Introduction

Dearth of well-articulated information on genetic parameter estimates and diversity of upland rice limits the genetic improvement of rice. This study assessed the genetic parameter estimates and genetic diversity among 40 rice accessions using 26 agro-morphological traits. The trial was conducted in 2020 at the Njala University experimental site using 5 × 8 triple lattice design. The agro-morphological traits were analyzed using various multivariate and genetic parameter estimate techniques. Classification based on qualitative and quantitative traits grouped the germplasm into ten and five distinct clusters, respectively. Genotypes Buttercup-ABC, Buttercup-RARC, Jewulay, NERICA L4, Ndomawai, Sewulie and Painipainie produced earliest days to heading (81.8-97.2 days) and maturity (111.2 - 120.7 days). Genotypes Jasmine (3.036 t.ha-1), Rok 34 (3.238 t.ha-1) and Parmoi (2.663 t.ha-1) exhibited the highest grain yields. Principal component analysis (PCA) of qualitative traits exhibited four principal components (PCs) with eigenvalues > 1.0 and cumulative variation of 68.04%, whilst the PCA of quantitative traits had five PCs accounting for 81.73% of the total genetic variation. The findings indicate the presence of enough variability that could be exploited for the genetic improvement of rice varieties and the studied traits can be used for selection. Leaf blade length and width, culm diameter at basal internode, culm length, days to 50% heading, flag leaf girth, panicle number per plant, grain yield, and 100 grain weight had high heritability and genetic advance indicating the presence of additive gene action. Findings are relevant for conservation, management, short term recommendation for release and genetic improvement of rice.