The information content of tensile tests of human hair (wet) is limited: Variables mainly cluster in just two principal components

Abstract

Tensile testing of keratin fibres, such as wool and hair, in the wet state is a well established tool in the academic as well as applied portfolio of mechanical analyses. For the tensile curve of a hair fibre, fourteen material-specific stress-, strain, and work-related variables were identified. Analysing twelve samples of cosmetically untreated hair, we show that the variables show good precision and a satisfactory degree of homogeneity across samples. Correlation analysis shows strong relationships between essentially all variables. Some strong correlations were in fact expected. However, their actual extent is very surprising, including even rather dissimilar variables (e.g., elastic-vs break-modulus). This observation led us to investigate whether variables may be clustered into a number of groups in order to define a set of underlying orthogonal factors, using Factor Analysis. The results show that 87% of data variance can be accounted for by just two factors, which we refer to as ‘stress-’ and ‘strain’- factor, respectively. The tensile properties of untreated hair (wet) are thus well described by just any pair of strong variables from each factor, such as elastic modulus and break strain. We hypothesize that similar principles may also be applicable for treated hair. This type of approach and its implications may also have relevance for other epidermal appendages (nails, claws, horn, etc) or even more generally for other biological composite materials, such as skin, tendon and bone.