The methods employed in the experimental work have been described in the _British Journal of Dermatology_ for August, 1913. By using a very constant focus tube and averaging a large number of readings and correlating the results with those obtained in clinical practice, we were able to construct the curves indicating the colour developed by the pastille. In these curves, shown in Fig. 10, the ordinates are the Lovibond colour units, the abscissa the time during which the current was actually passing. When using an interrupter working at a constant speed, the actual time was taken, otherwise the number of current interruptions as measured by a dipper tachymeter was used. As was to be expected, the daylight and electric light curves were quite different. In each case the standard yellow glasses employed were kept constant throughout the curve, that for daylight being 15 units, that for electric light 13 units. When these were combined with blue and red glasses in varying units and fractions of a unit, they gave a colour range which matched the pastille exactly in the changes it undergoes from the unexposed condition to the 2 B Tint. The curves are plotted in accordance with Mr. Lovibond’s practice--that is, not as a direct representation of the standard glasses used, but as showing the colour sensation received by the eye. [Illustration: Fig. 10.] First, as to the daylight curve: In order to match the unexposed pastille we interpose between the pastille and the observer’s eye a yellow and a blue glass, plus a certain amount of neutral tint (composed of three equal colour units). Thus the colour sensation received is a yellow green, together with a certain amount of white light. As the pastille darkens under irradiation, both the green and the white light disappear, until at a point just below the 1/2 B dose, there is no other colour present but yellow. After this red glasses are required--_i.e._, the colour sensation is a yellow orange, which gradually deepens owing to an increase in the proportion of red. The yellow curve thus rises till just below the 1/2 B point, and then falls as the orange increases. Next, as to the electric light curve. The unexposed pastille has but a trace of green, which is soon lost. The orange begins much earlier than in daylight, and thus at Tint B has reached a higher point than in daylight. From this point the orange and yellow parts of both curves run practically parallel with one another up to 2 B. Beyond 2 B the readings become more difficult. I have not determined the point when no more colour develops, as it has no great practical value, though it might well be of interest from a physico-chemical standpoint.