JOURNAL OF COSMETIC SCIENCE 444 bound fatty acids ranged from C16 through C20, with 6% uncharacterized. For human hair, Wertz and Downing (34) found 41% 18-MEA, 18%, palmitic acid, 7% stearic acid, and 4% oleic acid, and the remaining small percentages of fatty acids ranged from C16 through C20, with 9% uncharacterized. Negri et al. (42) found 72% 18-MEA, 8% pal- mitic acid, 6% stearic acid, and 5% oleic acid in wool fi ber. The variation in these data from different laboratories is quite large. Part of the variance has been suggested to be related to fi ber diameter, which determines the number of layers of covalently bound fatty acids in the fi bers however, certainly part of the variance is due to experimental error. The bottom line is that somewhere in the vicinity of 50% ± at least 10% of the covalently bound fatty acids in most keratin fi bers is 18-MEA (attached mainly on the top surface of cuticle cells). Furthermore, hair fi bers from sheep, humans, dogs, pigs, and cattle, and likely most keratin fi bers, contain the remaining nearly one half of the covalently bonded fatty acids, primarily as palmitic, stearic, and oleic acids attached mainly on the bottom of cuticle cells. In 1990, Kalkbrenner et al. (43) demonstrated with isolated cuticle cells that 18-MEA is essentially all in the cuticle. Since 18-MEA represents more than 40% of the total cova- lently bound fatty acids in human hair and about 50% in wool fi ber, and since 18-MEA is confi ned to the upper beta layer of the cuticle (23,24) while most (essentially an amount equal to the 18-MEA) of the other covalently bound fatty acids are confi ned to the lower beta layer, then most of the covalently bound fatty acids in wool and hair fi ber must be in the cuticle–cuticle CMC, with some in the cuticle–cortex CMC (to be described later) and virtually none in the cortex–cortex CMC. Therefore, if most of the covalently bound fatty acids are in the cuticle–cuticle CMC, then most of the lipids of the cortex–cortex CMC must be bound to the membranes on one side and to the delta layer on the other side by non-covalent bonds. The fact that most of the remaining lipids can be removed by solvent extraction confi rms that this is the case, as will be shown below and in the sec- tion entitled “Lipids in the CMC.” Leeder et al. (44) fi rst found that there are virtually no phospholipids in keratin fi bers. This fact was confi rmed by Schwan and Zahn (45) and by Rivett (46), casting doubt on whether lipid bilayers could be involved in the cell membranes of keratin fi bers (44). However, Wertz et al. (47) demonstrated that liposomes (lipid bilayers and a presumed precursor to the formation of lipid bilayers in the CMC of keratin fi bers) can form in the absence of phospholipids if an acid species such as cholesterol sulfate is present with other lipids. Table I Covalently Bound Fatty Acids in Wool and Human Hair Fiber Data for wool fi ber Data for human hair Fatty acid [13] [43] [44] [36] [22] Averages [35] 16:0 8 11 8 17 20 12.8 18 18:0 8 12 6 10 25 12.2 7 18:1 7 8 5 5 0 5 4 MEA 51 43 72 48 55 53.8 41 Others 26 26 9 20 trace 16.4 30 Data are expressed as percentages.

CELL MEMBRANE COMPLEX 445 Furthermore, evidence has been provided confi rming the existence of cholesterol sulfate in human hair by Wertz and Downing (34) and by Korner et al. in wool fi ber (48). The work of Korner et al. (48) builds upon the fi ndings of Wertz et al. on liposome forma- tion and lipids from stratum corneum (47), where Korner et al. (48) demonstrated that cell membrane lipids extracted from human hair and wool fi ber with chloroform/metha- nol/aqueous potassium chloride can form liposomes, providing evidence for a bilayer structure of the internal lipids of the beta layers of the cortical CMC in wool fi ber and in human hair (see Figure 3). Such extracts must come primarily from the cortex–cortex CMC because covalently bound MEA and the other covalently bound lipids of the cuticle CMC are not removed with this solvent system. Therefore, if the beta layers of the cuticle cells are primarily covalently bound fatty acids with some free lipids (see Figure 2) and the beta layers of cortical cells consist primarily of lipid bilayers (Figure 3), then it is highly likely that the proteins that these very differ- ent lipid layers are attached to, which are the cell membrane proteins, and the delta layer proteins of the cuticle cells and cortical cells are also different (see the section entitled “Proteins of the CMC” and the next part of this section). Additional differences between cuticle–cuticle, cuticle–cortex, and cortex–cortex CMC. As early as 1975, Nakamura et al. (8) provided evidence by staining reactions that the disulfi de con- tent in the delta layer in cuticle–cuticle CMC is lower than the disulfi de content of the delta layer in either cuticle–cortex or cortex–cortex CMC and that the carboxyl content in the cuticle–cuticle CMC is higher than that of the cortex–cortex CMC. In addition, Nakamura et al. added that the delta layer of the cuticle–cuticle CMC stains similar to the endocuticle. In 1983, Leeder et al. (44) in a TEM study involving the effect of solvents on wool fi bers found that formic acid treatment of wool fi ber modifi ed the CMC of the fi bers, but this effect was only observed between adjacent cortical cells and not between cuticle and corti- cal cells. These scientists suggested that these results are consistent with differences in the CMC between cuticle cells versus the CMC between cuticle and cortical cells. Peters and Bradbury (49) observed by electron microscopy that formic acid treatment of wool modifi ed the cell membrane complex of the cortex, “but that of the cuticle ap- pears unchanged.” They also analyzed “resistant membranes” from cuticle and cortical cells isolated by shaking wool fi bers in formic acid and oxidized with performic acid, and found differences, although their results for the Allworden membrane provided considerably lower values for cystine than the analysis of Allworden membranes by Allen et al. (30). Peters and Bradbury concluded that the “CMC of the cuticle differs from that of the cortex.” Leeder et al. in 1985 (50) described differences in the staining characteristics of the cuticle– cuticle CMC, the cuticle–cortex CMC, and the cortex–cortex CMC. After dyeing the fi bers with a uranyl dye, these scientists found a layer of dye around each cuticle cell, which was restricted to the CMC of the cuticle and not in the CMC of the cortex. They found only one dye layer at the cuticle–cortex junction and none in the cortex–cortex CMC, but two layers of dye in the cuticle–cuticle CMC, and they referred to the observa- tions of Nakamura et al. (8) on differences in the staining characteristics of these three types of CMC. Mansour and Jones, in 1989 (37), treated wool by Soxhlet extraction with chloroform/ methanol for fi ve hours, and subsequently in boiling water for 15 minutes, and examined