approach to him, and can never bring the slightest trouble to the solar system. But, if its course had chanced to be directed full upon the sun, the meteoric masses doubtless forming its head and train (not tail), falling in countless millions upon him at the rate of more than 300 miles per second, when they crossed his visible surface, and probably passing deep below that surface with ever and most rapidly-increasing velocity to reach his real nucleus, would have generated an intensity of heat far exceeding that which he constantly emits. The increased emission might not have lasted a month, or even a week, but it would have sufficed.

So, again, what we now know of cometic structure leads us to believe that the comet of 1858, called Donati's, whose head is shown in Fig. 2, would have proved a very dangerous visitor had its course led it directly towards the nucleus of the sun. Fortunately, the chance of any comet visiting our system from interstellar space, travelling directly towards the sun, is so small, that it may be reckoned "almost at naked nothing." As to comets already belonging to our system, if any such have orbits passing very close to the sun, so as to be checked in their career at every perihelion passage, it is clear (from the continuance of life during many hundreds of thousands of past years on the earth) that the mischief must long since have been taken out of them-unless we suppose (which is incredible) that the last perihelion passage of such a comet preceded the beginning of life on the earth.

HAS THE MOON AN ATMOSPHERE?

BY MR. RANYARD.

No lunar clouds are ever seen floating over the sharply-defined objects of the lunar landscape. As the sun rises upon the moon, the transition from darkness to light is quite sudden, no zone of twilight stretching beyond the sun-illuminated region can be detected, and as the moon passes over bright stars, there is no gradual diminution of their light as they approach the lunar limb, but they disappear suddenly, as if snuffed out, at the instant that the lunar limb covers them. It is quite certain, therefore, that the moon has not a vapour-laden atmosphere which is as dense, or nearly as dense, as the atmosphere which envelopes the earth. There is no evidence tending to show that rays of light are bent as they graze the lunar limb. Rays grazing the earth's surface in a similar manner would be turned through an angle of more than a degree from their original course by refraction in the earth's atmosphere, for it is known that stars, when seen upon the horizon, are raised more than half a degree above their true places by the bending which rays of light undergo before they reach the observer. If there were any such refraction of light by a lunar atmosphere, the sun would never be entirely hidden by the dark body of the moon

during an eclipse. During the recent total eclipse, as seen from Sohag, the moon's diameter was about half a minute of arc greater than the sun's diameter, and at the time of central eclipse, when the centre of the moon's disc coincided with the centre of the sun's disc, the sun's limb would have remained visible if there had been a lunar atmosphere giving a refraction equal to 1-240th part of the refraction which would have been caused by the earth's atmosphere. But not only was the sun's limb hidden at the time of central eclipse, but it remained hidden for more than a minute-in fact, the observed duration of totality agreed within a few seconds with the duration calculated on the supposition that there is no bending of light by a lunar atmosphere.

There is always some discordance amongst observers as to the actual duration of totality, for the disappearance of the sun's light is not an absolutely instantaneous phenomenon, so that occasionally trained observers standing beside one another differ as much as three or four seconds in their estimates of the time of the commencement of the total phase. There is also some uncertainty as to the real diameters of the sun and moon. But, after making allowance for these uncertainties, we may safely assert that the observed. duration of total eclipses agrees with the calculated duration sufficiently closely to enable us to be sure 'that there is not an atmosphere about the moon capable of causing a refraction equal to one twothousandth part of the refraction which would be

caused by the earth's atmosphere under similar circumstances.

But this does not prove that there is no lunar atmosphere, or even that there is not a lunar atmosphere equal to one five-hundredth part of the terrestrial atmosphere, for the mass of the moon is less than one-eightieth part of the earth's mass, so that the weight of a body at the moon's surface is less than one-sixth of the weight it would have upon the earth's surface. And an atmosphere like our own would, if removed to the moon (the temperature remaining the same), be more loosely packed, and would extend to a much greater height above the surface.

There is another method by which the existence of a lunar atmosphere may possibly be detected. Besides bending rays of light, the earth's atmosphere absorbs certain wave lengths more than others, so that in the tpectrum of sunlight which has passed through the lower strata of the atmosphere for some distance, as at sunrise or sunset, there are certain lines which are not to be found in the spectrum of a high sun. M. Thollon, who was one of the observers of the total eclipse at Sohag, has spent many years in making a great map of the solar spectrum, in which he has laid down the position and relative intensities of the solar lines, and he has also mapped down the places of many hundreds of atmospheric lines which are proved not to belong to the true solar spectrum, because they grow more intense as the sun sinks towards the horizon, and vary with the amount of aqueous vapour present in the atmosphere.

The instrument with which M. Thollon has made his observations is a bisulphide of carbon spectroscope, giving a dispersion equivalent to thirty-one prisms of 60° of ordinarily dense flint glass, and giving a much brighter spectrum and far better definition than could be obtained with such a cumbersome train of prisms, if it were possible to mount them and keep them in adjustment. M.Thollon has in a very simple manner contrived to make the solar rays pass three times through a system

1130

1220 1230

The B group as seen in M. Thollon's Instrument.

of prisms and half-prisms, so that the eye-piece of the viewing telescope and the slit remain fixed, while, by a motion of the prisms, different parts of the spectrum can be brought into view. M. Thollon has made with this instrument a map of the solar spectrum 49 feet long, showing more than 4,000 lines.

The wood-cut, copied from one of M. Thollon's drawings, represents the well-known B group, which lies towards the red end of the spectrum. It is a