NEAR-INFRARED SPECTROSCOPY 199 criteria, both before and after four weeks of treatment. On the basis of the results obtained, we attempted to answer the following questions: ß Is dry skin insufficiently hydrated, or is it simply rough? ß How good is the agreement between clinical evaluation, and results provided by the physical measurement? ß Are the results of the two physical methods redundant or complementary? MATERIALS AND METHODS MATERIALS Near-infrared spectrophotometer. Diffuse-reflection infrared spectroscopy has been considerably developed since the initial work by Norris in the 1970s (7). This rapid, noninvasive method can be used to obtain spectra of solid, opaque samples. In the agro-alimentary field, it is mainly used for in vitro analysis of water, lipids, and proteins (7). Apart from applications aimed at identifying raw materials and analyzing finished products, most uses in cosmetology (e.g., the assessment of hydration and the efficacy of moisturizing agents) involve in viva measurements. The apparatus we have adapted to spectral measurements in viva is an integrating sphere spectrophotometer (Infra-Alyser 500) coupled to a PC AT microcomputer. An external integrating sphere has been added (Figure 1) so as to acquire a full energy spectrum and to optimize the signal-to-noise ratio, without modifying the basic functions. The external integrating sphere is identical to the internal sphere and is connected to the optical elements of the Infra-Alyser 500 via two special optical fibers with a very low attenuation coefficient (0.5 db/m) throughout the study spectrum. The reference beam and the analytical beam are collected by the optical elements through a group of EXTERNAL MEASUREMENTS Internal integrating sphere "-'• • Sample beam Chopper • Reference beam :. %.. , /,,,, Optic •t fibers t• • ,• II • •'" ß II , •'/ External •,• , • Integrati Detectors • [ sphere Sample Figure 1. Modification scheme of the Infra-Alyser 500 to obtain in vivo measurement on all body sites.

200 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS adapted lenses, as well as under the internal sphere by a special optical system. The beam geometry (solid angles and incidence) in the external sphere is identical to that in the internal sphere. In preliminary studies, these modifications did not lead to a wavelength shift or change the absorption spectra of test materials. The fibers permit measurements to be made over the range 900-2500 nm with sufficient sensitivity, something that has not always been the case (6). The spectra were acquired from 1100 to 2500 nm in 4-nm steps and expressed as absorbance (log l/R), where R is the ratio between the energy reflected on the walls of the sphere and that retrodiffused by the sample. For homogeneous samples with little absorbance and no diffusion, the concentration of an analyte can be considered directly proportional to the absorbance (Ai) at a given wavelength (M) according to an equation based on Beer-Lambert's law: C = KAi = K log (1/Ri) In practice, particularly for measurements of the skin, the phenomena are far more complex. This law is no longer applicable because of interference due to analytes other than that under study and, above all, to diffusion due to the granulometry or surface state of the skin. The effects of this diffusion can be partly eliminated by calculation based on the difference in the absorbance at two wavelengths. The resulting equation is as follows: C = K (A2 - A1) = K log (R1/R2) (equation 1) Conductance measurements. Skin conductance was determined using a DermoDiag © apparatus (8), which operates at 10 MHz the result is related to the degree of hydration, the surface state (contact impedance between the skin and the electrode), and the thickness of the stratum corneum. EXPERIMENTAL STUDIES In vitro. Samples of stratum corneum isolated by heat-trypsin treatment were equili- brated at various degrees of relative humidity and analyzed by means of near-infrared spectroscopy. The difference in absorbance at 1936 and 1100 nm was recorded, and the water content was deduced from the sorption isotherms determined by differential calorimetry and weighing (9). Spectroscopic analysis of the skin in vitro was performed before and after removal of the epidermis by heat treatment, in order to study the influence of the most superficial layers of the skin. In vivo. Study population. Dry skin was characterized by studying the external surface of both legs (just below the middle of the calf) in a panel of 159 women with a mean age of 40 years (range, 18-67), who had given their informed consent. The efficacy of five cosmetic preparations was determined by studying five groups of about 20 women per group matched for age and the dryness score (2). The treatment was applied twice daily for four weeks to one leg, the other serving as a control. Products were different O/W (products B, C, D) and W/O (products A, E) formulations con- taining various proportions of moisturizing agents except product D. The measurements