( 349 ) 
Such a three-fold plate is made nearly as easily by the experimenter 
as a double one, whereas for measurements its use allows of a 
somewhat higher accuracy by reason of the symmetry of the field 
Using either these double or three-fold mica plates one immediately 
gets the total displacement of the bands, if a quartz plate be super¬ 
posed and the combination placed between crossed Nicols. An image 
of the horizontal lines of separation is projected by means of an 
achromatic lens upon the slit of a spectroscope. 
In fig. 1 of the Plate a photograph taken with a three-told mica- 
plate is given. The relative retardation produced by the mica is a 
quarter of a wave-length, if the distance of two bands in the outer 
fields is just halved by one in the middle field. 
A comparison spectrum from * = 6561 (H) to a _ 407 ( g) 
taken simultaneously. How largely the phase-difference in the mica 
depends upon the wavelength is very clearly shown. 
4. Fresnel’s rhomb gives, as is well known, circularly polarized 
light after two total reflexions of light polarized at an azimuth of 4o . 
This arrangement, if perfect, has the advantage of giving circularly 
polarized light for all colours of the whole visible spectrum. 1 
examined as to how far the rhomb at my disposition satisfied the 
requirements imposed by the object I had in view. 
The chief purpose of my investigation requires a rather close 
approach to the ideal condition. For this reason and also because, as 
far as I know, Fresnel’s beautiful invention was never investigated 
by the method of interference bands, I may be allowed to give a 
more extensive treatment of the results than I otherwise would have 
contemplated. . ,. , 
In order to obtain also with Fresnel’s rhomb two fields in which 
the bands are shifted in opposite directions, the rhomb is to be 
combined with a double plate. I used a double-plate made of a p ate 
of quartz, 1.7 m.m. thick, and cut parallel to the axis This plate 
is cut in two halves by a line making angles of 45 with the prin- 
cipal section; one half is then rotated round an axis perpendicular 
to the line of separation, through 180°. 
Fig. 2 was obtained by means of this double-plate and a rhomb, 
many years old and belonging to the collection of the Amsterdam 
laboratory. The deviation from an ideal apparatus is very apparent 
on inspection of the photograph. Measurement shows that the bands are 
4 ) cf. CoRHU l.C. 
Proceedings Royal Acad. Amsterdam. Vol. XII. 
24 
