40 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS Table IX Fracture Pattern Distribution for Fibers Subjected to Tensile Fatigue at Different Loads at 65% RH No. of Fracture Type (%) Load (g) Specimens Smooth Step Angle Fibril. Split End tinfatigued (Table VIII) 0 36 6 56 6 22 10 Nonsurvivors 10 17 0 29 0 53 18 20 16 19 0 6 75 0 30 30 17 7 30 33 13 40 45 2 36 29 22 11 Survivors 10 74 3 76 8 7 7 20 66 6 73 11 8 3 30 43 7 81 9 2 0 40 29 7 79 7 0 7 pattern. Comparison with the fracture patterns of unfatigued fibers, which show 22% of fibrillated fracture ends, suggests that a significant number of these fibers with low intercellular cohesion between cortical cells fail during fatiguing. Many of these fibrillated patterns appeared similar to those of nylon fibers subjected to torsional fatigue (3). The possibility cannot be discounted, however, that in a significant number of fibers fatiguing was responsible for the destruction of intercellular cohesion in the cortex. This type of destruction may occur as a result of shearing between cortical cells due to torsional deformation. The percentage of fibrillated fractures seems to go through a maximum (at 20 g) as a function of load that is, although the total number of failures increases with the load, the fraction of fibrillated fibers decreases at the 30 and 40 g loads. It would appear that torsional strains go through a maximum as a function of load and that at higher load levels the growth of existing cracks under the influence of increasing axial strains becomes the predominant cause for the additional failures. Typical SEMs of the fracture ends of some of the fatigued fibers are shown in Figure 19. Many of these fibers fracture in a fibrillated pattern as in Figure 14. The fractures often seem to occur in the twist regions of the fiber. Fibers that survived 11 kc of fatiguing were subjected to standard Instron © tensile fracture experiments, and the distributions of breaking elongation are shown in Figure 20. It would be expected that the number of premature failures in survivors would be less than among unfatigued fibers and would decrease with load during fatiguing. This seems to be happening as the load is increased from 10 to 30 g, but at a load of 40 g the number of premature failures (among the survivors) appears to increase again, suggesting that these fibers were damaged (or preexisting damage was intensified) during the fatiguing procedure. After fatiguing under 40 g the fibers showed an irreversible increase in length (creep) of 3.4 _+ 2%. CONCLUSIONS Negroid hair fibers show a highly twisted configuration with a flattening approaching a collapsed structure in the region of twist. This results in high levels of ellipticity and

BEHAVIOR OF NEGROID HAIR 41 Figure 19. Fracture ends of untreated fibers that failed in fatigue under a load of 30 g. a) 1BOx. b) 105 x. considerable variation in the shape of the cross-sectional area. The twist is not unidirectional but shows frequent, apparently random, reversals along the fiber length. Negroid hair contains a relatively high proportion of weak fibers which, at 65% relative humidity, fail at extensions below 20%. This premature failure is eliminated when the fibers are fully swollen in water, suggesting that plasticization and swelling lead to fast relaxation of stress concentrations near potentially critical flaws. The fracture behavior of Negroid hair fibers has been investigated and five modes of