6 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS an initial reading (pt. A on the abscissa). The control site (o) was left untreated. After two hours the velocity through the treated site had reduced significantly relative to the control. Immediately after this reading (B) both sites were tape stripped to remove the outer stratum corneum layers and the sonic velocity was measured again (C). The untreated control site velocity decreased to a level equivalent with the treated site, while on the treated site the velocity remained constant after stripping (the criterion of c• .05 was used to determine if statistically significant differences existed). In another experiment, application of glycerin/petrolatum products or pure water to freshly stripped skin was found not to produce a statistically significant alteration in the sonic propagation velocity compared to untreated stripped skin. These stripping studies are interpreted as indicating that the observed differences in sonic velocity which occur ajQer acute product treatment are the result of changes in the near surface layers of the stratum corneum induced by the treatment. Both product treatment and removal of the uppermost cell layers of the skin produce a sonic velocity decrease of the same magnitude (15%). Once these outer layers are removed, a constant lower velocity is observed which is not further decreased by continued cell removal (stripping) or by application of either products or water to the stripped site. 3. SONIC VELOCITY AND THE VISCOELASTIC PROPERTIES OF SKIN The physical properties of skin have been widely studied (18,19,21-29). While all of these workers agree that human skin is a viscoelastic material, an examination of this literature reveals different descriptions of its viscoelastic behavior. Some researchers (19,21,24) report that with continued stretching (increasing strain), skin goes through four distinct physical behavior patterns as follows: Phase 1: Removal of "slack" Phase 2: Purely elastic upon release, the skin immediately returns to its original form. Phase 3: Viscoelastic upon release, the skin slowly returns to its original form (hysteresis). Phase 4: Purely viscous the skin is permanently damaged and does not return to its original form (as manifested by the striae gravidarum resulting from pregnancy or obesity--ref. 30). The sonic velocity measurements in this work were taken while the skin was in the elastic region (phase 2 of the model), i.e., just after the slack was removed. In phase 2, the in vivo stress-strain characteristics of skin are linear (18). Indeed linearity was evidenced by the fact that slight further manual stretching of the skin did not change the observed sonic velocity which is related to the elastic modulus (vide infra.). Others (23,25) describe the skin's viscoelastic behavior as an interaction between predominantly elastic components such as collagen, elastin, and keratin in the outer layers, and the purely viscous interstitial fluids. It is this keratinous outer layer that skin care products are thought to affect and the skin stripping studies described above confirm that it is this very layer which affects the observed changes in sonic velocity. These models, together with the tape stripping results, suggest that the physical property measured in sonic transmission studies may be made up of two components: transmission through the viscous, inner layers of the skin (lower stratum corneum, epidermis, and perhaps the dermis) and transmission across or through the dehydrated,

SKIN CONDITION MEASURED BY SONIC VELOCITY 7 elastic outermost skin layers. Resting the transducer probes directly on the exposed under layers after stripping results in a decreased observed velocity, whereas shorter pulse transit times are recorded if the probes are on the outer layers. It is highly unlikely that removal of the stratum corneum would immediately lead to a decrease in the propagation velocity of sound through the lower skin layers (i.e., the transmission velocity through these layers would remain the same whether the stratum corneum was present or not). Therefore, the fastest propagating wavefronts, those which reach the receiving transducer first after an input pulse and trigger the timing circuit, must travel longitudinally (the geometry of the transducers on the skin insures that only longitudinal waves are generated and detected) for at least part of their path through the skin structure in which propagation is the fastest, the very outer layer of the stratum corneum. Because the stripping experiments indicate that this layer is extremely thin (being removed in five tape strippings), a one-dimensional, longitudinalphysical model for the sonic propagation process in the skin will be assumed. Applying this simple model, Young's modulus of elasticity, E (slope of the stress-strain curve, definitionally the stress required to produce unit strain), is directly related to the squared sonic transmission velocity, c2: E = d c 2 ß 10 3 where E is in newtons/meter 2, d is the density of skin in g/ml, and c is in m/sec (31). An experimentally determined stress-strain curve for skin from the data of Grahame (18) is set out in Figure 4. Referencing this figure provides a qualitative interpretation of 5 To • Stress (Newton/m X 10 2) 2 1 0 2 4 6 8 Strain (m X 10 -•) Figure 4. Experimental stress--strain curve for in vivo skin from the data of Grahame (ref. 18). sonic transmission. Before sonic velocity measurements, the skin is extended by application of a force, To, into the linear portion of the stress-strain curve. A time-dependent compression and rarification of the tissues involved in longitudinal wave propagation may be represented by the alteration AV, or volume change along the abscissa as pressure alterations AT are applied by the transmitting crystal. As long as one remains in the linear elastic region of the curve, the tension of the skin may be