JOURNAL OF COSMETIC SCIENCE 36 observations by the authors as well as with those reported by various informal sources, where water-waved hair tresses have been hanging in laboratories with constant climate for years, apparently without losing their set. Figure 2 shows the effects of aging on hair recovery curves at different relative humidities and a common, initial aging time (tA=100 min), calculated on the basis of the param- eter values given in Table I for non-aging and for aging hair, respectively. Again, aging greatly delays recovery, and the recovery values, which the curves apparently approach for long times, increase with increasing relative humidity. This is in agreement with the practical observation that the set of a hairstyle based on water waving decreases with in- creasing humidity. Figures 1 and 2 show that for long times recovery approaches an apparent constant value. In order to calculate the “equilibrium,” the apparent fi nal value of recovery, Rf, for the dif- ferent humidity conditions, and for all practical purposes a very long time (log t=11, t3000 years), was chosen. The values for Rf are summarized in Figure 3 versus water content. The values closely follow a straight line, which describes the loss of set with increasing hu- midity. The straight line fi t predicts Rf =0.32 for dry hair and total recovery at a water content of about 25%. The latter value is in good agreement with the maximum water content of hair. The extrapolation validates the assumption underlying the calculations that total recovery will be achieved in water, that is, at conditions above the glass transition. Furthermore, the observation is in good agreement with the practical observation that wet- ting of hair removes the water wave. The tools that have been developed for the above calculations also enable as to assess the effects of fi ber damage on bending recovery namely, DSC-investigations (21,22) have shown that, damage, e.g., through oxidative or reductive processing, is fi rst and most strongly imparted to the fi laments. Plausibly expecting that such damage will lower the elastic modulus of the fi laments, Figure 4 shows recovery curves for an undamaged hair at 65% RH and a constant, short aging time of tA=10 min, together with the recovery Figure 2. Recovery curves at constant aging time (tA=100 min) and for different relative humidities (20°C), as marked for the solid curves for the aging material. The broken lines show the related curves, expected for non-aging hair.

VISCOELASTIC BENDING RECOVERY OF HAIR 37 of the same material in which the bending stiffness of the fi laments, B∞, is reduced to relative values of 70%, 50%, and 30%, respectively, while that of the matrix ΔB is kept unchanged. Figure 4 shows that damage to the elastic fraction of fi ber bending stiffness is expected to signifi cantly promote and stabilize the set for the aging fi ber, while the ef- fect in the non-aging fi ber will eventually be lost. In practice, however, the benefi cial ef- fects of fi lament damage are expected to be offset through collateral processing damage to the matrix and other morphological components, as well as the expected higher suscepti- bility of damaged hair to humidity and temperature changes, which will remove the benefi ts of physical aging. Figure 3. Recovery values of aging hair at log t = 11 (t 3000 years), considered as representing “equilib- rium,” fi nal recovery, Rf, for all practical purpose, vs water content. A straight line is fi tted through the data, for which the equation is given on the graph in the usual x/y-notation. The broken line marks the extrapola- tion range. Figure 4. Bending recovery curves of non-aging (broken lines) and aging hair (solid lines) at 65% RH and tA=10 min, in which the elastic bending stiffness of the fi laments, B∞, has been reduced from the initial, relative value of 1 to 0.7, 0.5, and 0.3, respectively. The curves for the aging material are marked accord- ingly.