Chapter 8 267 Additional comments on the magnitude of the stress relative to the “Hookean region” of hair in varying humidity experiments: As mentioned already, the magnitude of the applied stress can create fatigue experiments that are performed inside or outside of the “linear” region for hair. Moreover, the position of the “linear” region yield point changes with the relative humidity. As such, as seen from Figure 9, a repeating 0.010 g/μm2 stress is within the “linear” region at 60% RH (solid data points), but beyond the yield point at 90% RH (hollow data points). Similarly, a repeating 0.014 g/μm2 stress is outside of the “linear” region at 60% RH (hollow data points), but within this region at 20% RH (solid data points). Therefore, care is needed in comparing results under different conditions as additional factors may be present. The 0.05 x 109 Pa stress in Table 3 was selected as it falls within the “linear region” under all three humidity conditions. Again it is noted that the data in Table 3 is based on stress-strain experiment performed at 40 mm/ min, and that the extension rate is much higher in these fatigue experiment. Weibull analysis of fatigue data: Whenever generating experimental results it is prudent to examine the statistical distribution of data points. Most commonly, we are familiar with a normal distribution, where data forms a symmetrical bell-shaped distribution about the mean. Figure 12 shows a histogram for breakage data arising from fatigue experiments on Caucasian hair at 60% relative humidity using a repeated stress in the range 0.010- 0.011 g/um2. Clearly, the data is not normal, but instead suggests the applicability of an exponential-like distribution. As outlined already, failure in a fatigue experiment is generally attributed to the propagation of flaws within a sample. Thus, with the presence of these flaws being statistical in nature, it becomes necessary to also treat failure as a statistical variable. Therefore, it will be shown how statistical analysis makes it possible to predict the probability of fibers breaking under different conditions.