For this reason also, it is important that the final filter glass be flat, and this is one of the difficulties in making filters, that is, to obtain six pieces of glass that are sufficiently flat to be serviceable. To test the glass for flatness, place it on a flat black support such as a focusing cloth or piece of velvet, this being placed at an angle of 45 degrees on a table about six feet from a window; then, on looking down on it, the image of the cross bars of the window will be seen reflected and usually a double image, one from the front and a fainter one from the back of the glass. If the glass or the eye be now moved so that the images are reflected from all over the surface, we can easily see whether the two surfaces are parallel, as then the two images will keep the same distance apart, whereas if lenticular they will either diverge or converge. Pieces showing this defect should be rejected. If the images are not straight, the glass is also curved. It has been assumed that the glass is bought cut to the required size; if it is bought in a large sheet it can be examined in the same way and the flat parts marked with a piece of soap and subsequently cut out; but this presumes knowledge of how to cut glass with a diamond or wheel. This, like everything else, is extremely easy when you know the trick, but it will probably be better to purchase the glass ready cut and select the good pieces.

The size of the filter is easily determined, and a rough and ready way is to measure the diameter of the lens hood, not the glass, and allow one-quarter inch beyond this; this is assuming that the filter is to be used on the lens, as it will be seen later that there are other positions. A more correct method is to use a diagram such as is shown in Fig. 9, in which L represents the lens racked out from the plate PP, to its equivalent focus; then, by drawing a line from the corner of the diagonal of the plate through the lens, the size of the filter can be at once determined by actual measurement, as the distance between the line from the corner of the plate to the straight line A which is a prolongation of the optical axis of the lens will give at once half of the necessary width of the filter. Obviously the further the filter is from the lens the larger it must be if it is not to curtail the field or the light.

Fig. 9.

The gelatine solution is prepared as follows: Distilled water must be used and it is well to make up a generous quantity of the solution, say about 400 ccm, as it is decidedly better to waste some than to run short. As a 10 per cent solution is required, we weigh out 40 g of gelatine, place in the 500 ccm beaker, cover with distilled water and stir it well for two or three minutes, then pour the water off and repeat the washing, allowing the gelatine to soak for ten minutes the second time and fifteen minutes the third, stirring it occasionally. Then as much water as possible should be pressed out by means of a glass rod and the beaker placed in hot water at 550 C. (1300 F.), when the gelatine will gradually melt in the water that it has absorbed and enough more is added to make the bulk up to 400 ccm. The solution should now be filtered through a double thickness of well washed and wetted linen; an old handkerchief does well for this.

The dye solutions are preferably prepared in the form of stock solutions as follows: 1 g to 100 ccm for rose Bengal, patent blue, acid rhodamin and naphthol green; 2 g to 100 ccm for tartrazin and 0.5 g to 100 ccm for toluidin blue. These quantities should be placed in clean, well-dried bottles, 100 ccm of hot distilled water added, the bottles well shaken for half an hour and then allowed to stand so that the solutions may settle. The quantities of the dye (not solutions) needed for the average size of filter are very small and, therefore, the amounts for a square meter are given, from which it will be easy to calculate that for any given size:

### For The Red Filter

Rose Bengal 1.25 g.

Tartrazin 2.0 g.

### For The Green Filter

Naphthol green 0.4 g.

Patent blue 0.2 g.

Tartrazin 1.0 g.

### For The Blue Filter

Acid rhodamin 0.6 g.

Toluidin blue 1.8 g.

The usual quantity of gelatine solution employed is 700 ccm per square meter, which gives a dry filter thickness of about 0.4 mm; it is advisable to adhere always to a given volume of dyed gelatine for a given area.

We can now prepare the glass. This should be placed in a mixture prepared as follows. Make the following solution:

Potassium bichromate 50 g.

Water 250 ccm.

Then add gradually:

Sulphuric acid 25 ccm.

Stir well and add:

Water to make 1000 ccm.

Fill a developing tray with this solution and immerse the glasses, using a flat strip of wood to lift them, as the solution bites the skin. Really the best way to clean the glasses is to immerse one at a time and scrub the surface with an old tooth brush or a temporary mop, made by tying some old rags round the end of a short stick, then turn the glass over and scrub the other side, lift out with a stick and drop it into hot water, then start cleaning another glass and remove the first from the hot water, give it a rinse in distilled water and put in a rack to dry. Then put the second glass in the hot water and finish all in this way. It is possible to put six or more glasses into the cleaning liquid at one time, but in doing so there is a very great chance of scratching one or more of the glasses with a sharp edge or corner, and although cleaning them singly is more trouble and takes more time, it pays in the end.