METEOROLOGY. 377 



zone, above which the ordinary blue or purple sky is visible, while be- 

 low we have yellow, then orange, and on the horizon brownish red. 



2. Opposite to the above is a dark segment which is simi)ly the asliy- 

 tinted shadow of the earth, and which is sharply distinguished from 

 the remaining portion of the heavens. 



3. A circular region of considerable diameter of a rosy-red that is 

 generally described as the purple light; this develops above the bright 

 segment a long time before sunrise or after sunset, so that the lower 

 part of this region seems to be hidden behind the bright segnieut. 

 The center of this region sinks rapidly after sunset, while at the same 

 time its radius increases so that finally the limit of the borders of the 

 puri)le region unites with that of the bright segment, the impression 

 being as if the purple light slipped in behind the bright segment. The 

 puri)le light plays the part of a very much magnified, very mucli diluted, 

 image of the sun ; at the time of its greatest development the general 

 brightness is increased so that objects on the earth's surface that were 

 invisible soon after sunset become again visible. The maximum of this 

 second illuminatioD occurs in the Alps when the sun is four or five de- 

 grees below the horizon, at which time the center of the ])urple light 

 is about 18° above the horizon, while its highest point reaches u[) to 40° 

 or 50O. 



As soon as the purple light has completely disappeared behind the 

 bright segment there appears a second dark segment on the ()pi)osite 

 side of the sky. Soon there further develops a second bright segment 

 above the slowly sinking first one, and only with difliculty distinguished 

 therefrom ; wirh very clear skies one can later occasionally also observe 

 a second purple light, and therewith a temporary increase of bright- 

 ness, so that a third illumination of objects favorably located can be ob- 

 served. [It would seem that the occurrence of two separate i)eriods of 

 the red or rosy or ])urple tint, as was observed in the winter of 1883-84 

 at many places throughout the world, is therefore simply an unusual de- 

 vel<)i)ment of an ordinary phenomenon. The twilight phenomena of 

 1883-84 have been specially due to the remarkable (level()i)ment of the 

 second purple light, which has frequently surpassed the first purple 

 light.] {D. M. Z., I, p. 32.) 



381. Dr. G. Elellmann contributes some observations on twilight made 

 by himself in Spain, 1875 to 1877, and follo)vs this by a comparison with 

 previous authorities, from allot which he deduces the following results: 



1. The depression of the sun under the horizon at the end of the as- 

 tronomical twilight is not, as has been generally assumed, constant at 

 about i8 degrees. 



2. The depression show 3 a decided annual period, a maximum in win- 

 ter and miniumm in summer. 



3. The depression is greater in the morning than in tln^. evening. 



4. It shows an intimate connection with the relative humidity of the 



