enveloping each other, and presenting an immense variety of the most graceful curves.

Auroral arches generally temd to divide into short rays running in the direction of the breadth of the arch, and converging toward the magnetic zenith. They frequently appear to be formed of transverse fibres terminating in a regular curve, which forms the lower edge of the arch. Arches of a uniform nebulous apprearance are not the most frequent; striated arches are very common; and auroral arches present every intermediate variety between these two extremes. Sometimes auroral beams arrange themselves in the form of an arch. Sometimes an auroral arch is formed of short streams parallel to each other, presenting the appearance of a row of comets' tails.

An auroral arch does not long maintain a fixed position. It is frequently displaced, and is transported parallel to itself from north to south, or from south to north. Sometimes an arch which is first seen near the northern horizon gradually rises, ascends to the zenith, and descends toward the southern horizon, where it remains for a time nearly stationary, and then perhaps retraces its course. In the United States, as well as in Europe, auroral arches more frequently move from north to south than from south to north. Sometimes there is also a movement of the arch from west to east, or from east to west. The rate of motion of auroral arches very variable. If we suppose the arch to he elevated 125 miles above the earth, the observed angular motion of translation would indicate an actual velocity of from 1000 to 3000 feet per second.

The motion of auroral beams is sometimes in a lateral direction, and sometimes it is upward or downward. The downard motion is the most common, and sometimes it takes place with very great velocity, and in a large number or beams simultaneously. When an auroral beam rises and falls alternately without much change of length, it is said to dance. This is a common occurence in high northern latitudes, where it is known by the name of "Merry Dancers."

When the sky is filled with a large number of separate beams all parallel to each other, according to the rules of perspective these beams will seem to converge to one point, as shown in the Figures 15 and 16; and if the beams are parallel to the direction of the dipping-needle, they will seem to converge to the magnetic zenith. [See Figure 17.] Hence results the appearance of a corona or crown of rays whose centre is less luminous than other portions of the sky. [See Figure 13.]

Sometimes the corona is incomplete, the beams on one side being deficient. When a striated arch passes the magnetic zenith it frequently presents the appearance of an incomplete corona. If the arch advance. from north to south, before reaching the magnetic zenith it forms a half crown on the northern side; when it passes the magnetic zenith we have a corona tolerably complete; and after the arch has passed the magnetic zenith it forms a half crown on the southern side.

When an aurora becomes less active its beams become less luminouss, their edges become more diffuse, they diminish in length while they increase in breadth, and assume the appearance of luminous clouds. Sometimes they exhibit a fibrous structure, and present a strong resemblance to those delicate clouds which are often seen in pleasant weather, and are designated by the term "cirrus."

During the exhibition of a brilliant aurora we frequently notice an appearance of general nebulosity or luminous vapor covering large portions of the heavens, and sometimes altnost the entire celestial vault. In the upper part of the sky the light is generally faint, sometimes not exceeding that of the milkyway; but near the horizon the light is sometimes so intense as to resemble a vast conflagration. The great disparity between the light of auroral vapor when near the zenith and near the horizon indicates that the vertical thickness of the auroral vapor is small in comparison with its horizontal dimensions.

The great auroral exhibition of August and September, 1859, was very carefully observed at a large number of stations, and these observations have enabled us to determine the height of the aurora above the earth's surface. At the most southern stations where these auroras were observed, the light rose only a few degrees above the northern horizon; at more northern stations the aurora appeared at a greater elevation; at certain stations it just attained the zenith; at stations further north the aurora covered the entire northern heavens as well as a portion of the southern; and at places still further north nearly the entire visible heavens front the northern to the southern horizon were overspread with the auroral light.

In Figure 14, An represents a portion of the earth's surface, and beneath are given the names of some of the places where observations were made upon the aurora of August 28, 1859, all at the same hour of the evening. The dotted lines drawn from the five most southern stations (Jamaica to Savannah) represent the elevations of the upper boundary of the auroral light above the northern horizon. The point 1) thus determined is then the upper edge of the auroral light near its southern margin, and this point is found to be 534 miles above the earth's surface.

The dotted lines from the five most northern stations (Sandy Spring, Maryland, to Lewiston, Maine) show the elevation of the lower limit of the auroral light above the south horizon. The point C thus determined is the lower edge of the auroral light near its southern margin, and this point is found to be 46 miles above the earth's surface. The line CD represents, therefore, the southern boundary of the auroral illumination.

These results, combined with a vast number of other observations, show that the aurora of August 28, 1859, formed a stratum of light encircling the northern hemisphere, extending southward to latitude 38 degrees in North America, and reaching to an unknown distance on the north; and it pervaded more or less the entire interval between the elevations of 46 miles and 500 miles above the earth's surface. This illumination consisted chiefly of luminous beams or columns every where nearly parallel to the direction of a magnetic needle when freely suspended.

At New York a magnetic needle, freely suspended, points about seven degrees westward of the true north; and if the needle be supported by its centre of gravity, so as to be free to move in a vertical plane, the north pole will incline downward, making an angle of about 17 degrees with a vertical line. Such a needle we call a dipping-needle, and the point nearly overhead toward which one pole of the dipping-needle is directed is called the magnetic zenith. In Southern Florida the dip of the magnetic needle is 55 degrees, and it increases as we proceed northward, being about 73 degrees at New York, and 78 degrees at Quebec.

The luminous beams in the aurora of August, 1859, were sensibly parallel to the direction of the dipping-needle; they were about 500 miles in length, while their diameters varied from 5 to 50 miles, and perhaps sometimes they were still greater.

The height of a large number of auroras has been computed by similar methods, and the average result for the upper limit of the streamera is 450 miles. From a multitude of observations it is concluded that the aurora seldom appears at an elevation less than about 45 miles above the earth's surface, and that it frequently extends upward to an elevation of 500 miles. Auroral arches having a well-defined border are generally less than 100 miles in height.

Some persons contend that the aurora is occasionally seen nt elevations of less than one mile above the earth's surface. It has been claimed that the aurora is sometimes seen between the obserYer and a cloud; but this appearance is believed to result from a cloud of very small density, thoroughly illumined by auroral light which shines through the cloud, so as to produce the same appearance as if the aurora prevailed On the under side of the cloud.

Sometimes the lower extremity of an auroral streamer appears to be prolonged below the summit of a neighboring mountain or hill. This appearance is probably an illusion. The same phenomenon has been noticed by more cautious observers, who traced the result to the reflection of the auroral light from the snow which covered the mountain. Although it is possible that the aurora may sometimes descend nearly to the earth's surface, there is no sufficient evidence to prove that the true polar light has ever descended so low as the region of ordinary clouds.

There is no satisfactory evidence that the aurora ever emits any audible sound. It is nevertheless a common impression, at least in high latitudes, that the aurora sometimes emits sound. This sound has been described as a rustling, hissing, crackling noise. But the most competent observers, who have spent several minters in the Arctic regions, where auroras are seen in their greatest brilliancy, have been convinced that this supposed rustling is a mere illusion. It is therefore inferred that the sounds which have been ascribed to the aurora must have been due to other causes, such as the motion of the wind, or the cracking of the snow and ice in consequence of their low temperature.

If the aurora emitted any audible sound this sound ought to follow the auroral movement after a considerable interval. Sound requires four minutes to travel a distance of 50 miles. But the observers who report noises succeeding auroral movementa make no mention of any interval. It is therefore inferred that the sounds which have been heard during auroral exhibitions are to be ascribed to other causes than the aurora.

Auroras are very unequally distributed over the earth's surface. They occur most frequently in the higher, latitudes, and arc almost unknown within the tropics. At Havana, in latitude 23 degrees, but six auroras have been recorded within a hundred years, and south of Havana auroras are still more unfrequent. As we travel northward from Cuba, auroras increase in frequency and brilliancy; they rise higher in the heavens, and oftener ascend to the zenith. Near the parallel of 40 degrees we find on an average only ten auroras annually. Near the parallel of 42 degrees the average number is twenty annually; near 45 degrees the number is forty; and near the parallel of 50 degrees it amounts to eighty annually. Between this point and the parallel of 62 degrees auroras, daring the winter, are seen almost every night. They appear high in the heavens, and as often to the south as the north. In regions further north they are seldom seen except in the south, and from this point they diminish in frequency and brilliancy as we advance toward the pole. Beyond latitude 62 degrees the average nullifier of auroras is reduced to forty annually. Beyond latitude 67 degrees it is reduced to twenty; and near latitude 78 degrees it is reduced to ten annually. If we make a like comparison for any European meridian we shall find a similar result, except that the auroral region is situated further northward than it is in America. Upon Figure 18 the dark shade indicates the region where the average number of auroras annually amounts to at least eighty; and the lighter shade indicates the region where the average number of auroras annually amounts to at least forty.

We thus see that the region of greatest auroral abundance is a zone of an oval form surrounding the north pole, and whose central line crosses the meridian of Washington in latitude 56 degrees, and the meridian of St. Petersburg in latitude 71 degrees. Accordingly, auroras are much more frequent in the United States than they are in the same latitudes of Europe. Within this auroral zone is a region 2000 miles in diameter, throughout which it is presumed that auroras are not more common than they are in New England.

Auroras in the southern hemisphere are nearly, if not quite, as frequent as they are in the corresponding magnetic latitudes of the northern hemisphere, and it is probable that the geographical distribution of auroras in the two hemispheres is somewhat similar.

By comparing the records of auroras in the two hemispheres we find a remarkable coincidence of dates, which seems to justify the conclusion that an unusual auroral display in the southern hemisphere is always accompanied by an unusual display in the northern hemisphere; that is, a great exhibition of auroral light about one magnetic pole of the earth is uniformly attended by a great exhibition of auroral light about the opposite magnetic pole.

The aurora is ordinarily accompanied by a considerable disturbance of the magnetic needle, and the effect increases with the extent and brilliancy of the aurora. Auroral beams cause a disturbance of the needle, particularly when the beams themselves are in active motion. Auroral waves or flashes, especially if they extend us high as the zenith, cause a violent agitation of the needle, consisting of an irregular oscillation on each ride of its mean position.

These extraordinary deflections of the needle prevail almost simultaneously over large portions of the globe, even where the aurora itself is not visible. During the great auroral display of September 2, 1859, the disturbances of the magnetic needle were very remarkable throughout North America, Europe, and Noethern Asia, as well as in New Holland. At Toronto, in Canada, the declination of the needle changed nearly four degrees in half an hour. The inclination was observed to change nearly three degrees when the needle passed beyond the limits of the graduated scale, so that the entire range of the needle could not be determined. At several observatories in Europe still more remarkatle disturbances were recorded. These irregular disturbances of the magnetic needle are not quite sitnultaneous at distant stations. Over the surface of Europe and also of North America they appear to be propagated from northeaat to southwest at the rate of about 100 miles per minute.

Auroras exert a remarkable influence upon the wires of the electric telegraph. During the prevalence of brilliant auroras the telegraph lines generally become unmanageable. The aurora develops electric currents upon the wires, and hence results a motion of the telegraph instruments similar to that which is employed in telegraphing; and since this movement is frequent and irregular, it ordinarily becomes impossible to transmit intelligible signals. During several remarkable auroras, however, the currents of electricity on the telegraph wires have been so steady and powerful that they have been used for telegraph purposes as a substitute for a voltaic battery; that is, messages have been transmitted by telegraph from the auroral influence alone. This result proves that the aurora develops on the telegraph wires an electric current similar to that of a voltaic battery, and differing only in its variable intensity.

Auroras appear at all hours of the night, but not with equal frequency. The aserugs number increases uninterruptedly from sunset till about midnight, from which time the number diminishes uninterruptedly till morning. In Canada the maximum occurs an hour before midnight; further north, in latitude 52 degrees, the maximum occurs at midnight; and still further north, to the Arctic Ocean, the maximum occurs an hour after midnight.

Auroras occur in each month of the year, but not with eqnal frequency. In New England and New York the least number of auroras is recorded in winter, and the greatest number in the autumn. It is difficult to make an entirely satisfactory comparison on account of the unequal length of the days in the different seasons of the year; but apparently the maximum occurs in September, and the minimum in December or January. The number of auroras seen in different years is extremely variable. Sometimes, for several years, auroras arc remarkable for their number and magnificence, and then there succeeds a barren interval during which auroras are almost entirely forgotten.

If we compare the observations made at any one station for a long period of years, we shall diseover that the inequality in the number of auroras upon successive years recurs periodically. In order to discover the law which governs auroral displays, it important to have observations made at the same station upon a uniform plan continued for a long period of time. A tolerably complete auroral record has been kept at New Haven for nearly one hundred years, and a similar record has been kept in the neighborhood of Boston since 1712. Similar records have been preserved at many places in Europe, extending back for a period of two centuries. In order to neutralize as far as possible the imperfections of any single record, I have taken the average of three different records, viz.: those at New Haven and Boston, representing New England, and that at St. Petersburg, representing the north of Europe. Instead of exhibiting these results in a tabular form I have represented them by a curve line in Figure 19. The years are indicated both at the top and bottom of the figure, and from the base line AB for each year a perpendicular is drawn whose length is proportional to the number of auroras recorded for that year, the number of auroras being indicated on the left of the figure. Through the points thus determined a curve line is drawn, and this curve represents the relative number of auroras in New England and Northern Europe for a period of 130 years. This curve dearly indicates a period of unusual auroral abundance from 1770 to 1790. Then followed a period of great barrenness from 1792 to 1826; and then succeeded another period of unusual abundance from 1836 to 1864. We also notice subordinate fluctuations which generally succeed each other at intervals of about eleven years. Thus auroras were uncommonly abundant in the years 1773, 1781, 1787, 1840, 1848, and 180, while during the barren period from 1790 to 1826 is slight increase will be remarked in the years 1804 and 1819, with a more decided increase in 1830. Most of these peculiarities are noticeable in the observations at each of the stations employed in this comparison, and the principal of them are clearly marked in the observations at every station where an auroral record has been long continued. These inequalities in the observed frequency of auroras are not accidental, nor are they local peculiarities, but they are characteristic of the northern hemisphere of our globe. It is then considered as established that periods of unusual auroral abundance succeed each other at intervals of from eight to sixteen years, the average interval being somewhat over eleven years. Moreover, these successive maxima are very unequal in intensity, showing generally a grand maximum at the end of five of the shorter periods; that is, at intervals of 55 or 56 years. These conclusions are confirmed by the records of auroral displays extending back a century earlier than is shown in Figure 19. Tte grand result, then, which we have deduced from the observations is, that auroras recur in unusual numbers every eleven years, and there is a maximum of unusual splendor every fifty-five years. During the last few years auroras have been less numerous than usual, but a considerable increase of splendor may be anticipated about 1870.

Theory of the Polar Light.

Some have ascribed the polar light to a rare nebulous matter occupying the interplanetary spaces, and revolving round the sun at such a distance that a portion of this matter occasionally falls into the upper regions of the atmosphere with a velocity sufficient to render it luminous from the condensation of the air before it. Upon this hypothesis the aurora would not differ essentially from a grand exhibition of shooting-starks unless, perhaps, in the density of the substance which occasions the phenomenon. But this hypothesis will not explain why auroras are always confined to certain districts of the earth, and are wholly unknown in other portions. We reject this hypothesis, therefore, as irreconcilable with the known geographical distribution of auroras.

Auroral exhibitions take place in the upper regions of the atmosphere, since they partake of the earth's rotation. All the celestial bodies have an apparent motion from east to west, arising from the rotation of the earth; but bodies belonging to the earth, including the atmosphere and the clouds which float in it, partake of the earth's rotation, so that their relative position is not affected by it. The same is true of auroral exhibitions. Whenever an auroral corona is formed, it maintains sensibly the same position in the heavens during the whole period of its continuance, although the stars meanwhile revolve at the rate of 15 degrees per hour.

The grosser part of the earth's atmosphere is limited to a moderate distance from the earth. At the height of a little over four miles, the density of the air is only onehalf what it is at the earth's surface. At the height of 50 miles the atmosphere is wellnigh inappreciable in its effect upon twilight. The phenomena of lunar eclipses indicate an appreciable atmosphere at the height of 66 miles. rime phenomena of shoutingstars indicate an atmosphere at the height of 200 or 300 miles, while the aurora indicates that the atmosphere does not entirely cease at the height of 500 miles. Auroral exhibitions take place, therefore, in an atmosphere of extreme rarity; so rare indeed that if, in experiments with an airpump, we could exhaust the air as completely, we should say that we had obtained a perfect vacuum.

The auroral light is electric light. Onr first reason for believing in this identity is derived front the appearance of the auroral light. The colors of the aurora are the same as those of ordinary electricity passed through rarefied air. When a spark is drawn from an ordinary electrical machine in air of the usual density, the light is intense and nearly white. If the electricity be passed through a Ow vessel in which the air hat been partially rarefied, the light is more diffuse, and inclines to a delicate rosy hue. If the air be still further rarefied, the light becomes very diffuse, and its color becomes a deep rose or purple. The same variety of colors is observed during auroral exhibitions. The transition from a white or pale straw-color to a rosy hue, and finally to a deep red, probably depends upon the height above the earth, and upon the amount of condensed vapor present in the air.

The emerald-green light which is seen in some auroras is ascribed to the projection of the yellow light of the aurora upon the blue sky, since green may be formed by a combination of yellow and blue light. A similar effect is often produced in the evening twilight by a combination of the yellow light of the sun with the blue of the celestial vault.

The light of electricity possesses certain properties which distinguish it front solar light. There are certain substances which, in ordinary solar light, appear almost entirely transparent, like pure water, but which, when illumined by an electric spark in a dark room, present a very peculiar appearance, as if they were self-luminous. This appearance is termed fluorescence. When such substances are illumined by auroral light, they exhibit the same peculiarity as when illumined by the spark of an ordinary electrical machine.

These considerations must he admitted to create a strong probability that auroral light is identical with electric light. This probability becomes a certainty when we study the effect of an aurora upon the telegraph wires. The electric telegraph is worked by a current of electricity generated by a voltaic battery, and flowing along the conducting wire which unites the distant. stations. This current, flowing round an electromagnet, renders it temporarily magnetic, so that its armature is attracted, and a mark is made upon a roll of paper. During a thunderstorm the electricity of the atmosphere affects the conducting wire in a similar manner, and a great auroral display produces a like effect. During the auroras of August and September, M59, there were remarked all those classes of effects which are considered as characteristic of electricity. We will enumerate the most remarkable of these effects:

(1.) In passing front one conductor to another, electricity exhibits a spark of light. This light is not like that of a burning coal or a heated iron, but a bright spark, without appreciable duration, which is renewed whenever the electricity passes. During the auroras of 1S59, at numerous stations both in America and Europe, similar sparks were drawn from 'the telegraph wires when no battery was attached.

(2.) In pawing through pour conductors electricity develops heat. In like manner, during the auroras of 1859, both in America and Europe, paper, and even wood, were set on fire by the aurorul influence alone.

(3.) When passed through the animal system, electricity communicates a well-known characteristic shock. This electric shock is unlike any effect which can be produced upon the nervous system by any other known method. During the auroras of 1859 several telegraph operators received similar shocks when they touched the telegraph wires.

(4.) A current of electricity decomposes compound substances, resolving them into their elements. Most of the objects with which we arc familiar in daily life are compound; that is, are formed by the union of two or more elementary substances. The current of an ordinary voltaic battery affords one of the most efficient means of resolving compound bodies into their elements. The aurora of 1859 was found to produce similar decompositions. One method of transmitting telegraph signals, which 1144 been successfully practiced, is known by the name of the electrochemical, in which a mark is made upon diemically prepared paper, this mark resulting from the decomposition of the substance with which the paper is impregnated. This substance is decomposed by the passage of an electric current, and the change of color of the paper is the visible proof of the decomposition. The aurora of 1s59 produced the same marks upon chemical paper as are produced by an ordinary voltaic battery.

(5.) A current of electricity develops magnetism in soft iron. The auroras of 1859 developed magnetism in a similar manner, and they developed it in such abundance that it was more than sufficient for the ordinary business of telegraphing.

(6.) A current of electricity deflects a magnetic needle from its ordinary position of rest. In England the usual telegraph signal is made by a magnetic needle, surrounded by a coil of copper wire, so that the needle is deflected by an electric current flowing through the wire. Similar deflections were caused by the auroras of l859, and these deflections were greater than those produced by the telegraph batteries.

These facts clearly demonstrate that the fluid developed by the aurora on telegraph wires is indeed electricity. This electricity may be supposed to be derived from the aurora, either by direct transfer from the air to the wires, or may be induced upon the wires by the action of the auroral fluid at a distance. If we adopt the former supposition, then the light is certainly electric light. If we adopt the latter supposition, then, since we know of but two agents, magnetism and electricity, capable of inducing electricity in a distant conductor, and since magnetism is not luminous, we seem compelled to admit that the auroral light is electric light.

The formation of an auroral corona near the magnetic zenith is the effect of perspective resulting from A great number of luminous beams nil parallel to each other. A large collection of vertical beams, as shown in Figure 15, would exhibit the appearance of a great number of beams diverging from a point directly overhead, as slim It in Figure 16; and a large collection of inclined beams, all parallel to each other, would produce a similar appearance, except that the point of divergence would not be in the zenith, but in that part of the sky toward which the beams were directed. Now the auroral beams arc all parallel to the direction of a magnetic needle freely suspended by its centre of gravity; and they all appear to diverge from that point of the sky toward which the pole of the dipping-needle is directed. The auroral corona or crown appears, therefore, always in the magnetic zenith; and it is not the same crown which is seen at different places any more than it is the same rainbow which is seen by different observers.

The auroral beams are simply spaces which are illumined by the flow of electricity through the upper regions of the atmosphere. During the auroras of 1859 these beams were nearly 500 miles in length, and their lower extremities were elevated about 45 miles above the earth's surface. Their tops inclined toward the south, about 17 degrees in the neighborhood of New York, this being the position which the dipping-needle there assumes.

It was formerly supposed that the electric current necessarily moved in the direction of the axis of the auroral beams; that is, that the electric discharge was from the upper regions of the atmosphere to the earth, or the reverse. Recent discoveries have suggested the possibility of a different explanation. When a stream of electricity flows through a vessel from which the air is almost wholly exhausted, under certain circumstances the light becomes stratified, exhibiting alternately bright and dark hands crossing the electric current at right angles, front which it might be inferred that electricity flowing horizontally through the upper regions of the atmosphere might exhibit alternately bright and dark bands like the auroral beams. Bui this stratification of the electric light is generally ascribed to rapid intermittences in the intensity of the electric discharge, and it is not probable that such intermittcnces can take place in nature with sufficient rapidity to produce it similar effect. It seems, therefore, more probable that auroral beams are the result of a current of electricity traveling in the direction of the axis of the beams.

The slaty appearance of the sky, which is a common feature of great auroral exhibitions, arises from the condensation of the vapor of the air, and this condensed vapor probably exists in the form of minute spicithe of ice or Hakes of snow. Fine flakes of snow have been repeatedly observed to fall during the exhibition of auroras, and this snow only slightly impairs the transparency of the atmosphere without presenting the appearance of clouds. It produces a turbid appearance of the atmosphere, and causes that dark bank which in the United States rests on the northern horizon. This turbidness is more noticeable near the horizon than it is at great elevations, because near the horizon the line of vision traverses a greater extent of this hazy atmosphere. When the aurora covers the whole heavens the entire atmosphere is filled with this haze, and it dark segment may be observed resting on the southern horizon.

Philosophers are by no means agreed as to he origin of atmospheric electricity. It has teen ascribed successively to friction, combusion, and vegetation, but these causes seem enirely inadequate to account for the enormous quantities of electricity sometimes present in he atmosphere.

Evaporation is probably the principal source of atmospheric electricity. The vapor which rises from the ocean in all latitudes, but most abundantly in the equatorial regions of theearth, carries into the upper regions of the atmosphere a considerable quantity of positive electricity, while the negative electricity remains in the earth. This positive electricity, after rising nearly vertically with the ascending currents of the atmosphere, would be conveyed toward either pole by the upper currents of the atmosphere.

The earth and the rarefied air of the upper atmosphere may be regarded as forming the two conducting plates of a condenser, which are separated by an insulating stratum, viz., the lower portion of the atmosphere. The two opposite electricities must then be condensed by their mutual influence, especially in the polar regions, where they approach nearest together; and whenever their tension reaches a certain limit there will be discharges from one conductor to the other. When the air is humid it becomes a partial conductor, and conveys a portion of the electricity of the atmosphere to the earth. On account of the low conducting power of the air, the neutralization of the opposite electricities would not be effected instantaneously, but by successive discharges more or less continuous and variable in intensity. These discharges should frequently occur simultaneously at the two poles, since the electric tension of the earth should he nearly the same at each pole.

Figure 20 represents the system of circulation here supposed; the north and south poles of the earth being denoted by the letters N and S, the direction of the currents being indicated by the direction of the arrows.

When electricity from the upper regions of the atmosphere discharges itself to the earth through an imperfectly conducting medium, the flow can not be every where uniform, but must take place chiefly along certain lines where the resistance is least; and this current mast develop light, forming thus an auroral beam. It might be expected that these beams would have a vertical position, but their position is controlled by the earth's magnetism. The earth is a magnet of vast dimensions, but feeble intensity. It is found that when magnetic forces act upon a perfectly flexible conductor, through which an electric current is passing, the conductor must assume the form of a magnetic curve. Now at each point of the earth's surface the dipping-needle shows the direction of the magnetic curve passing through that point. Hence the axis of an auroral streamer must lie in the magnetic curve which pasaes through its base; and since adjacent streamers are sensibly parallel the beams appear to converge toward the magnetic zenith.

Auroral arches assume a position at right angles to the magnetic meridian in consequence of the influence of the earth's magnetism. Auroral arches generally consist of a collection of short auroral beams all nearly parallel to each other. These beams tend to arrange themselves upon a curve which is perpendicular to the magnetic meridian, forming thus it ring about the magnetic pole. The same law has been discovered to hold true fur a stream of electricity under the influence of an artificial magnet. When electricity escapes front a metallic conductor under a receiver from which the air has been exhausted, and this conductor is the pole of a powerful magnet, the electric light forms a complete luminous ring around this conductor.

In like manner the auroral arch is a part of a luminous ring, nearly parallel to the earth's surface, having the magnetic pole for its centre, and cutting all the magnetic meridians at right angles; and this position results from the influence of the earth's magnetism.

The flashes of light observed in great almond displays are due to inequalities in the motion of the electric currents. On account of the imperfect conducting power of the air, the flow of electricity is not perfectly uniform, but escapes by paroxysms. The flashes of the aurora are therefore feeble flashes of lightning.

The disturbance of the magnetic needle during auroras is due to currents of electricity flowing through the atmosphere or through the earth. A magnetic needle is deflected from its mean position by an electric current flowing near it through a good conductor like a copper wire. A stream of electricity flowing through the earth or the atmosphere must produce a similar effect.

It is probable that the directive power of the magnetic needle is due to electric currents circulating around the globe from east to west. Such currents would cause the magnetic needle every where to assume a position corresponding with what is actually observed; and the existence of such currents has been proved by direct observation.

According to the theory already explained, positive electricity circulates from the equator toward either pole through the upper regions of the atmosphere, and thence through the earth toward the equator, to restore the equilibrium which is continually disturbed by evaporation from the waters of the equatorial seas. This current from the polar regions must modify the regular current which is supposed to be constantly circulating from east to west, resulting in a current from northeast to southwest, in conformity with observations. This current does not, however, flow uninterruptedly from northeast to southwest, but alternates at short intervals with a current in the contrary direction. Such currents of electricity roust produce a continual disturbance of the magnetic needle, and they seem sufficient to account for the disturbances actually observed.

The effect of the aurora upon the telegraph wires is similar to that of electricity in thunderstorms, except in the intensity and steadiness of its action. During thunderstorms the elsetricity of the wires is discharged instantly with a flash of lightning, while during auroras there is sometimes a strong and steady flow of electricity continuing for some minutes.

The geographical distribution of auroras depends chiefly upon the relative intensity of the earth's magnetism in different latitudes. According to experiments with artificial magnets, the electric light tends to form a ring around the pole, and at some distance from it. The electric light should, therefore, be most noticeable in the neighborhood of the earth's magnetic pole, but not directly over it. Auroras are accordingly most abundant along a certain zone which follows nearly a magnetic parallel, being every where nearly at right angles to the magnetic meridian of the place.

The electricity of the lower regions of the atmosphere within the tropics has great intensity, and moves with explosive violence in thunder-showers; and these exhibitions of electricity do not appear to be controlled by the earth's magnetism. But the electricity of the upper regions of the atmosphere is mainly controlled by the magnetic forces of the earth, and hence, in conformity with what we have observed in our experiments with artificial magnets, exhibitions of auroral light are almost entirely unknown in the equatorial regions of the earth.

The diurnal inequality in the frequency of auroras is probably due to the same causes as the diurnal variation in the intensity of atmospheric electricity. The intensity of atmospheric electricity varies with the hour of the day, being least about four o'clock in the morning, and greatest about ten o'clock in the evening. This variation is to be ascribed partly to real changes in the amount of electricity present in the air, and partly to variations in the conducting power of the air. Auroral displays are most frequent about midnight, probably because, on account of the increasing moisture of the air, the electricity accumulated in the upper regions of the atmosphere is most readily transmitted to the earth; and auroral displays become less frequent in the latter part of the night, because this accumulated electricity becomes partially exhausted by the steady discharge to the earth.

Similar considerations will explain in some measure the unequal frequency of auroras in the different months of the year; but it seems pretty well established that this inequality is partly due to the influence of extraterrestrial forces, as explained in the following paragraphs:

The secular inequality in the frequency of auroras seems clearly to indicate the influence of distant celestial bodies upon the electricity of our globe. This is inferred from the fact that the periods of auroras observe laws which are similar to those of two other phenomena, viz., the mean diurnal variation of the magnetic needle, and the frequency of black spots upon the sun's surface.

The magnetic needle has a small diurnal variation, the north end moving a little to the cast in the morning, and toward the west about the middle of the day. The mean daily change of the magnetic needle not only varies with the locality, but also varies from one year to another at the same locality, and these variations exhibit decided evidence of periodicity. In order to exhibit this fact readily to the eye I have drawn upon Figure 19 a curve line which represents these variations in Central Europe during a period of nearly a century. The curve is constructed in a manner similar to that representing the frequency of auroras, and which has been already described. The years are indicated at the top and bottom of the figure, and for each year is drawn a vertical line whose length is proportioned to the mean daily change of the magnetic needle for that year. A curve line is then drawn passing through the several points thus determined. The range of the magnetic needle is indicated by the scale on the left margin of the page; and it is seen that in 1829 the mean daily change of the magnetic needle was about fourteen minutes, while in 1834 it was less than eight minutes. Again, in 1838, it attained another maximum, and in 1844 another minimum, and so on. The undulations of this curve bear a remarkable resemblanee to the curve representing the frequency of auroras. The maxima and the minima of the two phenomena generally occur on the same years, and always nearly at the same date. We can not doubt, then, that one of these phenomena is dependent upon the other, or both are dependent upon a common cause.

The frequency and the extent of black spots upon the sun's surface exhibit a similar periodicity. Some years the sun's disc is never seen entirely free from spots, while in other years, for weeks and even months together, no spots of any kind can be perceived. On Figure 19 is drawn a curve line representing the relative number of spots seen on the sun's surface in different years from 1740 to the present time. It will be perceived that the times of maximum and minimum of the solar spots correspond almost exactly with the times of maximum and minimum of the magnetic variation, and both agree in a remarkable manner with the times of maximum and minimum frequency of auroral displays. We must therefore conclude that these three phenomena — the solar spots, the mean daily range of the magnetic needle, and the frequency of auroras — are somehow dependent the one upon the other, or all are dependent upon a common cause.

The interval from one maximum of the soy lar spots to another maximutn is somewhat variable; but its average value deduced from observations of more than a century is 11 1/5 years. Now what cause can be supposed to operate upon the sun to produce a grand display of black spots every 11 1/9 years? Jupiter makes one revolution about the sun in 11 7/8 years; and there is no other known celestial body having about the same period which could be supposed to exert an influence upon this phenomenon. In what way Jupiter should be capable of disturbing the surface of the sun we do not know, but if this disturbance results from the action of any of the planets, Jupiter is the one to be first suspected on account of his enormous mass. It must, however, bo admitted that time period of Jupiter is a little longer than the average period of the solar spots; whereas, if Jupiter is the cause of these changes, we should expect that the two periods would be identical. It is possible, however, that this small difference may result from a change in the condition of the sun analogous to a change which has been observed in the magnetism of the earth. The earth has the properties of a vast magnet of feeble intensity, and the position of its magnetic poles changes from century to century. In 1576 the magnetic needle at London pointed 11 degrees cast of north; in 1660 it pointed exactly north, and in 1810 it pointed 24 degrees west of north, since which time the needle has been slowly returning to the north. These observations indicate a movement in the magnetic poles of the earth, extending through it period of several centuries.

There are many facts which seem to indicate that the sun is endowed with a magnetic force similar to the earth; and if the sun is really a great magnet, the analogy of our earth would lead us to admit that the position of the poles of this magnet may be subject to a gradual change. Such a supposition would enable us to explain the small difference between the period of Jupiter and that of the solar spots.

If Jupiter exerts so palpable an influence upon the sun's luminous envelope, then we should anticipate a sensible influence from several of the other planets. If the influence of the different planets upon the sun is supposed to follow the generally received law of gravitation, then if we represent the effect of Jupiter upon the sun by 100, that of Venus will be represented by 14, that of Saturn by 9, that of the earth by 8, that of Mercury by 4, and that of Mars, Uranus, and Neptune by less than unity. We are thus naturally led to inquire whether Venus exerts a sensible influence upon the solar spots. The periodic time of Venus is 7½ months, and a careful measurement of the area of the solar spots has shown that the amount of spotted surface upon the sun is subject to a small inequality having a period of 7½ months; and the amount of this inequality is about one-tenth of that ascribed to Jupiter, which is a near approach to the ratio above computed. We find also that the amount of spotted surface upon the sun is subject to a small inequality having a period of 12 months, and this inequality (as deduced from several years' observations) is more than one-tenth of that ascribed to Jupiter, which somewhat exceeds the influence above computed for the earth. The effect of Saturn is apparent in modifying the action of Jupiter. Two revolutions of Saturn are very nearly equal to five of Jupiter; so that after five revolutions of Jupiter (ranking a period of somewhat less than 60 years) the two planets return again to nearly the same relative positions. This gives rise to large disturbances of the sun's surface at intervals of nearly 60 years, and to smaller disturbances during the intermediate period.

Not only are the solar spots most numerous and extensive upon those years in which great auroral displays are most common, but the most remarkable auroral displays have usually been attended by an unusual and nearly simultaneous exhihition of solar spots, if the aurora were the immediate effect of the spotted condition of the sun. If we select the most remarkable auroras of the present century, and compare the condition of the sun's surface during a few days preceding and a few days following the aurora, we shall find that the solar spots were more extensive before than after these auroral displays, and that the spots were most remarkable two or three days before the aurora. The great auroral display of August 28, 1859, was specially remarkable on this account, the solar spots during the week preceding the aurora having been more extensive than they had been for many previous years. Rapid changes were seen to take place in the appearance of these spots, and two observers, independently of each other, noticed patches of intensely bright light to move across a large spot nt the very moment when the magnetic disturbance commenced at Greenwich; and a few hours afterward there succeeded one of the most remarkable magnetic storms, which was felt sitnultaneously over the entire northern and southern hemispheres.

Moreover, if we select all those days in which a very unusual disturbance of the magnetic needle was recorded at the magnetic Observatory of Greenwich, and note time condition of the sun's surface far a few of the preceding and following days, we shall find that the solar spots were generally more extensive before than after these magnetic disturbances; and the greatest exhibition of solar spots preceded by one or two days the unusual disturbances of the magnetic needle. These facts will scarcely permit us to doubt that an unusually disturbed condition of the sun's surface is one of the causes (if not the principal cause) of magnetic disturbances and also of great auroral displays upon the earth.

We seem then naturally conducted to the following hypothesis: not only the earth but each of the planets and also the sun is endoved with a magnetic force, having poles which at each instant occupy a determinate position; but this position is subject to a slow change front year to year. As these magnetic bodies advance in their orbits, each body disturbs the magnetism of every other body in the solar system. The disturbance of the sun's magnetism gives rise to commotiona in its luminous envelope, causing openings of variable extent, and these disturbances follow periods corresponding to the times of revolution of the disturbing bodies. Jupiter is the great disturber, and accordingly the solar spots exhibit alternately maxima and minima at intervals corresponding to the time of one revolution of Jupiter. Saturn exerts a small but neverthelesss appreciable influence, resulting in unusually large disturbances of the sun's envelope at intervals of five revolutions of Jupiter. Venus and the earth also produce small disturbances of the sun's envelope, causing small undulations in the curve which represents the amount of spotted surface of the sun. The hypothesis thus stated enables us to explain with tolerable precision the principal fluctuations in the sun's luminous envelope during the last 150 years.

These disturbances of the sun's surface are accompanied by disturbances in the electrical condition of the earth. The phenomena might perhaps be best explained by supposing a flow of electricity from the sun to the earth, and that this flow is proportioned to the extent of the disturbance of the sun's surface as indicated by the prevalence of dark spots. If such an hypothesis should be thought inadmissible, then it seems necessary to suppose that during these periods of unusual solar disturbance the sun's magnetism acts with unusual efficiency upon the earth, decomposing its natural electricities, causing an accumulation of positive electricity about one magnetic pole and negative electricity about the opposite magnetic pole. This would lead to grand auroral displays recurring at intervals corresponding with the periods of the solar spots.

By the daily rotation of the earth the position of the great solar magnet with reference to a magnetic needle upon the earth, is continually changing, and this causes a daily oscillation in the position of our magnetic needle. An unusual disturbance of the electricity of the earth causes corresponding disturbances in the position of the magnetic needle, and thus the mean daily range of the magnetic needle exhibits fluctuations whose periods correspond to those of the solar spots.

It will thus be seen that our hypothesis conneets together in a simple manner three different phenomena, which apparently are quite unlike, and enables us to render a satisfactory account of their principal peculiarities. If this hypothesis is correct in its essential features, then we can no longer regard auroral displays as exclusively atmospheric phenomena, but they are to a great extent the result of the influence of celestial forces, while their movements are controlled by the magnetic power of the earth. We should then naturally expect that opposite electricities would be driven toward the opposite magnetic poles of the earth, and that the system of circulation of electric currents would be, not such as is exhibited in Figure 20, but such as is shown iu Figure 21, where N and S are supposed to represent the north and south magnetic poles of the earth, a and s the poles of an imaginary magnet representing the magnetism of the earth. The east and west bands represent auroral arches, upon which stand auroral streamers. The dotted lines represent magnetic curves passing from auroral streamers in the southern hemisphere to streamers in the northern hemisphere, showing the path pursued by the currents of electricity in passing from one hemisphere to the other above the atmosphere. This is understood to be the system of circulation advocated by Mr. B. V. Marsh of Philadelphia. It agrees substantially with that represented by Figure 20, so far as the phenomena can be observed in the northern hemisphere; but they lead to different results in the southern hemisphere. We have not the requisite observations from the southern hemisphere to enable us to decide between these two hypotheses. Such observations might easily be made upon the telegraph wires of Australia; and if during some future auroral display such observation, could be obtained, they would furnish a true experimentum crucis to decide between the two hypotheses.

The hypothesis which has now been stated readily explains the simultaneous displays of great auroras in both hemispheres. We can not explain the great auroral displays in the northern hemisphere by supposing that the electricity of the atmosphere is temporarily diverted from one hemisphere to the other, for the mean daily range of the magnetic needle exhibits its maxima simultaneously in both hemispheres; neither can we suppose that the absolute amount of electricity for the entire globe, as developed by evaporation from the water of the ocean, should undergo great periodical variations, fur the mean temperature of the earth's surface does not change sensibly from one year to another. But if these great auroral displays result from the direct action of the sun, through the agency of its magnetism, such an effect should take place simultaneously in both hemispheres, conformably with the results of observation.

It is not claimed that the hypothesis which has here been proposed to explain the inequalities In the frequency and brilliancy of auroras is to be regarded as fully established. Further researches and discoveries may require us to modify it in some important particulars, or even to abandon it altogether. Such an hypothesis, however, is not without its value, since it enables us to clarify the known facts, and even to predict results which have not hitherto attracted the attention of observers. The true philosopher will not undervalue hypotheses, which have often proved of great value in the promotion of science, but he should be ready to abandon any hypothesis as soon as the progress of science shows it to be no longer tenable.