204 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS 1.400 1.200 • 1.000 z 0 .800 .600 .400 .200 I I I I I I I I I r I I I 1200 1400 1600 1800 2000 2200 2400 WAVELENGTH (nm) Figure 4. NIR absorption spectra of skin showing different scores of dryness (intermediate scores not presented). 0.94 0.92 1.10 1.08 r = - 0.536 n = 310 pO,001 1.06 o • 1.04 (• 1,02 C) 1.00 r,... 0.98 0 0.96 0.90 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 GLOBAL SCORE Figure 5. Linear correlation between the NIR absorbance (measured at 1936-1100 nm) and the global score. skin dryness: above an overall score of 3, no further changes in conductance were observed. A closer examination of the clinical data shows that electrical conductance was related to the roughness of the skin: skin roughness no longer varied above an overall

NEAR-INFRARED SPECTROSCOPY 205 o 3.5- 3.0 2.5 2.0 1.5- 0.5• O.O• -0.5 • . = . ) ß I I I I I I I 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 GLOBAL' SCORE Figure 6. Correlation between roughness or scales scores and the global score (note that beyond a global score of 3, the roughness score no longer increases). score of 3 higher scores were mainly due to marked scaling. The lack of correlation between conductance and roughness was clearly due mainly to the poor quality of the electrical contact between the skin surface and the measurement electrode. Needless to say, this poses a problem for the precise interpretation of the results of electrical conductance measurements. Such difficulties have previously been pointed out (8), but our data provide a concrete illustration of the problem. The use of an integration sphere that collects all the radiation re-emitted probably explains why the NIR absorbance method was so sensitive. Above a clinical score of 3, scaling probably corresponds to a reorganization of the stratum corneum or the epider- mis in response to relatively deep dehydration, which the Infra-Alyser recorded linearly up to scores of about 4.5. With regard to the efficacy of the cosmetic preparations, product D, which does not contain moisturizing agents, was totally ineffective, regardless of the parameter con- sidered. The results for the other products varied accoMing to the measurement tech- niques used. The clinicians considered them to be equally effective, whereas product B was ranked between the group A/C/E and product D in the NIR method and on a level with product D in the electrical conductance method. This latter result was probably erroneous since, despite a degree of subjectivity and a lack of accuracy, the clinical scores could not have been so far from the real situation. It is more likely that the unexpected classification of product D by the impedance method was due to its lack of sensitivity above clinical scores of 3.