Key points are not available for this paper at this time.
Abstract Roselle (Hibiscus sabdariffa) is a Malvaceae with multiple economic and pharmacological potential, little or poorly known and under-exploited in Cameroon. The aim of this study was to assess the genetic variability of 36 collections from several countries in the Central African sub-region, in an alpha lattice design (6x6) with two replications at the University of Yaoundé I. The data collected from five plants in each plot per replication were randomly selected from the middle row, excluding the border plants, and harvested at the stage of 50% flowering. Principal Component Analysis (PCA) and Hierarchical Clustering Analysis (HCA) were estimated using the R software variability package (version 3.5.1). The analysis of variance results from the 20 traits evaluated including 13 qualitative and 7 quantitative showed significant differences (P < 0.0001) between accessions at 60 days after sowing. Average heights to first branch ranged from 2.62 to 14.70 cm and the plant heights ranged from 10.5 to 86 cm. Mean collar diameter values varied between 0.40 and 1.36 cm. Collar diameter was strongly and positively correlated with the number of branches (r = 0.8) and with the 50% flowering date. Flower bud initiation date (DIBF) was negatively correlated with chlorophyll index (-0.6) and plant height (-0.4). Hierarchical clustering grouped the collections into 03 groups of 10 collections, 12 collections and 14 collections respectively. However, no differences were observed between collection origins for all parameters studied. High heritability was recorded for growth parameters such as collar diameter (0.6143), number of branches (0.6363), plant height (0.7240) and height of first branch (0.7653), revealing strong potential for selection and genetic improvement. The first two axes of the PCA which explained 64.7% were taken into account. The association of quantitative variables and collections showed the top 10 collections which represented potential parents for breeding programs.
BENOIT-CONSTANT et al. (Tue,) studied this question.