intensity when the red is very brilliant, so also is the green. The red remains the brighter of the two tints, and is also that which fades last when fog obliterates the aurora by degrees. Independently of the relative intensity of the two tints, the fact that red light penetrates fog more easily than green may have something to do with this.

In very brilliant auroras the intermediary yellow tint may disappear altogether, and only the red and green remain. This frequently happens in the crown more particularly, of which the centre is then green while the whole periphery is red.

We have said above that isolated rays are frequently entirely red; it is, on the contrary, very rare to see rays which are entirely green. This phenomenon only presented itself once to the French Commission during the whole of its stay at Bossekop; on January 2, 1839, at 8.45 P.M., an arc composed solely of green rays was observed.

According to some observers it would seem that the green colour is in certain auroras replaced by a blue or violet tint; but this last colour is extremely

rare.

The brilliancy of the colours of the aurora seems to have a definite relation to the state of the atmosphere. In high latitudes particularly, the observations of Sir John Franklin, McClintock, Weyprecht, &c., have shown that the colouring of the aurora is less strong when the air is very pure, and increases

when the atmosphere becomes foggy and its transparency is diminished. Another circumstance which should be mentioned in this connection is that the auroras in the form of drapery, which are generally the richest in colour, are usually seen only in regions where the seas are open in winter and free from ice, and where, consequently, fogs are of very frequent occurrence; such are the north of Norway, Spitzbergen, and Newfoundland.

2. Intensity of the light of polar auroras.-The intensity of the light emitted by polar auroras is usually feeble, even in the most brilliant auroras. By the light of a very bright boreal crown, Bravais could hardly read a few words printed in the character known to printers as small text; it was very easy, on the contrary, to read the same characters by the light of the full moon, of which the apparent dimension is however much smaller than that of the aurora. The greater number of scientific men who have observed auroras in polar regions, where they are brightest, Parry, Bravais, Kane, Hayes, Weyprecht, Nordenskioeld, agree in the statement that the total illumination produced by the finest auroras is generally inferior to that of the full moon, and rarely exceeds that of the moon in her first quarter. It is clear, then, that on an equal superficial area the brilliancy of the polar aurora is far inferior to that of the

moon.

An indirect proof of this fact is found when we

note the periods of the appearances of the aurora. Their frequency always diminishes when the moon is full, which shows that the general illumination of the sky produced by the moon at the full completely drowns a great number of auroras and prevents their being visible. This influence of the age of the moon. on the apparent frequency of the aurora borealis had already been noted by Mairan, who found, on an average, three times as many auroras visible in that half of the lunar month which comprehends the new moon—that is, from the beginning of the last quarter to the end of the first-as in the half which comprehends the full moon, from the end of the first quarter to the beginning of the last.

There are, however, exceptional auroras which are sufficiently brilliant to appear such even when the full moon is shining; witness the aurora which was observed all over Central Europe on February 4, 1872. It has occasionally been averred that auroras have been seen in broad daylight, but the fact appears to be more than disputable. It is probable that there has been some confusion between the aurora and certain cirrus clouds which resemble it, and may even be associated with it, as we have already shown.

Another proof of the faint light given by the aurora is to be found in the facility with which the light of the stars penetrates it, without sensible diminution. The stars of the first and second