Interest in hemp (Cannabis sativa L.) is increasing due to the development of a new range of industrial applications based on bast fibers. However the variability of bast fiber yield and quality represents an important barrier to further exploitation. Primary and secondary fiber content was examined in two commercial hemp varieties (Fedora 17, Santhica 27) grown under contrasted sowing density and irrigation conditions. Both growing conditions and hemp varieties impact stem tissue architecture with a large effect on the proportion of secondary fibers but not primary fibers. Attenuated total reflectance infrared spectroscopy allowed the discrimination of manually-isolated native primary fibers and secondary fibers but did not reveal any clustering according to growing conditions and variety. Infrared data were confirmed by wet chemistry analyses that revealed slight but significant differences between primary and secondary fiber cell wall composition. Infrared spectroscopy of technical fibers obtained after mechanical defibering revealed differences with native primary, but not secondary fibers and also discriminated samples obtained from plants grown under different conditions. Altogether the results suggested that the observed variability of hemp technical fibers could be partially explained by i) differences in secondary fiber production and ii) differential behavior during mechanical defibering resulting in unequal separation of primary and secondary fibers.

The distribution of lignin, 8-5' and 8-8' linked lignin substructure, and noncellulosic polysaccharides in hemp (Cannabis sativa L.) phloem fibers were explored based on histochemical and immunological methods. Ultraviolet absorption and potassium permanganate staining were observed mainly in the compound middle lamella (CML) and S1 layers, and rarely in the G-layer of phloem fibers, suggesting that lignin concentration is high at the CML and S1 layers, and very low at the G-layer of hemp fibers. Acriflavine staining, uniform KM1 labeling (8-5' linked lignin substructure), and no KM2 labeling (8-8' linked structure) were observed in the G-layer, suggesting that there is a small amount of lignin-like compound with 8-5' linked structure in the G-layer. In addition, some fiber cells showed a multilayered structure. Uniform arabinogalactan protein (AGP) labeling was observed on the S1 layers and G-layers using JIM14, but little appeared in the CML of hemp fibers, indicating that these layers of the phloem fibers contain AGP. Immunogold labeling of xylan (LM11) and glucomannan (LM21) showed that xylan and glucomannan were mainly present in the S1 layers and the G-layers, respectively. In some phloem fibers, LM21 immunofluorescence labeling showed multilayered structure, suggesting the heterogeneous distribution of glucomannan.