Abstract: For the identification for cashmere and fine wool based scale pattern gene codes, the accuracy was studied. These researches show that though the same gene codes of cashmere and fine wool follow the same distribution, the numerical characteristics are different. The numerical characteristics of gene codes show that the scale of cashmere is more like a square or narrow rectangle and that of fine wool more like a wide rectangle. Among scale pattern gene codes, the distribution curves of scale area, perimeter and rectangle factor between two types of fiber are serious overlap, which imply that the three gene codes can’t be used to distinguish cashmere from fine wool. Except for that, the distribution curves of other scale pattern gene codes are partial overlap. So the different identification standard can be developed on the basis of minimum recognition errors for each scale pattern gene code and the maximum recognition probability of 88.8% for cashmere and 92% for fine wool can be got. These results provide a theoretical basis for the finding of optimal gene group for distinguish cashmere from fine wool.

Key words: cashmere, fine wool, scale pattern gene code, numerical characteristic, distribution