590 



NATURE 



[October 20, 1892 



Generally at the same time stars become visible on the 

 photographic plate, through which, in association with 

 the time of photographing, the direction of adjustment of 

 the apparatus is ascertained (that is to say, the position 

 of the axle of the apparatus is ascertained). 



With regard to equatorial regions, it is of great import- 

 ance that the exact time in which luminous night-clouds 

 pass through them should be determined. According to 

 the observations hitherto made, the passing through the 

 equator may take place in the time between the beginning 

 of September and the end of October, and the return 

 between the beginning of March and the end of April. 

 Under 20'' S. lat., the time of passing through will, in that 

 case, be from the middle of September until the middle of 

 November, and from the middle of February until the 

 middle of April, and under id" N. lat. from about the 

 middle of March until the middle of M ay, and from the 

 middle of August until the middle of October. In con- 

 sequence of the daily rotation of the earth round its axis- 

 together with the distinct movements of the earth, 

 atmosphere, it may be that the passing through the 

 equator does not take place in the simple manner here 

 described. It does not seem to be unlikely that the 

 periods are not limited as exactly as stated. 



Moreover, it is probable that the luminous night-clouds 

 consist of a gas which is condensed in consequence of 

 the lower temperature prevailing in the altitude of 82 

 kilometres. On the question relating to the nature of this 

 . gas depend several other cosmical questions : for instance, 

 with respect to the temperature of the air of the mundane 

 space and the temperature of the atmosphere at the 

 altitude of 82 kilometres, which will be answered through 

 comparing experiments in the laboratory. For this 

 reason, spectrographs of the sunlight at low altitudes of 

 the sun, in the season in which the phenomenon of the 

 luminous night-clouds is seen, are of great value. Such 

 spectrographs should be taken in the evening shortly 

 before sunset, and in the morning shortly after sunrise. 



It appears that in the northern regions of the earth, in 

 about 70° latitude, in the period from the middle of June 

 until the middle of July, an especially great accumulation ' 

 of clouds takes place, which, however, on account of the 

 sun standing constantly above the horizon during this 

 period, will be hardly visible. It will, therefore, be of 

 special advantage for these regions to take spectrographs 

 of the sunlight at low positions of the sun. 



These short remarks regarding the importance of the 

 phenomenon with reference to cosmical problems may 

 serve to show that the observations necessary for the ex- 

 ploration of the subject are well within the sphere of as- 

 tronomers and geophysicists. There can be no doubt 

 that the observations necessary for the solving of these 

 questions are far beyond the capacity of a single institu- 

 tion. Those who take interest in the furtherance of the 

 questions we have indicated are therefore requested to 

 assist through one or other of the kinds of observation 

 above noted in the investigation of the luminous night- 

 clouds.^ 



W. FOERSTER. 



O. Jesse. 

 Berlin Royal Observatory, September 1892. 



adjoining pair seems to converge at one end, and to 

 diverge at the other, whereas in reality the lines are all 

 parallel. The first step in an explanation of the illusion 

 would be the determination of its essential factors, of its 

 various forms, and of some general principle embracing 

 under one formula its several varieties. The next step 

 would be to correlate this formulation with some recog- 

 nized psychological principle. The generalization is found 



yxxxy^ 



yy.ooo'. 



W^^^K'^K^^K^^K^^K^K 



y ^y O 'y ^, ^y O r^ ^x O 'y "y 'y ". 



^0000000000000 



in the statement, that the direction of the sides of an 

 angle are deviated toward the direction of the angle, and 

 may be illustrated by reference to Fig. 2. In this figure 

 the continuation of the left horizontal line seems to fall 

 below the right horizontal line, and the continuation of 

 the latter above the former ; in reality the two are con- 

 tinuous. Similarly, if the continuations of the oblique 

 lines be added, they will not seem continuous, but diver- 



gent slightly to one side or the other. If now we call the 

 direction of an angle the direction of the line that bisects 

 it, then the deviation is what would result from a draw- 

 ing up of the sides of the angle towards this central 

 bisecting line ; the left end of the left horizontal line 

 would be drawn up, and the right end of the right hori- 

 zontal line would be drawn down, and thus the two seem 

 discontinuous. The same would happen, though to a less 



SOME OPTICAL ILLUSIONS? 

 A STRIKING illusion, first described by Zullner 

 -^^- some thirty years ago, and usually called by his 

 name, appears in Fig. i. Of the four main lines each 



' A publication, "Die leuchtenden Nachtwolken," by O. Jesse, whichmay \ 

 be expected within the next months, will contain detailsregarding the entire i 

 present position of these questions. 



- Abstract of a paper on "A Study of ZoUner's Figures and other Re- 

 lated Illusions," by Joseph Jastrow, Ph.D. (with the assistance of Helen i 

 VVest), being a part of "Studies from the Laboratory of Experimental I 

 i'sychology of the University of Wisconsin."— .^/wcr/caw Journal of \ 

 Psychology, vol. iv. No. 3. 



NO. I J 99, VOL. 46] 



Fig. 3. 



degree, if either oblique line were omitted. There are 

 many other ways of illustrating this fact. Instead of 

 drawing the right line horizontal, we may incline its right 

 end downwards slightly, and then it will seem continuous 

 with the left horizontal line. We may apparently incline 

 both lines so that they would converge towards a point 

 between and below them, as in Fig. 3 and the like. Two 

 further points or corollaries should be noted : (i) that the 



