is important that the camera should be mounted facing downwards, as in use, so 
that the weight of the pressure pad acts normally. The pressure pad in this 
case had a lapped surface of hard composition material; the pressure was about 
30 lb. For convenience in observation mirrors were used in the set-up as 
shown. Later the projection lens was replaced by a camera so as to photograph 
the Newton's rings. The light source for most of the work was a mercury lamp 
with a filter to select the line at 54-61 °A. 
At any particular location on the format, interference takes place between 
the rays reflected from the emulsion surface and from the adjacent upper surface 
of the register glass; the path difference between the two rays is twice the 
thickness of the air space between film and glass. Allowing for the phase 
change which takes place on reflection at the emulsion surface, destructive 
interference will occur at places where the thickness of the air film is an 
integral number of half-wavelengths and all points with the same number will 
appear to be connected by a dark band. At adjacent points where the air space 
is a quarter wavelength thicker or thinner, a bright band will occur. Thus a 
whole fringe system will arise, the fringes forming curves of constant level of 
the emulsion surface relative to register glass. The change in level between 
two fringes of the same kind, dark or bright, will be 0.27 micron. Fig.3 shows 
a typical record of a fringe system. 
There are two points to note about this technique. The first is that it 
does not reveal the order of any particular fringe, i.e, the actual number of 
wavelengths of path difference at any point. The second is that it does not 
indicate whether, in passing from one fringe to another, the air space is 
increasing or decreasing. The first limitation, although inconvenient, is not 
vital; the absolute separation is not of primary interest because, as already 
pointed out, the reseau image formed on the film during photography provides a 
calibrated co-ordinate system. The second limitation complicates, but does not 
prevent adequate interpretation of the fringe pattern. 
Fig.3 is strongly suggestive of a contour map, and one recognises the 
centres of closed fringe systems as peaks or troughs, where the height changes 
in the same way, irrespective of the direction of leaving the centre. Since 
also each fringe corresponds to a fixed level, their horizontal separation 
indicates the slope of the surface and at a point of zero slope the fringes 
must broaden out. This occurs not only at peaks and troughs but also at points 
such as S in Fig.3 where a fringe splits into two branches. The interpretation 
of the diagram may be illustrated by considering the set of features A, B, C and 
S. If one assumes that A represents a peak, then B must represent a trough, 
because the slope between them is uni-directional (no broad fringes). Leaving 
B in the direction of C, the slope is positive and again uni-directional, leading 
to a peak at C. S, which is above B, but below A and C, is a saddle-point, and 
the "ground" falls away again on the side opposite to B, as may be seen by the 
nature of the contours surrounding A and C on this side. By such considerations 
it is possible to establish the relations between all the main features of the 
interferogram. It is not possible to say with certainty which point, A or B, 
corresponds to the greater separation between film and register glass. The use 
of polychromatic light enables this question to be answered, because in this case 
the distinctness of the fringes decreases as the thickness of the air space 
increases. However, this same fact makes it more difficult to use the inter- 
ferogram where the air space exceeds a few microns, and since the main interest