BEHAVIOR OF NEGROID HAIR 27 Therefore, these measurements were repeated on much shorter fiber specimens (1-2 mm) with a single region of twist. Angles of rotation for these shorter specimens are shown in 'Fable II. The position of the crossbar with a weight of 5 g was considered as the initial position (0ø), and the changes in angle with subsequent loading were measured with reference to this position. Angles of rotation in Table II usually increase in a given direction with load until the breaking point is reached. This indirectly confirms that the results in Table I were due to the presence of multiple twist reversals in the specimen. On some fibers, however, the angle of rotation decreases prior to break (specimens 1, 6, 8, 9, and others). It is suggested that this decrease is due to the initiation of cracks which relieve the torsional stress on the specimen, and that the growth of these cracks subsequently leads to fracture. In order to check this hypothesis, some fibers were saved as soon as the decrease in rotation angle was observed and examined in the scanning electron microscope. In a few of these fibers, cracks were indeed observed (Figure 5). It should Figure 5. Generation of a crack in the region of twist in torsional untwisting under the influence of a tensile load. a) 200x. b) 800x. c) 2400x. be noted, however, that this does not provide definitive evidence that they were formed during the stretching of the fiber, since occasional cracks or flaws have been found before stretching.

28 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS Table III Angles of Axial Rotation and Extension of Negroid Hair in the Region of Twist (65% RH, 21øC) (Specimen lengths 1-2 mm) Weight (g) Specimen Property 10 20 30 40 50 60 70 80 90 100 A -18 -44 --82 --222 -289 --305 -327 brk. 1 E 15 -- 23 28 38 49 54 A 7 15 6 6 saved 2 E 14 17 17 20 A 22 60 121 188 202 205 209 3 brk. E 7 7 17 42 45 46 55 A -8 -20 --72 --72 --67 4 saved E 1 1 4 12 24 A 3 9 25 25 13 5 saved E 2 7 13 40 42 A 2 8 13 17 35 38 42 42 51 6 brk. E 0 7 7 15 31 35 42 63 90 A 10 18 34 65 89 110 89 7 saved E 3 5 8 21 48 49 83 A 11 14 39 80 46 8 saved E 0 16 26 31 42 A --9 --24 --39 --66 --88 --88 --88 9 --105 brk. E 10 16 23 26 51 57 66 A ? 16 22 49 39 10 brk. E 6 7 2O 2O 36 A = Angle (deg). E = Extension (%). Initial condition: weight of 5 g corresponds to 0 ø. In an extension of this study, fiber elongation was measured along with measurements of the angles of rotation. The results are shown in Table Ill. Although angle of rotation and degree of extension increase together, both level (per unit length) and direction of twist vary considerably in different fibers. For example, very high levels of twist in opposite directions are observed in fibers 1 and 3 at relatively moderate extension levels, while in fiber 6 rather low angles of rotation are found even though the extensibility reaches much higher levels. These high levels of extension may be associated with the absence of critical flaws within the short length of fiber specimen. Fibers 2 and 4, on the other hand, show relatively low extension and twist angles. FRACTURE BEHAVIOR AND MECHANICAL PROPERTIES Fracture ends of fibers from Tables II and III were examined microscopically in an effort to understand the role of torsion in the fracture of these fibers. Obviously, in most cases fracture occurred in the region of twist. The most frequent fracture pattern was the step fracture, only one fiber showing a smooth fracture. There was no simple way of assessing the role of torsional deformation in these fractures except that a considerable number of fibers showed a slanted step fracture such as is shown schematically in Figure 6. A comparison of slanted and level step fracture patterns is