Identifying, characterising and predicting visually objectionable and non-objectionable alpaca fibers by means of their cross-sectional characteristics

Visually objectionable alpaca fibers adversely affect fabric colour, contribute to undesirable sensations of prickle and itch in garments composed of these fibers and reduce fiber value. In this study, fibers from 50 Peruvian Huacaya alpacas were examined for fiber cross-sectional structural variables including any medulla, using the visualisations and algorithms of the Fiber Mult device. The following variables were analyzed: area, perimeter, minor axis, major axis, and ellipticity index, for both the fiber and the medulla. It was possible to successfully identify, characterise and predict visually objectionable alpaca fibers followed by the use of logistic regression models. Almost all visually objectionable fibers had medulla (99.6%), while the non-objectionable fibers mostly had no medulla (76.5%). All the physical characteristics measured in the objectionable medullated fibers were greater compared with the non-objectionable medullated fibers. These differences make it possible to reliably distinguish between visually objectionable and visually no objectionable alpaca fibers, and to predict their presence and levels in alpaca fiber sample, with high accuracy, sensitivity and specificity, using logistic regression models that can include a single variable (perimeter or area) or two predictor variables (perimeter and ellipticity index; medulla area and net area).