QUANTITATIVE STUDY OF DEPILATORIES UPON IIAIR 703 (a) Comparisons are made within a given hair type. Caucasian should be compared to caucasian, negroid to negroid, and orieutal to oriental. This is done since the ellipsoidal cross-sectional characteristics of hair vary with hair type (e.g., oriental being much more circular than caucasian). Diameters will have meaning only if observed in some con- sistent manner in terms of the reference type chosen. (b) Only sections of hair of fairly uniform width, in which the tapering effect is minimal, are utilized for the determination of T•/and If these two precautions are taken, the normalization process then consists of dividing Td and T• by the mean diameter of the hair, as aver- aged from two points at its extremes. For convenience of measurement, the diameter of the major axis is utilized. The resulting normalized quantities are denoted by To and T•. In hair with a marked difference between major and nilnor axes, the degree of swelling of the minor axis is appreciably greater than that of the major axis on completion of swelling, the cross-sectional area will in- variably be more circular. The authors were concerned that the hair, while swelling in diameter, might at the same time also be uncoiling longitudinally. This would create two problems: (a) Longitudinal swelling would mean that the point on the hair shaft, whose diameter was determined at one position of the traveling microscope, would no longer be at that position when the next determi- nation was made. This, however, did not prove to be a serious source o[ error. From Fig. 2 it may be seen that most of the cross-sectional swell- ing is complete before appreciable longitudinal swelling sets in. (b) The possible twisting of hair on longitudinal swelling could create apparent changes in diameter merely by reorienting the axis being measured. The following two experiments were devised to deternfine whether this twisting did in fact exist: A thin steel wire in the form of a pointer was cemented perpendicular to the long axis and at the lower portion of the hair. A reference point was marked on the outside of the test tube. The position of the pointer, after swelling, was noted and found unchanged. The second, more sensitive, experiment utilized a small mirror ce- mented to the hair in the manner of the pointer in the first determination. This created, in effect, a configuration similar to that found in ballistic galvanometers, where a light beam is utilized to create a long pointer

704 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS Fig•re 4. 0.6 --•_ 443 - I 2 :5 4 5 6 TIME, MIN. Differential longitudinal swelling of two hairs in the same depilatory preparation arm. Very small changes in are may be discerned by this procedure. The result of this experiment was in agreement with the first. No rota- tion was detected. In contrast to the determination of cross-sectional swelling, the measurement of maximum rate of longitudinal swelling is quite straight- forward, independent of hair orientation, and practically independent of surface damage. The independent nature of the maximum rate of ex- tension is readily apparent if one considers its time of occurrence to be a functiou of the summation of the individual rates of extension along the entire hair length. If a small section of hair is damaged such that the time at which its maximu• rate of extension occurs differs appreciably from that which would be expected of the hair as a whole, its extension, being small by comparison to the total extension, will not affect a change iu T•,•. This may be seen in Fig. 4, where two hairs are placed in thio- glycolate solutions of identical concentration and pH. Although the plots are quite dissimilar (T/(443) = 4.0 minutes, T/(444) : 4.5 minutes) when T• is normalized to T•, the values obtained are equal (Tt(44a) = 4.0 minutes, and T/(444) : 4.2 minutes). These values approach the maximum precision attainable when T is small since readings are normally taken every 15 seconds. Sections of several hairs were abraded with wooden spatulas, and their T• values compared with the corresponding unabraded section. No appreciable change in T• was discernible.