ZOOLOGY AND BOTANY, MICROSCOPY, ETC. 
239 
In order to dispense with the use of homogeneous light, the author 
prefers to the above method one, no less exact, depending on the obser- 
vation of the bands of Mtiller, which appear in the spectrum when the 
interference colours presented by doubly refracting plates in polarized 
light are analysed by the spectroscope. In this method, the wedge of 
selenite is placed, as before, between the crossed nicols with its axes 
orientated at 45° to the principal sections of the nicols. Above the 
analyser, which is directly above the objective, is placed a direct-vision 
spectroscope provided with a special apparatus for determining the wave- 
length in the different parts of the spectrum. The lateral displacement 
of the wedge in the azimuth 45° is effected by a micrometer screw, and 
can be measured to 0*05 mm. nearly. In these conditions the dark 
bands in the spectrum are observed and their position accurately deter- 
mined by means of the scale of the micro-spectroscope. The position 
and number of these bands varies with the nature and thickness of the 
double-refracting plate. 
With daylight instead of homogeneous light the interference colour 
is composed of all the colours for which the difference of path 0 of the 
two interfering rays is equal to 
* = (2l+l)j, 
while in this tint there are wanting all the colours for which 
The parts of the spectrum corresponding to these absent colours are 
therefore dark, and form the bands of Muller. 
Denoting by & the constant difference N e — N 0 , we have then for the 
first dark band corresponding to the wave-length A 
8 D, = A, 
which gives the thickness of the wedge at the point considered, and 
similarly for the second, third .... bands. 
Having determined the thickness of the wedge at each point of the 
micrometer division, the tints of the Newton rings which correspond to 
these thicknesses are given by 
2 e = D (N 0 — N«), 
i. e. for the selenite 2 e = 0*0091 D, where e is the thickness of the layer 
of air corresponding to the ring considered. 
The exact determination of the position of the bands of Muller in 
the spectrum offers some difficulty owing to their considerable width 
when thin layers are observed. For this reason 6 to 10 observations 
must be made and the mean taken. The bands are so much narrower, 
the thicker the section. This leads to a slight modification of the method 
of calibration described. This modification consists in superposing the 
f wedge on a doubly refracting plate of the same nature as that of the 
wedge and of such a thickness that it takes the white colour between 
crossed nicols and gives a certain number (4 to 6) of the bands of Muller 
throughout the length of the spectrum. 
