enough quantities to affect hair swelling behaviour. So, despite the damage to the hair caused by the relaxer treatment and the pres- ence of cracks in the hair, only the leucine and low-MW hydrol- ysed keratins were able to penetrate deep into the hair in large enough amounts to change protein structures. We propose that although the hair was drying, these compounds, being a charged species, competed with native hair charged side chains over forma- tion of salt-bridges, as well as contributed to the repulsion of like charged species, which caused the hair cortex to swell by approxi- mately 11%. Investigation of mid- and high-MW peptide penetration using fluorescence microscopy showed that the mid-range MW keratin peptide was able to penetrate into the cortex. The high-MW keratin peptide only penetrated the outer layers of the cortex. We have to conclude that penetration into the hair cortex, despite being visible in the fluorescence microscopy, was not enough from either treat- ment to affect hair swelling. Although this study provided us a glimpse into the behaviour of peptide penetration, we would like to perform more studies which investigate the kinetics of peptide pene- tration into the hair in the future. Effects of treatments on hair stiffness Young’s modulus measures the resistance of a material to elastic (recoverable) deformation under load. A stiff material has a high Young’s modulus and changes its shape only slightly under elastic loads. A flexible material has a low Young’s modulus and changes its shape considerably. Young’s modulus of hair changes as a func- tion of relative humidity. At dry conditions, both matrix and inter- mediate filaments contribute to the hair resistance to stresses. The amorphous matrix, especially of damaged hair, is very sensitive to plasticization by water therefore, its contribution to hair strength decreases with increasing water content. Despite significant damage to intermediate filaments (as shown by denaturation enthalpy results), Young’s modulus results suggest that at 20% RH, matrix rigidity masked the damage extent thus, even severely damaged, relaxed hair was able to resist elastic defor- mations similarly to virgin hair. There have been suggestions that changes in amino acid composition can occur as a result of the relaxation process [30-33]. Cleavage of disulphide bond and forma- tion of lanthionine are the most commonly proposed changes [32, 33]. Moreover, studies [30, 31] agree that the levels of negatively charged amino acids remain unchanged. Unfortunately, these data contradict each other, as to whether there is a reduction in lysine [30] or arginine [31] content. Because of these uncertainties, it is entirely reasonable for us to propose that the composition of charged side chains remained the same. Based on this hypothesis, we propose that matrix rigidity of relaxed hair at 20% RH was maintained by (i) the positively and negatively charged side chains (unchanged content suggest similar level of salt-bridges and hydro- gen bonds), and (ii) increased contraction in the matrix due to par- tial disulphides replacement by lanthionine (shorter chains). This makes the elastic modulus an unsuitable parameter to assess the extent of damage, even between drastically different samples at very dry conditions. With increasing humidity, differences in Young’s modulus between virgin and differently treated relaxed hair continued to increase, revealing damage to intermediate filaments and to some extent, the matrix. As discussed in the results section, only high- MW peptide treated hair remained significantly stiffer than the relaxed hair control. Previously published studies investigating effect of peptides on mechanical properties of hair have shortcom- ings in mechanical testing or its description, such as very low numbers of replicates [10] and unspecified critical testing condi- tions (such as humidity) [15], making it difficult to make direct comparisons to our results. However, a significant increase in Young’s modulus, which is consistent with our results, was observed after (i) applying mid-MW peptide (10–12 amino acid) to relaxed hair (tested at 55% RH) [34] (ii) applying cD-crystallin pro- tein of 21 kDa MW on bleached hair (tested at 62% RH) [35]. Effects of treatments on hair breakage Obtained stress-strain curves shared a common theme – prema- ture breakage. Premature breakage is considered when the hair breaks at extensions below 20% Strain [18]. Kamath et al. postu- lated that premature breakage may occur because of weak point propagation during tensile experiments. Weak points may be gen- erated by grooming practices in the areas where cross-section structure changes (fibreflattens or collapses within a short dis- tance in the twist region) [18]. This parameter is highly consumer relevant as real-life premature breakage occurs under relatively low loads, such as while pulling hair during combing procedures. Performed preliminary SEM screening of several fibres of virgin and relaxed hair controls helped to understand the reason behind such high premature breakage. Fibres contained long longitudinal cracks of various depths. Relaxed hair treated with mid- and high-MW peptides had a fewer number of prematurely broken fibres than other differently treated relaxed hair or virgin hair. It is understandable that fibres with large cracks, like those imaged at 65x magnification, cannot be easily repaired. However, our results suggest that fibres with smaller cracks could have been repaired by mid- and high-MW peptides to mitigate premature breakage. At 80% RH, premature breakage in all of the tested hair types was significantly reduced because of fibre plasticization, especially at stress accumulation points [18]. Nevertheless, mid- and high-MW peptides significantly contributed to the reduction of premature breakage of relaxed hair. Interesting breakage behaviour was observed while comparing virgin hair at high and low humidities. With other types of hair, it is expected that the average break stress of hair broken at high humidity is lower than that at low humidity [36]. This trend was not observed in either study done by us or by Kamath et al. [18]. The anticipated higher average break stress at low humidity was reduced by the significant presence of prematurely broken fibres (broke at low stresses). Meanwhile, the average break stress at high humidity did not decrease proportionally as observed in Caucasian hair [36] because less textured hair fibres broke prematurely. This compensated for the anticipated lower average break stress at high humidity. Thus, average break stresses of textured hair at high and low humidities are not different. Although dimensional data suggest significant penetration by low-MW compounds, it appears that they did not have any pro- tein stabilizing effect due to their small molecular size. Meanwhile, both mid- and high-MW peptides were effective at improving breakage performance of relaxed hair. As fluorescence work revealed different depth of actives penetration, we hypothesize that the reduction in breakage for mid- and high-MW peptide treated hair may be due to different mechanisms. Mid-MW peptides, com- prised of ~20 amino acids, were made using a mild proprietary process that leaves the natural cystine content of the keratin in an active, S-sulpho form. Besides potentially stabilizing the hair 10 © 2020 The Authors. International Journal of Cosmetic Science published by John Wiley & Sons Ltd on behalf of Society of Cosmetic Scientists and Societe Francaise de Cosmetologie International Journal of Cosmetic Science, 1–12 Larger hydrolyzed keratins reduce hair breakage E. Malinauskyte et al.