of a base of convenient length and having telephonic connection.  The instruments used were altazimuths, constructed under the direction of Prof.  Mohn, specially for measuring the parallax of the aurora borealis.  A full description of these instruments and of the calculations will be found in the Acta Reg Soc.  Sc.  Ups., 1884.  The results now in question are based upon nearly 1,500 measurements of heights; the motions will form the subject of a future paper.  It was found that clouds are formed at all levels, but that they occur most frequently at certain elevations or stages.  The following are, approximately, the mean heights, in feet, of the principal forms:  Stratus, 2,000; nimbus, 5,000; cumulus (base) 4,500, (summit) 6,000; cumulo-stratus (base), 4,600; “false-cirrus” (a form which often accompanies the cumulo-stratus), 12,800; cirro cumulus, 21,000; cirrus, 29,000 (the highest being 41,000).  The maximum of cloud frequency was found to be at levels of 2,300 and 5,500 feet.

Generally speaking, all the forms of cloud have a tendency to rise during the course of the day; the change, excepting for the cumulus form, amounting to nearly 6,500 feet.  In the morning, when the cirrus clouds are at their lowest level, the frequency of their lowest forms—­the cirro-cumulus—­is greatest; and in the evening, when the height of the cirrus is greatest, the frequency of its highest forms—­the cirro-stratus—­is also greatest.  With regard to the connection between the character of the weather and the height of the clouds, the heights of the bases of the cumulus are nearly constant in all conditions.  The summits, however, are lowest in the vicinity of a barometric maximum.  They increase in the region of a depression, and attain their greatest height in thunderstorms, the thickness of the cumulo stratus stretching sometimes for several miles.  The highest forms of clouds appear to float at their lowest levels in the region of a depression.  The forms of clouds are identical in all parts of the world, as has been shown in papers lately read by the Hon. R. Abercromby before the English and Scottish Meteorological Societies.—­Nature.

* * * * *

ON THE CAUSE OF IRIDESCENCE IN CLOUDS.

By G. JOHNSTONE STONEY.

When the sky is occupied by light cirro-cumulus cloud, an optical phenomenon of the most delicate beauty sometimes presents itself, in which the borders of the clouds and their lighter portions are suffused with soft shades of color like those of mother-of-pearl, among which lovely pinks and greens are the most conspicuous.  Usually these colors are distributed in irregular patches, just as in mother-of-pearl; but occasionally they are seen to form round the denser patches of cloud a regular colored fringe, in which the several tints are arranged in stripes following the sinuosities of the outline of the cloud.