Analysis of morphological characteristics of six varieties of cassava using discriminant analytical approach

Abstract Cassava (Manihot esculenta Crantz) is a vital staple crop in sub-Saharan Africa, yet accurate varietal identification is often hindered by overlapping morphological traits influenced by environmental variability. This study applied Multiple Discriminant Analysis (MDA) to evaluate the discriminatory capacity of six morphological traits—leaf width, leaf length, leaf area, leaf index, plant height, and mean angle—across six cassava varieties cultivated in Nigeria. Data from 96 plants were analyzed using descriptive statistics, MANOVA, Mahalanobis distances, and canonical discriminant functions. Results showed that leaf area, leaf index, and mean angle were the most influential traits for separation. However, classification accuracy was modest, with an overall accuracy of 45.8% (APER = 54.2%). TMS98212 (75%) and NR100235 (56.7%) achieved the highest correct classification, while B515, COB61, IITA-IBA980581, and NWAGERI were completely misclassified. Canonical discriminant plots confirmed that TMS98212 and NR100235 clustered apart, whereas the other varieties overlapped extensively. These findings highlight the limitations of morphological traits as standalone classifiers and underscore the need to integrate molecular markers and advanced computational approaches for robust varietal identification.