EXPERIMENTS WITH THE DISPLACEMENT INTERFEROMETER. 89 



for illumination, as the source must be very intense. The results obtained are 

 surprising and are summarized in fig. 49 in which the ordinates 5 are the 

 chords of the radius of R = $o cm. subtending the aperture 26 of the coronal 

 disk or ring specified, so that s/2R = sind. The abscissas are the ordinal 

 numbers, z, of the successive partial exhaustions, all of them identical. Ar- 

 ranged in this way, all the apertures s, curiously enough, vary linearly with 

 the number of the exhaustions z, while the fog-particles are gradually increas- 

 ing in diameter d, exponentially. If the quantity 5 were laid off in terms of d, 

 the curves would be roughly hyperbolic, and less serviceable for exposition. 

 The three curves selected from many similar results refer, respectively, to the 

 aperture of the edge of the green disk, and to that of the inner and outer edge 

 of the first green ring, the latter value of 5 being about twice as large as the 

 s for the edge of the disk. During the earlier exhaustions, 2=1, 2,3, etc., the 

 coronas are very large and filmy, and a sharp value of s is out of the question. 

 Consequently observations in the regions A can not be expected to agree 

 closely. The graph for the outer ring, moreover, had to be taken from a 

 separate series of observations. 



The feature of these experiments is this, that the disk and the first ring are 

 alternately vividly colored (green) and alternately dark (yellowish, due to 

 the dull mercury line) ; i.e., when there is interference in the disk there is rein- 

 forcement in the ring, and vice versa. When the number of the exhaustion z 

 is high and the fog-particles therefore large, these alternations are crowded 

 so closely together that the successive partial exhaustions necessarily skip one 

 or more cases; but for the smaller particles, i.e., as far as exhaustion 0=15, 

 the results are invariably definite. The curves for the disk and inner edge are 

 joined by arrows, showing the successive position of vivid green color. 



There are two points of view from which these results may be interpreted. 

 We may lay off the intensity of green color in the disk and ring separately, 

 in which case maxima of the one will coincide with minima of the other as 

 shown at the bottom of the chart, fig. 49, i referring to the disk, 2 to the ring. 

 Again, the phenomenon may be regarded as a series of contracting rings, which 

 are first seen in a ring and successively in the disk. Lines a, b, and c have been 

 drawn from this point of view. The latter can not, however, be correct, since 

 it frequently happens that vivid green color is absent both from the ring and 

 from the disk, a condition not suggested by the lines a, b, c. On the other hand, 

 the lower intensity curves of the figure call for just this result at e, f, etc., 

 so that the former point of view is in accordance with observation. Apart 

 from the outer rings, which are too large for observation, we may even add 

 that the axial colors treated at the beginning of the paragraph probably 

 again reverse the alternations of the disk, or again approach the case of the 

 ring, though direct observations on this feature were not made. 



In fig. 50 the same attempt is made to represent the periodicity in disk and 

 inner ring, by representing the apertures of coronas 



s=zR sin 6; ^ = 30 cm. 



