EVALUATION OF HAIR FIXATIVES 73 CURL RETENTION ANALYSIS--SET I Resins A through E were evaluated by Curl Retention for 21 hr. Fourteen replicates were used to insure maximum statistical significance. Plots of average curl retention (per cent) versus time are shown in Figure 6. CURL RETENTION ANALYSIS--SET II Resin D, E and Formulation D1, El, E2 were evaluated by Curl Retention for 21 hr us- ing 7 duplicate samples. Plots of average curl retention (per cent) versus time are shown in Figure 7. TWIST RETENTION ANALYSIS--SET I Resins A through E were evaluated by TRA using 2 per cent resin solutions. The results from 5 replicates were averaged and plotted against time (Fig. 8). TWIST RETENTION ANALYSIS--SET II Resins D, E and Formulations D 1, E 1 and E2 were evaluated by TRA using 2 per cent resin solution. The results from 2 replicates were averaged and plotted against time (Fig. 9). 90 80 70 3O 2(•- i - X ION E1 F O R M'""•M UX•L L A E2 FORMULATION D1 20 40 60 80 100 120 TIME (MINUTES) Figure 9. Twist Retention Analysis--Set II--Average Twist Retention as a function of time for Resins D, E and Formulations D1, El, E2 at 90 to 93 per cent R.H., 21 to 25øC and 2 per cent fixative solids (0.19 per cent pickup)

74 JOURNAL OF THE SOCIETY OF COSMETIC CHEMISTS RESULTS AND DISCUSSIONS TWIST RETENTION ANALYSIS--TEST VARIABLES During the study of Twist Retention Analysis, a number of variables were investigated, i.e., the weight of hair per swatch, amount of applied twist, air flow rate, and swatch reuse characteristics after shampooing. It was found that at a fixed degree of twist, as the number of hair fibers (weight) per swatch was increased, the Twist Retention increased. This was reasonable since, as the number of hair fibers in the sample is increased, the number of cross-over points per unit length was also increased. This resulted in an increased number of hair-to-hair bonding sites. To eliminate hair weight as a test variable, a constant weight of 0.20 to 0.21 g per swatch was selected. The amount of applied twist was varied and studied between 180 and 900 degrees in 180 degree increments. Five replicate samples were evaluated at each twist increment and statistically compared. As the twist was increased, the coefficient of variation (For- mula IV) decreased (increased precision) and was minimized above 540 degrees. An applied twist of 720 degrees was selected as optimum. Air flow rate was varied between 15 and 200 ml/min. Flow rates below 40 ml/min gave rise to non-reproducible results, while the flow rates above 55 ml/min gave constant Twist Retention/time relationships. Sample-to-sample reproducibility at 40 ml/min or below was poor, while at or above 55 ml/min it was significantly (acceptably) improved. Under low flow rate conditions it is suspected that the hair/fixative com- posite partially dehydrates the air, giving a non-reproducible Twist Retention response. Swatches were tested after 20 fixative application/shampoo cycles to study the effect of possible swatch deterioration. No significant differences were observed between the first and the last shampoo cycle. Swatches were, however, discarded after 20 shampoo cycles. STATISTICAL ANALYSIS Resins A through E, evaluated by both Curl and Twist Retention Analysis--Set I, were compared using standard statistical techniques. Coefficient of variation, correlation coefficients, and confidence of difference between resins were calculatd as a function of time. TABLE II COEFFICIENT of VARIATION RESIN TWIST RETENTION CURL RETENTION SET I SET I AVERAGE 0 VERALL AVERAGE 0 VERALL TIME INTER VALS TIME INTER VAL S A 9.2 17.1 B 10.7 20.5 C 9.1 24.2 D 7.9 20.0 E 4.7 11.9