means of a chronographic diapason, one can determine, with precision, the exact number of the revolutions of the wing which are effected in a second. That which I used, gave a graphic delineation of five hundred simple vibrations per second.

A continual rubbing of the wing on the cylinder, presents a resistance to this organ, which retards its frequency; so in order to have the nearest approach to the truth, I selected those drawings in which the contact of the wing with the cylinder was at a minimum, so that the diagrams were reduced to a series of points.

The frequency of the movements diminishes, also, when we load the wing with a little weight. It diminishes, equally, by fatigue, and the action of cold. Everything occurs in this case, as in the rythmic movements of the muscular system in different animals. Under equally favourable conditions for observation, the frequency of the beats which different species of insects produce, brings before us curious results. I have only been able to observe a small number of species, because of the lateness of the season. These are the numbers I have found for each second :

A more complete study of a great number of well-determined species would, doubtless, furnish much higher figures, as the maximum frequency.

I wish to add that the wing movements in this sort of captivity on account of the greater resistance will be reduced in number. My figures then must be below those representing vibrations in a free flight.

B. Form of the movements of the wing:

The graphic method does not answer very well to determine the course of the wing at each of its revolutions; indeed the tracings which the point of the wing of an insect describes in space are wrong; they are inscribed on the surface of an ideal sphere, which has for its radius the length of the wing, and for its centre the point at which this organ is implanted in the thorax of the insect. A spherical surface of this nature could only be tangential at one * We do not know what species is designated by the French name.

point to the surface of the registering cylinder, and every fuller contact risks deforming the drawing more or less, in re-producing the curvature of the wing. To obtain an exact notion of the course of the wing in space, I have had to resort to Wheatstone's optical method. It is well known that this celebrated English physicist terminates vibrating rods with bright metallic balls, whose gleam leaves upon the retina persistant impressions of the periodical movements they execute.

By fixing with varnish a little piece of gold leaf to the end of an insect's wing, and placing the animal in a ray of sunlight, I obtained a bright luminous image in the form of the figure 8, which indicates the different points in space traversed at each revolution, by the gilt spot.

Among different sorts of insects, I have almost always met with the same form.

Resuming then the graphic method to verify this result, I have succeeded in obtaining successively portions of the drawing, in some cases giving me the upper loop of the 8, in others the lower, and in others the double point, where is the intersection of the two halves of the 8.

By way of further confirmation, I have sought to register the contact of the wing with the cylinder, not only by its point, but by its anterior margin. The theory anticipates that under these conditions, the figure 8 ought to disappear, and in its place one should obtain a double contact of the wing with the cylinder. One of these contacts took place at the instant the upper loop of the 8 was formed, and at the point where this loop presents its convexity to the cylinder. The other contact took place where the lower loop was formed under the same conditions.

In a future paper I shall show that this complex movement does not result from a series of periodic muscular acts executed by the insect, acts which would produce in one case a simple oscillation in a vertical direction, whilst in the horizontal direction, other muscles would produce, at the same time, two oscillations. In reality, the insect only executes one movement of lowering the wing, to which succeeds a movement of elevation, and if in consequence of these two contrary movements, the wing is not limited to oscillate in one plane, this results from the resistance of the air, which imposes upon the wing a deviation in each half of its course.

GLACIERS OF THE CAUCASUS.

THE "Archives des Sciences" for January 7th, 1869, contains some observations on the Glaciers of the Caucasus, by M. Ernest Favre, introducing a paper by M. Boleslas Statkowski, on the causes of the avalanches of the glacier of Kosbek.

M. Favre contrasts the Caucasus with the Alps, the former though narrower greatly exceed the latter in height. The bold truncated forms of Elbruz and Kosbek, together with the serrated ridge, which separates these two giants strikes the imagination, but penetrating into the interior of this region somewhat effaces the first impression. Extensive views are rare, the horizon is bounded by great escarpments, and it is necessary to climb to greater heights, than in the Alps, in order to contemplate vast panoramas like those which constitute the beauty of the scenery of the latter. The valleys are much shut in, and difficult of access in their upper portions, and the glaciers seldom descend lower than from 2,000 to 2,400 metres, while many of the Alps reach a lower level of from 1,000 to 1,300 metres. M. Abich calls many of the Caucasian valleys "cauldron valleys." They descend from circles of high mountains, and form vast basins with very narrow outlets. The same authority gives an height above the sea level of several glaciers; facing the north, those of Bekson, Zea, Khaltschi, Devdoroc have their bases at elevations of 7,070, 6,575, 5,702, 7,540 feet respectively; facing the south, Kilde, Tschkharr, and Zzauner 7,912, 7,935, 6,612 feet respectively.

These glaciers only possess the character of ice rivers for a length of a few thousand yards. Descending from the snow-fields -nevées-with a more or less rapid slope, they soon reach the limits of vegetation, and stop soon after it is passed.

In Souanetti, M. Abich estimates the snow line at 9,527 feet above the sea, and he places the limit of arborescent vegetation at 7,298 feet. M. Radde, Director of the Museum of Tiflis, as the mean of numerous observations places it at 7,600 feet. The mean height of the base of the glaciers of the first order is 7,198 feet, so that they descend on the average about 188 metres below the limit of eternal snow, and only 58 metres below the line of arborescent vegetation.

M. Abich reports the existence of a great erratic deposit of trachytic and crystalline rocks, at a height of 9,200 feet, in a mass of secondary rocks separated from the central chain by a deep valley situated between the valleys of Ardon and Terek.

The glacier of Devdoroc descends on the N.E. flank of Kosbek. It is four versts long, and four or five hundred metres in its greatest width. It takes its rise in a grand plateau of snow-fields two versts wide, and from 12,000 to 13,000 above the sea level. It becomes rapidly narrow, and suddenly takes to a steep incline, which it continues for 10,500 feet, presenting a great number of peaks (aiguilles) and deep crevasses. It then slopes more gently and spreads out in the valley to a width of 8,200 feet, after which it narrows again and terminates at 7,540 feet above the sea level. Two affluents separated from the glacier in their upper portions by great masses of trachyte and black schists, make a rapid fall and precipitate themselves on its right flank. They come down from vast snow-fields and terminate in narrow channels. They are only accessary branches to the principal glacier, to which they are united lower down by great masses of snow. This glacier is covered by moraines over the greater part of its surface, and the torrent of Amilichka which flows from the glacier has dug for itself a deep bed in the erratic deposits.

Like other glaciers of the Caucasus, which are similarly situated, that of Devdoroc produces at intervals tremendous avalanches. In 1882, M. Statkowski tells us that one of these avalanches filled the valley with ice, stones, and mud, which extended 2,170 metres, and reached a height of 300 feet. Captain Grauert, an engineer officer, who witnessed the catastrophe, estimated the volume of this mass at 1,600,000 cubic sagenes. This avalanche divided the Russian military road to Georgia, and obstructed the traffic for two years.

The gorge of Devdoroc is 375 metres wide, and contracts at the end of the glacier to 32 metres. A promontory covered with rocks, which are incessantly falling, advances into the valley from the left, and almost closes it. The glacier in its changes of volume must frequently reach this promontory, and as its motion is thus arrested it must form an ice dyke of considerable height, and give rise to a lake by damming up the Amilichka. The pressure goes on increasing, and ends in breaking the dyke, when an immense mass is suddenly precipitated in a long tortuous valley. Arriving at Terek it forms a barrier of ice, stones, and mud, like that of 1832. The snows accumulate and enlarge this avalanche, and if one of the great storms, so frequent in the summer, occurs at the the time of its bursting, the destruction is all the greater.

The inhabitants fancy that these avalanches occur about once in seven years, but facts do not coincide with this notion. The people * A sagene is about seven English feet.

in the valley know when the avalanche is coming, and in 1832, 1842, and 1845, the Russian officers say that they predicted it several weeks in advance, and took away their flocks.

Major Mylow states that some time before the avalanche falls, the ice, subjected to enormous pressure, makes a boiling noise, together with detonations and rattling sounds like peals of thunder. These noises and a change in the colour of the water of the Koboby from white to yellow and black, are the symptoms chiefly noted by the inhabitants.

Some

The violence of the storms adds to the mass of matter in motion, but does not produce the avalanche which is caused by water action, and M. Statkowski advises measures to secure a more regular flow to the Amilichka. During the storm of 1832 a sentinel in the gorge of the Terek was carried to a distance of 140 feet. times the noises and discolourations of the water occur without being followed by an avalanche. Either the force of the descending matter is not sufficient to break the obstacles, or the waters have found some means of escape.

SOLAR HEAT AS A MOTIVE POWER.

BY M. MOUCHOT.

("Comptes Rendus.")

ACCORDING to my experiments, it is easy to collect, at a cheap rate, more than three-fifths of the solar heat arriving at the surface of the globe. The intensity of this calorific source, so feeble in appearance, was revealed by Pouillet, more than thirty years ago. At Paris, a surface of one square metre, normally exposed to the sun's rays, receives, at least, whatever may be the season, during the greater part of a fine day, ten heat units (calorics) per minute.* To appreciate such an amount of heat, it is sufficient to observe that it will boil, in ten minutes, one litre of water, taken at the temperature of melting ice, and it is almost equal to the theoretical power of a one-horse steam-engine. Under the same conditions, a superficies of one are (119-603 square yards) would receive, during ten hours of insolation, as much heat as results from the combustion of 120

* The unit of heat adopted by English physicists is the quantity necessary to raise one pound of water from 0° to 1° C. We suppose M. Mouchot adopts the same standard.