THE CHARACTERISTIC CURVE The above proposition was experimentally demonstrated in the following way. Thickly coated slow plates were found best for experi ment. Upon one of these a series of exposures was made beginning with one second and doub ling the time at each successive exposure. The times of exposure with a constant source of illumination were thus in geometrical progres sion. After development the density resulting from each exposure was determined and its value with the corresponding exposure plotted as shown in the curve, figure 9, the geometric series of exposures being laid off equidistant on the horizontal axis. This curve is to be known as the "characteristic curve" of the plate. On examination the curve is found to consist of four distinct parts, corresponding respectively to the exposures 1 to 32 seconds, 32 to 1024 seconds, :and two other parts corresponding to greater exposures still. It will be noted that the second part is a straight line which signifies that the vertical distances are proportional to the hori zontal distances, but the vertical distances which represent the densities actually measured are in arithmetical series while the corresponding horizontal distances are made to represent num bers in geometrical series, these numbers denot ing the exposures. Thus the densities are seen to be proportional to the logarithms of the expo sures in this part of the curve, and therefore this part of the curve expresses the law of true tone gradations.
'For exposures 1 to 32 seconds the density increases approximately with the exposure and Would continue to do so for larger exposures liut for the fact that the effect of the light can only result on the unaltered silver present at any instant, and this of course is gradually decreasing so that a second of exposure to a given light agent at the end of a minute would add a far less increment to the density than it would at the end of 5 seconds when there would be so much more unaltered emulsion to act upon. As then at first the densities vary as
the exposures, the opacities will increase much more rapidly and the effect of printing will be one characteristic of an under-exposed negative, i.e., a too great contrast of light and shade.
Similarly it can be shown that the third part of the curve corresponds to the period of over exposure in a plate in which the growth of dens ity in the brightest lighted parts is so slow compared to the exposure as to result in a plate of too little contrast. The more strongly lighted parts have become nearly saturated with altered silver and do not increase much in density with time, while the less lighted parts are still increas ing rapidly with time. The result is obviously a flat negative.
The fourth branch of the curve corresponds to a period of excessive over-exposure resulting in a reversal of the image. This has been stud ied in detail by Nipher and others.
Referring again to the straight part of the characteristic curve it will be noticed that for the particular plate used, any exposure from 32 to 1024 seconas would have produced a technic ally correct negative. This range or "latitude". (see chapter V) is different for different plates and is taken into account in selecting a plate for a particular purpose. Further, to produce a prop erly graded negative the exposure must be within the limits assigned by the latitude. By no special method of development is it possible to change the density ratios on any particular plate. The actual densities at the different' points of a plate are of course dependent on the kind of developer and the time of development, but not the ratio of the densities at any two points.