250 JOURNAL OF COSMETIC SCIENCE more readily than impacting a hair fiber over a comb tooth that is about ten times as thick. In another attached-loop experiment (see Table II), 11 of the top impacting hairs broke on 16 impacts while seven of the bottom hairs broke, all on the first impact. There is no significant difference in breakage by top versus bottom hair using the chi square statistical test procedure. The diameters of the hair fibers used in this experiment were measured, and a significant diameter effect was found by chi square, showing that the smaller-diameter hairs break more readily whether on the top or the bottom of the impact. Twelve finer-diameter hairs broke, and two thicker hairs, and four broke si­ multaneously by impact loading under these conditions. In this experiment, the top hair was in the form of a loop, but the bottom hair was essentially straight, suggesting that severe bending is not necessary for breakage to occur because loops do not appear to break more easily than straight sections of hair. This effect will be examined in more detail in another experiment. FRAGMENTATION PATTERNS OF HAIRS FROM COMBING AND IMPACT LOADING The broken hairs from all of these experiments were collected, labeled, and held for microscopic examination of the ends. Hair fiber fragments from the combing experiment described in the first paper of this series (1), which had been separated by fragment length, were also examined under the light microscope. The ends of the 6.35-12.7-cm fragments and the fragments less than 1.27 cm in length from the combing experiment, and the broken hairs from attached-loop and equalized-loop experiments, were examined and classified according to smooth fractures, step fractures, fibrillated ends, and splits as described by Karnath, Hornby, and W eigmann (3 ,4). The results are summarized in Table III. The ends of the longer hair fragments from the combing experiment showed more fibrillation and splits than the ends produced by the attached-loop impact-loading experiments. Therefore, additional impact-loading experiments were run, where the ends of the hairs were both bound or equal (equalized loop), and the effects of abrasion immediately before impacting on the fragmented ends were also examined. Rubbing over the spot (to be impacted) prior to impact produced more fibrillation and splitting similar to the ends produced from the combing experiments. In addition, the equalized-loop type of breakage provided fragmentation percentages Table II Effect of Fiber Diameter on Impact Loading and Top-Versus-Bottom Hair Breakage (attached-loop setup) Top hair (loop) Bottom hair (straight) 11 Breaks of 16 impacts 7 Breaks of 16 impacts 16 Impacts and 18 broken hairs (above). No significant top-vs-bottom effect by chi square. Diameters measured and 12 thinner-vs-thicker broke, 2 thicker broke, and 2 broke simultaneously (sig­ nificant diameter effect by chi square). 33-gram load, 3-4-cm drop, 40% RH, fibers 200 ± 20 mm.

IMPACT LOADING AND HAIR BREAKAGE 251 Table III Classification of Broken or Fragmented Ends From Combing versus Impact-Loading Experiments Size 6.35 to 12.7 cm 1.27 cm Combing (N = 3 5) Steps 56% Splits 14% Fibrous end 14% Smooth 16% Very fibrous with splits Atrached loops (N =22)* Steps 82% Splits 9% Fibrous end 9% Equalized loops (N = 20)* Steps 60% Splits 10% Fibrous end 20% Smooth 10% * Rubbing prior to impact produced mainly fibrillation and splitting with very fibrous fragments. more similar to those of the larger fragments from combing. However, once again there was not as much splitting or fibrillation, and the step fractures were shorter, leading to the conclusion that the fragmented ends are probably a combination of these two types of breakage, with some abrasion prior to and even after impact producing more fibril­ lation, splitting, and fibrous fragments. The smallest fragments from combing (those less than 1.27-cm length) were even more fibrous and fibrillated and they contained larger segments of exposed cortex than the longer fragments (Figure 4). In fact, some of the shorter fragments were largely exposed cortex. These differences in fragmentation collectively suggested that the mechanism for breakage for the shortest fragments might be different than for the longer fragments. Therefore, the short fragment formation was re-examined, starting with another comb­ ing experiment. FRAGMENATION BY COMBING In this experiment, tresses from 14-inch Caucasian hair were combed in 25 comb stroke increments up to 100 comb strokes, and the broken fragments were collected on a large Split & All Cortex Figure 4. Light micrographs of the shortest fragments from combing tresses.