MEASUREMENT OF SKIN COLOUR IN VIVO 727 as pigs, as shown in Fig. 1. This instrument is most suitable for measuring changes in pigmentation it can be used to quantify erythema, but lacks sensitivity. When erythema is present, with or without changes in pigmen- tation, it is necessary to make measurements at two wavelengths, one of which should be sensitive to changes in the degree of erythema, and the other to pigmentation. Estimation of the degree of erythema present, is made by comparing the reflectance between erythematous and control sites, using light selected by the green filter which is heavily absorbed by blood. The extent of pigmentation is measured using red light where changes in blood content of the skin have little influence on reflectance. A number of workers have made measurements of this sort using commercially available instruments in order to study pigmentation in the presence of erythema, or to quantify erythema in patients with different basic skin colours. Daniels and Imbrie (1) in studies on erythema produced by sun- light used a Photovolt Meter Model t310 which is fitted with wide band glass colour filters similar to the EEL reflectometer a similar instrument was used by Kahn and Wilcox (2) for comparing in vivo sunscreen testing methods, and by Lerner and McGuire (3) in endocrinological studies on pigmentation. Studies have also been made using instruments constructed in workers' own laboratories, but based on similar principles to those described above. Tronnier (4) and Runge, McHugh and Johnson (5) used wide band red and green glass filters to isolate their measuring wavelengths, while Breit and Kligman (13) used narrow-band interference filters with peak transmissions at 542 nm and 13131 nm for measuring erythema and melanin respectively. The advantage of using interference filters, rather than glass filters, is that a much narrower band width is possible for a given trans- mission factor. This is useful when measuring the degree of erythema, but is of less importance with respect to melanin. Spectrophotometry To obtain the spectral reflectance curve of a coloured surface it is necessary to measure photoelectrically the amount of light reflected at each wavelength through the visible spectrum, and to compare this with the light reflected from a standard surface, at each wavelength. Though it is possible to get only a very approximate idea of the colour of an ob iect from the reflectance curve, nevertheless, the spectrophotometer is the funda- mental instrument for colour measurement since it gives complete in- formation about the spectral composition of light reflected from the sample.

728 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS If the energy distribution of the illuminant is known, then the C.I.E. tristimulus specification can be derived, that is, the position of the sample can be plotted on the chromaticity diagram which gives a better indication of the colour appearance of the sample. Most commercially available spectrophotometers are fitted with an integrating sphere for making measurements of diffuse reflectance from surfaces, but it is difficult to position some portions of the body against the sample aperture of most instruments. However, extensive measurements of human skin pigmentation were carried out by Edwards and Duntley (7) using the Hardy recording spectrophotometer although this is a complex instrument and not suitable for general laboratory or field work. They obtained curves over the visible spectrum from 400-700 rim, which were sufficiently discriminating to detect the presence of reduced haemoglobin and oxyhaemoglobin as well as melanin. Similar curves were obtained in endocrinological studies by Lerner and McGuire (3) using a Cary model 14 spectrophotometer, and over a wider spectral range (from 235-700 nm) by Jacquez, Kuppenheim, Dimitroff, McKeehah and Huss (8) using a Beckman DR recording spectrophotometer. If visual properties of skin are being studied it is unnecessary to go beyond the limits of the visual spectrum (from 400-700 nm) when plotting reflection curves, but the wider range used by Jacquez yielded information about the nature of the pigments present in skin. For some investigations a full spectrophotometric curve is not required and useful information can be obtained by abridged spectrophotometry in which a small number of broad bands in the spectrum are isolated with suitable colour filters and the reflection of the sample and the standard white surface are compared at each band. Measurements of this type are useful when the general character of the spectral reflection is required, as when comparing racial differences in pigmentation, although the discrimin- ation of sharp changes in absorption is lost. The main advantages of abridged spectrophotometers are simplicity of operation and portability, and various instruments of this type have been used by a number of workers. Weiner (9), Barnicot (10), and Harrison and Owen (11) used a simple EEL reflectance spectrophotometer (Evans Electroselenium Ltd., Harlow, Essex) for anthropological studies of different races. This is a typical instru- ment of its kind and gives a satisfactory reflectance curve for skin, con- sidering that measurements are made with wide band filters at only nine points through the spectrum brightness, dominant wavelength and purity can also be obtained. The instrument consists of a reflectance head, con-