JOURNAL OF COSMETIC SCIENCE 458 lower the relative humidity or the moisture content of hair, the lower the strain level at which beta-delta failure occurs. Beta-delta failure was observed by Negri et al. (39) on wool fi ber, and they noted disruption of the cuticle–cuticle CMC along the upper beta layer in TEM sections. With this type of fracture, the delta layer and the lower beta layer are both retained on the underside of the “old” outermost cuticle cell, leaving 18-MEA as the “new” hair surface once the “old” outermost cuticle cell is abraded away. This type of fragmentation has been described in detail by Feughelman and Willis (33), who proposed that the failure of adhesion between overlapping scales involves 18-MEA, and that be- cause of its branching, it provides mobility and a reduction in adhesion between scales, leading to beta-delta failure. Therefore, the degradation of 18-MEA between scales as occurs in chemical (81) or photochemical bleaching (81) or even permanent waving treat- ments (82) leads to further weakening of this structure and more rapid beta-delta failure, leading to faster cuticle fragmentation and cuticle loss. WET STATE VS DRY STATE FAILURE Deformations such as stretching (4,83) (including extension cycling (32), bending, or twisting in the wet state) are very different from deformations in the dry state. This is because failure in the wet state generally involves fractures or breaking bonds in hydro- philic layers, e.g., the endocuticle or the central contact zone of the CMC, whereas failure in the dry state generally involves fractures in or between hydrophobic layers, e.g., beta- delta failure (4,33). Failure in the wet state generally involves hydrophilic regions be- cause when a layer or region is completely swollen, less mechanical stress is required to distort that layer and to produce a fracture. On the other hand, the lower the relative humidity or swelling condition of a hydrophilic layer, the more mechanical stress and strain required to distort hydrophilic vs hydrophobic layers, and therefore fractures are generally produced in hydrophobic layers. Extension of undamaged hair-to-break generally produces smooth fractures (32). How- ever, as the hair becomes more damaged or as the relative humidity is decreased, and es- pecially at lower humidities, more step fractures are produced (32), and step fractures involve extensive fracturing in the cortex–cortex CMC, most likely in the beta layers. Kamath and Weigmann (83) have shown, for human hair at low moisture content, that crack initiation occurs most often in the cortex, whereas at high moisture content, frac- tures almost always initiate at or near the surface of the fi ber because of the high pressure of the swollen cortex against the cuticle. Step fractures involve the axial propagation of cracks either through the cortex–cortex CMC or the medulla (83) and therefore occur more frequently in the dry state than when hair fi bers are wet (83). Kamath and Weigmann (83) also concluded that the CMC seems to “play an important role in stress transfer and axial splitting” of human hair fi bers. The abrasion resistance of human hair is decreased by most chemical treatments and photo-oxidation, as shown by the “protein loss” test of Sandhu and Robbins (84) or by the release of labile and eluted proteins as described by Inoue et al. (85). These tests are both wet-state methods. The inside of cuticle and cortical cells is degraded by alkaline peroxide, thereby weakening the cuticle and cortex cells internally, and the cuticle–cuticle CMC is degraded, weakening the cellular cohesion or the resistance of scales to break apart. Cuticle fragmentation in the dry state is caused primarily by the rupture between cuticle cells

CELL MEMBRANE COMPLEX 459 through beta-delta failure (32,33) and the resultant chipping of cuticle from the hair via abrasive actions. Cuticle loss in the wet state is primarily caused by the rupturing of cuticle cells internally and is greater in chemically damaged hair such as alkaline peroxide-treated or permanent-waved hair than in chemically untreated hair (84). Fatigue testing, a method developed at TRI-Princeton by Ruetsch, Kamath and, Weigmann (involving attaching a weight to a hair and dropping the weight multiple times to con- tinuously shock or jar the fi ber), shows that alkaline peroxide treatment of human hair fi bers when fatigued produces numerous scale-edge fractures with scale-edge chipping. Ruetsch (86) fatigue tested peroxide-treated hair, followed by extension, and found ex- tensive fracturing in the CMC between the scales due to a weakened cuticle–cuticle CMC. This effect is most likely due to peroxide attack on thioester linkages that disrupts the beta layers of the cuticle–cuticle CMC. Takahashi et al. (87) have provided evidence that wet cuticle wear in Asian hair is due more to CMC failure (possibly involving the central hydrophilic “contact zone” of the delta layer) rather than failure inside cuticle cells as in Caucasian hair (most likely endo- cuticular failure). Takahashi et al. (87) showed that wet cuticle wear on Asian hair occurs at a faster rate than on Caucasian hair because of differences in elasticity of the different layers inside cuticle scales. These scientists showed that the scales of Asian hair are re- moved faster by wet sonication after extension to 35% or by bleaching the hair followed by shampooing and combing the hair over a large number of cycles. In the latter case, after 90 times for four cycles, fewer scales were found on Asian hair relative to Caucasian hair (3.2 vs 1.3 scales, respectively). On further examination of the hair using an atomic force microscopic probe, these scien- tists found a greater difference in elasticity as a function of depth for the Caucasian hair (1.41 vs 1.26), and they concluded that the scales of Asian hair are more uniform inside and that therefore the intracellular matter of Asian hair cuticle is more resistant to frac- turing. Therefore, they concluded that the scales of Asian hair are removed more by frac- turing in the cuticle–cuticle CMC (even in the wet state), while the scales of Caucasian hair fractures inside the scales are most likely in the swollen endocuticle. It is interesting to note here that Nakamura et al. (8) by staining reactions has concluded that the com- position of the proteins of the delta layer of the cuticle–cuticle CMC is very much like that of the very hydrophilic endocuticle. CMC LIPIDS DEGRADED BY VISIBLE LIGHT MORE THAN BY UV Hoting and Zimmerman (88) have studied radiation damage as a function of wavelength and have shown that the CMC lipids of hair fi bers are degraded most by visible light, but also by UV-A and UV-B light, helping to explain the weakened CMC (of cuticle and cortex) and the multiple-step fractures that result from the axial propagation of cracks through the cortex–cortex CMC in sunlight-oxidized hair. Obvious weak links to photochemical attack on lipid structures are the tertiary hydrogen atoms of 18-methyl eicosanoic acid (89) and cholesterol and cholesterol sulfate. The allylic hydrogen atoms of oleic and palmitoleic acids and of cholesterol and cholesterol sulfate in the cortex–cortex CMC are also vulnerable to photo-oxidative reactions. Long-term irradiation does not provide for clean breakage between structural compo- nents of human hair, as was observed for peroxide-oxidized hair, but leads to cross-linking