As was shown originally by Abney, an intermittent exposure does not give the same effect as a continuous exposure for the same total time. The use of a rapidly rotating sector is, there fore, undesirable, and instruments giving continuous exposures should be used.
Another important factor is the failure of the reciprocity law, E=It. As was first shown by Abney (Proc. Royal Soc., 54, PP. 143-147, 1894), the effect of an exposure is not independent of the intensity. Recent work has shown that the curve showing the relation between density and intensity, It being constant, has a maximum, this point being referred to as the "optimum" intensity. (See fig. 12.) The shape of the curve as well as its absolute values vary with the type of the emulsion and also with the wave-length of the light that is used. The work of Jones, Hall and Webb has shown the great importance of the reciprocity law in sensitometric work.
It is therefore desirable that exposures in sensitometric instru ments should not only be continuous but of the intensity level likely to be used in practice.
The original photometer with two light sources used by Hurter and Driffield has been superseded by various optical instruments for measuring the densities, and these in turn appear likely to be superseded by physical photometers, such as those employing photo-electric cells. The density obtained depends on whether it is read in contact with a diffusing screen, such as opal glass, or whether it is observed by parallel light from a collimator. With the latter, as was pointed out by Abney, Chapman Jones, and Callier, much of the incident light is scattered, so that the density has a higher value than when measured by completely density of parallel light diffused light. The ratio between density of diffused light was termed Q by Callier ; it varied in amount from approximately unity to values as high as 1.6. The value of Q for average nega tives is about Recently graduated wedges of neutral tint have been used very largely in sensitometry both for exposing and for the measure ment of densities.
A point of considerable importance in speed determination is that of the fog inherent in the emulsion and developed without light action. Hurter and Driffield assumed that this was constant
throughout all densities, or, in other words, through all exposure periods; but such is not the case, there being actually more fog in the underexposed parts than in the more exposed parts, because there is less silver bromide available and also more bromide set free here than in the less exposed parts.
Returning to fig. 11, it will be observed that the tangent of the angle at which the straight line meets the exposure axis is marked a. This angle is of great importance in photography, since it defines the contrast of the image. The tangent of the angle, that is, the slope of the straight line, was termed y by Hurter and Driffield and was called the "development" factor, since its value depends upon the time of development. During development, the value of y increases, the increase tending to reach a limiting term, gamma infinity (7. ), which measures the extreme contrast of which a plate is capable. Thus, if we develop for different times plates which have been given a series of exposures increasing in geometrical proportion, measure the densities, and plot the resulting curves, we shall get the result shown in fig. 13, in which are shown curves corresponding to development times of 1 min., 2 min., 3 min., 5 min., and 6 min., while Tx corresponds to the maximum development that can be given before the contrast diminishes owing to fog.
If now we plot 7 or density as a function of time, we get a curve of the type shown in fig. 14. It will be seen that the value of y increases rapidly at first and then more slowly, finally reaching a limit . This is an exponential curve, the rate of development being proportional to the amount of exposed halide remaining to be developed, so that the reaction is identical in form with that of a chemical reaction of the first order. The velocity of the reaction of development depends therefore upon the value of 'y co which is a property chiefly of the sensitive material although also, to a minor extent, of the developing solution, and of K, the velocity constant of development, which is dependent upon the composition and temperature of the developer.