JOURNAL OF COSMETIC SCIENCE 106 to its matrix cross-link density and α-helical content. In fact, this is what is observed with African-American hair however, Jamaican hair, which had the second highest mechanical integrity, has the lowest thermal values. Additional studies are needed to assess chemical and structural differences that may be responsible for these results. Protein variations as a function of ethnic group were explored using 2D-gel electrophoresis and the results are shown in Figure 8. In all cases, minor variations in staining intensity occur when the IF proteins have MWs between 50–75 kDa. Similar results were reported by Nagai et al. and Khawar et al. (32,33) and Nappe and Kermici (34). In our study, most Figure 7. Thermal properties according to ethnicity n = 5 hair samples / ethnic group. (a) Mean denatur- ation temperature ± standard error. (b) The mean change in denaturation enthalpy ± standard error. Figure 8. 2-D gel electrophoresis of hair from different countries n = approximately four fi bers from 23–25 subjects / ethnic group. The presence and the intensity of spots give insight about specifi c proteins.

2008 TRI/PRINCETON CONFERENCE 107 Table II Two-Way ANOVA p-Values Property Ethnicity Curl class Cross-sectional area 0.038∗ 0.162 Ellipticity 0.212 0.082 Young’s modulus 0.001∗ 0.165 Break stress 0.001∗ 0.001∗ *Signifi cant when one-and two-way ANOVA are consistent and p 0.05. variations were found with keratin associated proteins (KAPs) at basic isoelectric points pH 9. Particular to our observations, more KAPs spots are present in the circled por- tion of the gel for hair from African countries compared to Jamaican and African-Amer- ican hair. Additionally, the spots indicated by arrows vary with ethnic group indicating different protein concentrations. These differences may be attributed to possible ethnic or origin-related variations. From the above studies it can be seen that hair characteristics change according to degree in curl and ethnicity. However, since some ethnic groups have curlier hair compared to others, it is diffi cult to assess whether the results are related to ethnicity or to type of curl or to both. So, answer this question, statistics using one-way and two-way ANOVA were conducted on hair from types VI and VII where there was a larger number of hair samples representative of each ethnic group. The results are shown in Table II. It was found that ethnicity was the sole driver for values in cross-sectional area and Young’s modulus, whereas curl class did not solely infl uence any of the parameters discussed in this paper. Both ethnicity and curl type affected variations in break stress but neither of them affected ellipticity. It seems that these statistics do not agree with our earlier results of them affected ellipticity. It seems that these statistics do not agree with our earlier results since Figure 3 shows a statistically signifi cant correlation between ellipticity and degree of curl. However, when the reduced range in ellipticity values for the latter statistical method is considered (1.67–1.71), it is clear that the reduced range is responsible for the inconsistency. In order to better understand the infl uence of curl type using this method, more data is needed where more curl types, and hence, ellipticity values, are represented for the different ethnic groups being studied. Nev- ertheless, the amount of samples represented within both ethnic groups, particularly curl type VI, statistically demonstrates the effect of ethnicity. The infl uence of geometry can easily explain why the mechanical properties in hair differ. The geometrical twists and kinks that are more prevalent in curlier hair can be local points of weakness that act as stress concentrators and result in higher fragility. However, when hair from different global regions that has the same curl is compared, as was done in the latter statistical method, the impact of geometry is signifi cantly reduced and one has to consider other possible factors for the observed differences. Regional and cultural diversities, such as diet, environment and grooming practices, may contribute to these differences however, more research is needed to understand the effect of ethnicity. CONCLUSIONS This study has shown that hair fi bers with a higher degree in curl are typically thinner, more elliptical and weaker than hair that has looser curls. Based on curl classifi cation, in