216 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS 60 minutes. The fiber was extended 20% in water and held for five seconds. Only those fibers exhibiting a linear Hookean region and an acceptable Yield region (turnover of 2-5 % and slope approximately equal to 0.1 times the slope in the Hookean region) were used for SFTK measurements. After the untreated hair was measured, it was allowed to re-equilibrate overnight in water to recover its tensile properties. Chemical stress relaxation. When a hair fiber is extended in water by less than 2% of its original length, stress relaxation is complete in approximately 30 minutes. A strain- cycling procedure was used to reduce the time necessary for stress relaxation (1,26,27). The fiber was extended to 2.0% strain in a buffer solution of the same pH and tem- perature as the reducing solution. Following extension, the hair fiber was allowed to stress relax for 60 seconds. The amount of strain was reduced to 1.50% in order to approximate permanent waving conditions, and the fiber was allowed to stress relax until a constant level of stress was achieved (2-3 minutes). After stress relaxation, the buffer solution was replaced rapidly by the reducing solution. SFTK data were collected to monitor the stress relaxation of the disulfide bonds caused by the reducing agent in the hair fiber. The data were collected as grams of force vs time and analyzed using the pseudo first-order kinetics model first described by Reese and Eyring (21). This model has been discussed in detail by Wickett (1,3). The force-vs-time curves follow an exponential decay: F t = F o e -kcøt (4) where F t is the force at time, t, F o is the force at time 0 (before applying the reducing agent), F c is the final force value reached (if reduction is complete, then F c = 0), C O is the concentration of the reducing agent, and k is the pseudo first-order rate constant for the reaction (1,21,26). Rate constants were obtained from the slopes of plots of the natural logarithm of [(F c - Fc)/(F o - Fc)] vs time, which are linear with an intercept of zero when the reaction is truly pseudo first-order. AMINO ACID ANALYSIS METHODOLOGY Mobile phases. A two-solvent system was used with a gradient elution program. Mobile phase A consisted of a 20-mM sodium acetate trihydrate buffer with 0.018% (v/v) triethylamine, and 1.5 ml of tetrahydrofuran adjusted to pH 7.20 + 0.05 with a 1% solution of glacial acetic acid. Mobile phase B consisted of 20% of a 100-mM sodium acetate trihydrate buffer, 40% acetonitrile, and 40% methanol adjusted to pH 7.20 + 0.05 with a 1% solution of glacial acetic acid. Both mobile phases were degassed using continuous helium purging for 15 minutes. Hair sample preparation. Hair from a single individual was extracted with ethanol in a Soxhlet apparatus for 24 hours to remove grease and other materials that may be deposited on the fibers. The extracted hair was soaked in an excess of methanol, rinsed with Millipore water, and allowed to dry. S-carboxymethylation of cysteine. Bundles of hair weighing 200 mg were prepared and reduced for a specified time length (0-30 minutes). Temperature during reduction was held constant at 23øC with a circulating water bath to reduce temperature and concen- tration variation of the reducing solution. Each tress was immersed in an excess (50 ml) of reducing solution. When reduction was complete, the tresses were immediately immersed in a cold 1% solution of iodoacetic acid. This step effectively blocked RS-

REDUCTION OF HUMAN HAIR 217 groups of the reducing agent and prevented further reduction from occurring by reacting slowly enough to effectively wash off the reducing solution only. The tresses were reacted with a 1% sodium iodoacetate solution, pH = 8.3, for 30 minutes at 75-80øC to form S-carboxymethyl derivatives (SCMC) of all cysteine residues (Eq. 5). The sam- ples were then dried in an oven at 110øC until constant weight was reached. Ker - S-H + I - CH 2 - CO - O-Na + Ker - S - CH 2 - + -• + Na + I- (5) cysteine sodium iodoacetate CO -- O- Na + S- carboxymethyl cysteine Hydrolysis and derivatization. The methodology employed was a modification of the recommended procedure for vapor phase hydrolysis given in the AminoQuant II oper- ator's manual (for a detailed description, refer to reference 27). The samples were hydrolyzed for 24 hours at 110øC with dithiodipropionic acid (DTDPA) added as a protection agent for unreacted cystine, re-dissolved in 0.1 N HCI containing a known amount of an internal standard solution, and filtered with a 0.45-}xm filter. As no residue was left behind after redissolution in 0.1 N HC1, the amino acid content was calculated on the assumption that the hydrolyzed material represented the amino acid composition of the total fiber. The use of DTDPA as a capping agent for cystine was twofold. First, DTDPA protects against the loss of sulfur compounds by oxidation during hydrolysis (32). Second, by converting all non-carboxymethylated cysteine (i.e., mixed disulfide) and cystine to cysteine mercaptopropionic acid (Cysteine-MPA), DT- DPA allows for the separation of cystine from any salts of mercaptan that may remain in the fibers despite the carboxymethylation step (27). DTDPA will react with these by-products and combine with OPA to form isoindole derivatives that may be detected separately from the cysteine-MPA and SCMC derivative (27). This method thereby eliminates the introduction of considerable error into the cystine assay by formation of separate derivatives. However, the percentage of mixed disulfide included in the cys- teine-MPA derivative must be accounted for in the SCMC derivative. Figure 1 shows a sample chromatogram for untreated and treated hair samples. Data collection. The data were collected at a wavelength of 338 nm for the OPA deriv- atives having a bandwidth of 10 nm. A baseline value for total combined cystine and cysteine content in the hair sample was determined by the analysis of an untreated sample. This value was used as a control for the combined value of SCMC and cystine- MPA derivatives isolated in treated hair samples. The area of the SCMC and cystine- MPA peaks was used to determine the increasing amounts of cysteine and decreasing amounts of cystine with increasing reduction time, respectively. The sample concen- tration was detected as picomoles/mg and converted to }xm/g of sample in order to maintain uniformity with previous amino acid studies reported in the literature. RESULTS AND DISCUSSION SFTK MEASUREMENTS A pre-stretch in water or a buffer solution to a constant level of force will effectively break the hydrogen bonding network in the hair ( 1-3,21-25,28,29). Once this network