SENSORY AND PHYSICOCHEMICAL PROPERTIES OF EMOLLIENTS 181 factors, explaining about 87 .9% of the total variance in y. For the individual descriptors, percent variance of experimental values accounted for by PLS Factor 1 and Factor 2 are shown in Table II. The correlation coefficients of sensory descriptors and instrumental measurements with the two PLS factors are presented graphically (Figure 3). Table II and Figure 3 show the PLS correlation to be bi-dimensional. This was confirmed by Osten's F-test, which determined that Factor 1 and Factor 2 were valid predictions (p 0.001). Table II shows that all attributes were well predicted by PLS regression on the instru- mental parameters. Figure 3 shows that gloss, residue, and oiliness were positively correlated to instrumental surface tension and negatively correlated to spreadability at one-half minute and one minute. Instrumental viscosity was positively correlated with difficulty of spreading and stickiness and negatively correlated with softness and slip- periness. These results suggest that sensory attributes related to mechanical instances of appli- cation or evaluation, such as difficulty of spreading, stickiness, and slipperiness, are related to physicochemical properties such as viscosity, that is, emollient resistance to flow and the forces needed to achieve it. Besides, sensory properties related to the film that the product leaves on the skin (gloss, oiliness, and residue) are related to physico- chemical properties such as spreadability at one-half minute and one minute, and surface tension, properties that depend on the capacity of the emollient to develop films. CONCLUSIONS In considering their physicochemical characteristics, the studied emollients were sorted into three groups, in which the two silicones are distinctly separated from the rest. The sensory characteristics enabled the discrimination of four groups of emollients, where, besides the two silicones, IMP also differentiates. 0.8 sof 811 0.4 0.2 glo oH N 0.4 '88 I -0.8 -0.8 -DA -0.2 0.2 0.8 0.8 -0.2 -DA da et1 -0.1 -1 PLS FaclDr 1 Figure 3. PLS2 loadings for the seven sensory variables (y-block) and four instrumental variables (x-block) analyzed on the eight samples. ds: difficulty of spreading. glo: gloss. res: residue. sti: stickiness. sli: slipperiness. sof: softness. oil: oiliness. SO.S: spreadability at one-half minute. Sl: spreadability at one minute. ST: surface tension. VIS: viscosity.

182 JOURNAL OF COSMETIC SCIENCE Table II Percent Variance of Experimental Values Accounted for by First Two Partial Least Squares Factors for Sensory Descriptors Factor 1 Factor 2 Difficulty of spreading 5.6 88.6 Gloss 81.8 5.9 Residue 90.8 1.4 Stickiness 0.1 95.6 Slipperiness 12.3 78.1 Softness 0.1 62.6 Oiliness 84.0 0.2 PLS revealed that emollient sensory attributes could be well predicted by instrumental measurements. Gloss, residue, and oiliness were correlated with surface tension and spreadability at one-half minute and one minute, whereas instrumental viscosity was correlated with difficulty of spreading, stickiness, softness, and slipperiness. REFERENCES (1) J. B. Wilkinson and R. J. Moore. Cosmetologfa de Harry (Ed. Dfaz Santos S. A., Madrid, 1990), pp. 73. (2) B. Salka, Choosing emollients: Four factors will help you decide, Cosmet. Toiletr., 112, 101-112 (1997). (3) Cosmetic Bench Reference, Cosmetics & Toiletries (Allured Publishing Corporation, 1998). (4) R. L. Goldember and C. De la Rosa, Correlation of skin feel of emollients to their chemical structure, J. Soc. Cosmet. Chern., 22, 635-654 (1971). (5) H. M. Brand and E. E. Brand-Garnys, Practical application of quantitative emolliency, Cosrnet. Toiletr., 107, 93-99 (1992). (6) V. Kamershwarl and N. Mistry, Propriedade sensoriais dos emolientes, Cosrnet. Toiletr., 13, 52-57 (2001). (7) L.B. Aust, L. P. Oddo, J. E. Wild, 0. H. Mills, and]. S. Deupree, The descriptive analysis of skin care products by a trained panel of judges, J. Soc. Cosrnet. Chern. 38, 443-449 (1987). (8) G. V. Civille and A. A. Dus, Evaluating tactile properties of skincare products: A descriptive analysis technique, Cosrnet. Toiletr., 106, 83-88 (1991). (9) C. E. Clum, Oils as moisturizers and emollients, Cosrnet. Toiletr., 93, 43-44 (1978). (10) E. Primo, Qufrnica Agricola Ill: Alirnentos. (Ed. Alambra, Madrid, 1979), pp. 161. (11) ASTM E 1490-92 (Reproved 1997), Standard Practice for Descriptive Skin/eel Analysis of Creams and Lotions (American Society for Testing and Materials, 1997). (12) ISO 8589, Sensory analysis: General guidance for the design of test rooms (International Standards Organization (ISO), Switzerland, 1988). (13) S. Glasstone, Tratado de Qufrnica Ffsica (Ed. Aguilar, Madrid, 1964), pp. 443. (14) E. D. Schukin, A. V. Pertsov, and L. A. Amelina, Qufmica Coloidal (Ed. Mir, Moscow, 1988), pp. 47. (15) M. Martens and H. Martens. "Partial Least Squares Regression," in Statistical Procedures in Food Research, J. R-. Piggot, Ed. (Elsevier Applied Science, London, 1986), pp. 293-359. (16) G. Hough, A. N. Califaro, N. C. Bertola, A. E. Bevilacqua, E. Martinez, M. J. Vega, and N. E. Zaritzky, Partial least squares correlations between sensory and instrumental measurements of flavor and texture for Reggianito grating cheese, Food Qua!. Prefer., 7, 47-53 (1996). (17) D. W. Osten. Selection of optimal regression models via cross-validation,]. Chem., 2, 39 (1988). (18) Genstat 5, Release 3 Reference Manual (Genstat 5 Committee, Rothamstead Experimental Station, Oxford, UK, 1993 ).