MECHANICAL HYSTERESIS OF CHEMICALLY MODIFIED HAIR 407 this study. If the previously given model is correct, the viscosity of the matrix in the case of esterfried fibers must decrease up to the transition point and then must increase at a slower rate. As yet, no explanation can be given for such behavior. Alexander and co-workers (34) noted no main chain hydrolysis in wool after four hours of treatment at 65 øC with 0.01N HC1 in methanol, although 70% of the carboxyl groups were esterfried. The present study suggests that some hydrolysis has taken place after 20 hours as shown by a 22% lowering of the work to extend the fiber to 20%. However, this effect is not apparent in the hysteresis ratio. Reduction and Blocking or Crosslinking W20 and H20 of hair reduced with either thioglycolic acid, phenyl mercaptan, or ethyl mercaptan and subsequently blocked with iodoacetic acid (treatments 9a-d in Table I) are essentially the same, although Maclaren (24) suggests differences in the degree of reduction of wool by these disulfide bond breakers (Table II). On the other hand, benzyl Table II % Reduction of Wool (by Reagent in 1:1 n-PropanolfWater) • Thioglycolic acid 28% Phenyl mercaptan 40% Ethyl mercaptan 72% Benzyl mercaptan 85% • After Maclaren (24). mercaptan causes a much larger reduction in W20 and also effects an appreciable change in H20 at various temperatures. The relatively minor effect of the damaging treatments on the hysteresis ratios can probably be explained by the fact that the introduction of free carboxyl groups (by alkylation with iodoacetic acid) into hair may also establish new salt links or other bonds within the fiber. It is also noted that large differences in H20 between the control fiber and the reduced and blocked fibers occur only at temperatures below the transition temperature. This finding suggests that disulfide bonds play a definite role in the slope of the H20 vs. T curve at low temperature. In the absence of disulfide bonds, there may be no transition point and a completely reduced keratin fiber may thus not exhibit a transition point at all. Crosslinking of the reduced fibers (treatments 10a-c in Table I) not only raises W20 but also "normalizes" the hysteresis ratio and the transi-

408 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS tion temperature to values which approach those of the control fibers more closely. It appears, therefore, that crosslinks, just like disulfide links, have an effect on the transition temperature but that the length of the crosslink is relatively unimportant. Deamination Speakman and Stott (26) used the fact that a completely deaminated fiber gives the same extension curve in water as in 0.1N HC1 as an indica- tion of the extent of deamination. Deaminated fibers would be expected to exhibit the same extension curve in water as in 0.1•V HC1 because HC1 breaks the salt links in the fiber and, therefore, prevents the inter- action of free carboxyl groups with the free amino groups in the fiber (27). The present study introduces a possible second criterion for the degree of deamination, i.e., the linear relationship between hysteresis ratio and temperature. It should be noted that treatment with ninhydrin also causes deamina- tion. As soon as the temperature is high enough to overcome the in- fluence o[ deposition of bulky groups, i.e., beyond 30 øC, the hysteresis- temperature plot approaches the slope of hair that has been deaminated with nitrous acid. As noted previously, the hysteresis-temperature plot for PFO treatment is also linear and could be due to both a-helix disruption and breaking of salt links. Stress-Strain Curves The authors are not aware of any published report which relates processes at the atomic or molecular level with the macroscopic picture shown by the unloading curves. One of the striking features of these curves is the fact that they differ in shape: One extreme is the shape of untreated fibers (Fig. 2B) the second shape is that of fibers treated with phenyl isocyanate (Fig. 2B). In the former, an immediate loss in stress is followed by a slow decrease in stress (almost parallel to the slope of the yield region in the loading curve) and finally by a rapid decrease of stress to the starting point. The final decrease probably occurs at the point at which reformation of hydrogen bonds takes place. The shape of the second type of curve is different. In this case, the stress drops very rapidly to zero while the strain is still relatively large, as much as 15%. An analogy might be drawn between this behavior and that of a spring which is extended (or is allowed to contract) in a very viscous medium. During the loading cycle, a large external force is required to extend the spring, due to the surrounding viscous medium. During the unloading cycle, only internal forces are available to con-