This explanation might be subjected to a severe proof, were it possible to measure the diameters of halos as accurately as we do those of the colors of the rainbow; but unfortunately the phenomenon is seldom presented with sufficient regularity and distinctness for such a measurement.

Nevertheless, M. Arago has made another kind of observation, which proves with certainty that the light of halos is refracted light; for, studying by a peculiar process the state in which this light is found with respect to polarization, he ascertained that it is always polarized by refraction and not by reflection.

The exterior halo may be explained in many ways, but we shall not hazard a theory, while so uncertain as to its true dimensions.

Without doubt, many interesting researches are still to be made on this phenomenon, but, after the observation of M. Arago, we may affirm that the condition necessary to its production, is the presence of icy particles in the higher regions of the atmosphere. This conclusion is of importance, as bearing on the temperature of the air at a great height in different seasons of the year; and we would advise observers to attend to the temperature of the air at the time of halos, whether round the sun or moon. During the experiment of M. Arago, the surrounding temperature was 15° [centigrade] at the Observatory in Paris.

VI. PARHELIA OR FALSE SUNS.

These im

PARHELIA are the simultaneous appearance of several suns. ages always appear at the same height with the real sun, and are always connected together by a white horizontal circle having its pole at the zenith. This circle ascends and descends with the real sun, and its apparent semi-diameter is always equal to the distance of the sun from the zenith. Those images which appear on the same side of the circle with the real sun, present the colors of the rainbow; and sometimes those parts of the circle which are near them are also colored. On the contrary, the images formed opposite to the real sun, are always colorless; whence we may conjecture that these images as well as the great circle are produced by reflection, and the others by refraction. Besides, at the time of these phenomena, one or several circular concentric coronæ, having the colors of the rainbow, may ordinarily be seen about the sun; and, finally, we sometimes observe in these very coronæ, or in points of the great circle, other traces of rainbows, and even entire rainbows. The most complete meteor of this kind on record is that observed by Hevelius at Dantzic, on the 20th of February, 1661.

In order to understand the process by which this phenomenon is produced, we must, as Huygens did, first consider the white horizontal circle on which the real sun is always found. The uniform whiteness of this circle proves it to be produced by reflexion; thus the problem is reduced to this:

Suppose an infinite number of corpuscles suspended in the air; what form must we attribute to them in order that the solar rays, reflected from their surfaces, may form invariably the same angle with the horizon as the incident rays from which they are derived? It is evident that this condition can be fulfilled only by giving to these corpuscles the form of small vertical cylinders; and indeed, if we suppose the sun to shine upon an infinity of such cylinders, the resulting effect must be a white horizontal circle having its pole in the zenith, its semi-diameter being the complement of the sun's altitude. Now to account for the phenomenon of the colored suns which appear on either side of the real sun, it is sufficient to suppose these cylinders composed of an outer transparent part and an opaque cylindrical nucleus; for then, by a lateral refraction taking place perpendicularly to their axes, they will produce an effect analogous to that of the globules of hail in the coronæ, and with more light, on account of their elongated form and their parallelism.

Finally, if we suppose, as is very likely, that the extremities of these cylinders are both rounded, they will produce in this direction the effects resulting from sphericity, and hence may arise the colored coronæ concentric to the real sun. Now Descartes assures us, in his book on Meteors, that he has sometimes observed such cylinders of hail containing an interior nucleus, snowy, opaque, and in like manner cylindrical. Indeed, Huygens has, as it were, imitated this formation by an experiment, placing at different angular distances from his eye, and from the sun, a cylinder of thin glass filled with water, with an opaque cylindrical nucleus in the interior, and he has thus been able to verify by experiment all the phenomena indicated by the calculation. He has also shown how these calculations would faithfully represent the characteristic circumstances of the phenomenon. But in order to arrive at all the minute particulars observed by Hevelius, it was necessary for him to distribute in the atmosphere, in a variety of positions, the cylindrical and globular particles which he had imagined. This complication, which would seem to belong to this kind of phenomena, is no reason for rejecting the idea of Huygens, but rather an inducement to observe their appearances exactly for the purpose of making the comparison. The law of double refraction, so long misunderstood, has taught us that we must not treat lightly the speculations of so great a man; and Newton himself seems to have adopted them, in this case, since in speaking of parhelia, in his Optics, he refers us to the explanation of Huygens.

VII. LIGHTNING-RODS.*

THE most advantageous form that can be given to lightning-rods appears evidently to be that of a very sharpe cone; and the higher it is elevated in

*From a Memoir of M. Gay-Lussac. Annales de Chime.

the air, other circumstances being the same, the more its efficacy will be increased, as is clearly proved by the experiments with electrical kites, made by M. Romas and M. Charles.

It has not been accurately ascertained how far the sphere of action of a lightning-rod extends, but, in several instances, the more remote parts of large buildings on which they have been erected, have been struck by lightning at the distance of three or four times the length of the conductor from the rod. It is calculated by M. Charles, that each lightning-rod will effectually protect a circular space, whose radius is twice the height of the conductor; and they are now attached to buildings according to this principle.

A current of electric matter, whether luminous or not, is always accompanied by heat, the intensity of which depends on the velocity of the current. This heat is sufficient to make a wire red-hot, or to fuse or disperse it, if sufficiently slender; but it scarcely raises the temperature of a bar of metal, on account of its large mass. It is by the heat of the electric current, as well as by that disengaged from the air, condensed by the passage of the lightning through it, when not conveyed by a good conductor, that buildings struck by it are frequently set on fire.

No instance has yet occurred of an iron bar rather more than half an inch square, or of a cylinder of the same diameter, having been fused, or even heated red-hot by lightning. A bar of this size would therefore be sufficient for a lightning-rod; but as its stem ought to rise from 15 to 20 feet above the building, it would not be strong enough to resist the action of the wind, unless the lower part were made much thicker.

An iron bar about three quarters of an inch square, is sufficient for conductors. It might even be made still smaller, and consist merely of a wire, provided it were connected at the surface of the ground with a bar of metal, about half an inch square, immersed in water, or in a moist soil. The wire indeed would pretty certainly be dispersed by the lightning, but it would direct it to the ground, and protect the surrounding objects from the stroke. However, it is always better to make the conductor so large as not to be destroyed by the stroke; and the only motive for substituting a wire for a stout bar is the saving of expense.

The noise of the thunder generally occasions much alarm, although the danger is then passed; it is over, indeed, on the appearance of the lightning, for any one struck by it neither sees the flash nor hears the report. The noise is never heard till after the flash, and its distance may be estimated at so many times 1136 feet as there are seconds between the appearance of the lightning and the sound of the thunder.

Lightning often strikes solitary trees, because, rising to a great height, and burying their roots deep in the soil, they are true lightning-rods; and they are often fatal to the individuals who seek them for shelter, since they do not convey the lightning with sufficient rapidity to the ground, and are worse conductors than men and animals. When the lightning reaches

the foot of the tree, it divides itself amongst the neighbouring conductors, or strikes some in preference to others, according to circumstances. Sometimes it has been known to kill'every animal that had sought shelter under the tree; at others, only a single one out of many has perished by the stroke.

A lightning-rod, on the contrary, well connected with the ground, is a certain security against the effects of lightning, which will never leave it to strike a person at its foot; though it would not be prudent to station one's self close to it, for fear of some accidental break in the conductor, or of its not being in perfect communication with the ground.

When the lightning strikes a house, it usually falls on the chimneys, either from their being the most elevated parts, or because they are lined with soot, which is a better conductor than dry wood, stone, or brick. The neighborhood of the fireplace is consequently the most insecure spot in a room during a thunder-storm. It is best to station one's self in a corner opposite the windows, at a distance from every article of iron or other metal of any considerable size.

Persons are often struck by lightning without being killed; and others have been wholly saved from injury by silk dresses, which serve to insulate the body, and prevent the access of the electric matter.

The stem, or part of the rod above the building, should be a square bar of iron, tapering from its base to the summit, in the form of a pyramid. For a height of from 20 to 30 feet, which is the mean length of the stems placed on large buildings, the base should be about 24 inches square.

Iron being exposed to rust by the action of the air and moisture, the point of the stem is liable to become blunt; to prevent this, a portion is cut off from the upper end, about 20 inches in length, and replaced by a conical stem of brass or copper, gilt at its extremity, or terminated by a small platina needle, two inches long.* The platina needle should be soldered with silver solder to the copper stem; and to prevent its separating from it, which might sometimes happen notwithstanding the solder, it is secured by a small collar of copper. The copper stem is united to the iron one by means of a gudgeon which screws into both. If the gilding of the point cannot easily be performed on the spot, nor the platina be readily obtained, we may dispense with both without any inconvenience, and employ only a plain, conical, copper stem. Copper does not rust in the air to any considerable depth, and even if the point becomes somewhat blunt, the rod will not thereby lose its efficacy.

Below the stem, three inches from the roof, is a cap, soldered to the body of the stem, and intended to throw off the rain-water, which would flow down the stem, and tend to injure the building.

* Instead of a platina needle, one of standard silver may be substituted, composed of nine parts of silver, and one of copper.

Immediately above the cap, the stem is rounded for about two inches to receive a split collar, with a hinge and two ears, between which the extremity of the conductor of the lightning-rod is fixed by a bolt. Instead of the collar, we may make use of a square stirrup, embracing the stem closely. The stem of the lightning-rod is fixed on the roof of buildings, according to circumstances. If it is to be placed above a rafter, the ridge must be pierced with a hole through which the foot of the stem passes, and is steadily fixed against the king-post by means of several clamps. This disposition is very firm, and should be preferred if circumstances admit of it.

If the stem be fixed on the ridge, a square hole must be made through it of the same dimensions as the foot of the stem; and above and below we fix, by means of bolts, or two bolted stirrups, which embrace and draw the ridge together, two iron plates about three-quarters of an inch thick, each having a hole corresponding to that in the woodwork. The stem rests by a small collet on the upper plate, against which it is strongly pressed by a nut, made to screw on the end of the stem against the lower plate.

Lastly, if the lightning-rod is to be fixed on a vaulted roof, it should be terminated by three or four feet, or spurs, which must be soldered into the stone, with lead, in the usual manner.

The lower part of the conductor should be an iron bar or rod about threequarters of an inch thick, reaching from the bottom of the stem to the ground. It should be firmly united to the stem by means of a collar, screw, or bolt, and its several parts should be connected together in a similar manner. After penetrating into the ground for about two feet it should be bent at right angles to the wall of the building, and after being carried in that direction for twelve or fifteen feet, it should be made to communicate with a well, drain, aqueduct, or permanently moist earth. If the soil be dry, it should extend to the depth of twelve or fifteen feet; and to secure it from rust, it should be surrounded with charcoal, which is indestructible, and which, while it preserves the iron, facilitates the passage of the electricity into the ground by its conducting property.

Both the bottom and top of a lightning-rod are sometimes made to terminate in several branches, and its efficacy is thus increased. It is also recommended to connect with the lightning-rod any large masses of iron that may be in the building, as metal pipes and gutters, iron braces, &c.; without this precaution the lightning might strike from the lightning-rod to the metal, especially if there happened to be any interruptions in the former, and thus occasion serious injury to the building, and danger to its inhabitants.

In the case of powder-magazines, the lightning-rod should not be attached to the building, but to poles eight or ten feet from it. If the building be large, several should be used, arranged according to the rule, that a lightning-rod may be considered as protecting a circular space whose radius is twice the height of the rod. If the magazine be in a tower or other very