Impaginato 277 Adv. Hort. Sci., 2021 35(3): 277­284 DOI: 10.36253/ahsc­10523 Genotypic diversity and trait profiles of some Amaranthus accessions O.A. Oyetunde 1 (*), M.O. Olayiwola 2, B.T. Osho 3 1 Department of Crop Production and Horticulture, Lagos State Polytechnic, Ikorodu, Nigeria. 2 Department of Crop Production, College of Agricultural Sciences, Olabisi Onabanjo University, Ago‐Iwoye, Nigeria. 3 Department of Agronomy, Pan African University for Life and Earth Sciences (Including Health and Agriculture), University of Ibadan, Ibadan, Nigeria. Key words: Amaranthus, biplot, improvement, trait profile, variability. Abstract: Knowledge of the pattern of trait variation among accessions, and the trait profiles of the accessions is crucial for improvement of a crop. Twenty­one Amaranthus accessions were evaluated in 2018 and 2019 to investigate the extent of genotypic diversity among the amaranth accessions and their trait profiles. Data were subjected to analysis of variance, and correlation and prin­ cipal component analyses. Taking stem weight (SWT) as the yield, the accession × yield­trait combination (GYT) biplot was employed to investigate the trait profiles of the accessions. Accession, year, and accession × year mean squares were significant (P≤0.05/0.01) for most of the measured traits. The first three principal components explained 88.18% of observable variation among the accessions and identified plant height (PHT), number of leaves per plant (NOL), and root weight (RWT) as the major contributors. Significant (P≤0.01) correla­ tion was observed in the association of SWT with NOL, TBM, and RWT. Accessions NGB00019 and NGB00061 were associated with the sector contain­ ing all the YT combinations considered. However, NGB00019 was identified as the best combiner of yield with other traits. Further studies involving more traits should determine the trait profiles of the remaining accessions. 1. Introduction Amaranthus is a member of the Amaranthaceae family and is believed to have originated from South America (Janovská et al., 2012). Despite the level of under­development, the genus is one of the most diverse among cultivated crops, with about 70 species (Ebert et al., 2011). Amaranths are classified based on the part of the plant for which they are grown. The leaf types including A. hybridus and A. tricolor are grown for their leaves consumed as vegetables while the grain amaranths such as A. caudatus, A. cruentus, and A. hypochondriacus are popular for the grains. In addition, some amaranth like A. tricolor are valued as ornamental, while some (e.g. A. palmeri S. Wats., A. powellii S. Watt., A. retroflexus L., and A. (*) Corresponding author: biyi.oyetunde@gmail.com Citation: OYETUNDE O.A., OLAYIWOLA M.O., OSHO B.T., 2021 ­ Effect of vine and fruit pruning on yield attributes of two watermelon Citrullus lanatus) cultivars. ­ Adv. Hort. Sci., 35(3): 277­284 Copyright: © 2021 Oyetunde O.A., Olayiwola M.O., Osho B.T. This is an open access, peer reviewed article published by Firenze University Press (http://www.fupress.net/index.php/ahs/) and distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Data Availability Statement: All relevant data are within the paper and its Supporting Information files. Competing Interests: The authors declare no competing interests. Received for publication 26 February 2021 Accepted for publication 2 August 2021 AHS Advances in Horticultural Science https://doi.org/10.36253/ahsc-10523 http://www.fupress.net/index.php/ahs/ http://creativecommons.org/licenses/by/4.0/ http://creativecommons.org/licenses/by/4.0/ http://creativecommons.org/licenses/by/4.0/ Adv. Hort. Sci., 2021 35(3): 277­284 278 spinosus L.) are considered as weeds. In general, cul­ tivated amaranths are of nutritional importance because they produce edible leaves, stems, and grains (e.g., A. hybridus and A. tricolor) (Akin­Idowu et al., 2016; Neelesh and Pratibha, 2018). The leaves and stem are rich natural stores of vitamins (A, B, and C), and dietary minerals including calcium and iron (Stallknecht and Schulz­Schaeffer, 1993). Amaranthus has also been found to be a source of lysine, an essential amino acid that is lacking in diets based on cereals and tubers (Schippers, 2000). The vegetable is reported to be adapted to a wide range of agro­ecol­ ogy habitat (Katiyar et al., 2000) and can tolerate biotic (i.e. diseases and pests) and abiotic (i.e. heat and drought) stresses (Shukla et al., 2010). As is with many other under­utilized leaf vegetables, the genet­ ic and economic potentials of Amaranthus have not been fully harnessed. Although the amaranth is essentially self­pollinat­ ing, there is significant level of natural outcrossing and inter­specific and inter­varietal hybridization, leading to the occurrence of wide differences among accessions (Akin­Idowu, 2016). There is a growing interest in research on amaranth because of the high genetic diversity and phenotypic variability of the crop which are of immense use in designing strate­ gies for improvement of the crop. Plant breeders have found differences in plant observations among genetic materials as effective means to deduce esti­ mates of genetic diversity (Akin­Idowu, 2016) and to determine genotypic superiority. However, identifica­ tion of genetic superiority is hindered by unfavourable associations among a set of target traits since the decision is often based on multiple traits (Yan and Fregeau­Reid, 2018). This implies that a successful cultivar must attain desirable levels for several key traits. Thus, knowledge of the pattern of trait variation among accessions and the trait profile (strength and weakness) of available Amaranth accessions will aid the exploitation of existing geno­ typic diversity for improvement of the crop for desired traits. Recently, the accession x yield­trait (GYT) biplot approach was proposed by Yan and Fregeau­Reid (2018) to combat the challenge posed by evaluation of accessions based on multiple traits, In this approach, the worth of a accession is deter­ mined by its value for yield in combination with other traits (Y­T) rather than its levels for an individual trait. Therefore, this study investigated the extent of geno­ typic diversity among amaranth accessions, as well as the trait profiles of the accessions with a view to pro­ viding information to aid effective future improve­ ment strategies. 2. Materials and Methods Genetic materials Twenty­one (21) amaranthus accessions (NGB00001, NGB00005, NGB00019, NGB00022, NGB00025, NGB00027, NGB00028, NGB00029, NGB00031, NGB00024, NGB00058, NGB00059, NGB00060, NGB00061, NGB00070, NGB00078, NGB00082, NGB00108, NGB00111, NGB00112, and a local check named LASPO­COL­001). All the acces­ sions, except the local check, were obtained from the National Agency for Crop Genetic Resources and Biotechnology (NACGRAB), Ibadan, Nigeria. Seeds of the local check were collected from a reputable farmer in Ikorodu. All the accessions belong to the species A. hybridus, and are of Nigerian origin (Table 1). Field evaluation and phenotyping The 21 amaranthus accessions were evaluated on the field during the main seasons of 2018 and 2019. Table 1 ­ Origin of the 21 Amaranthus hybridus accessions used in the study Serial number Accession name Origin 1 NGB00001 Katsina State, Nigeria 2 NGB00005 Kebbi State, Nigeria 3 NGB00019 Lagis State, Nigeria 4 NGB00022 Zamfara State, Nigeria 5 NGB00025 Niger State Nigeria 6 NGB00027 Osun State, Nigeria 7 NGB00028 Osun State, Nigeria 8 NGB00029 Ondo State, Nigeria 9 NGB00031 Oyo State, Niigeria 10 NGB00034 Oyo State, Niigeria 11 NGB00058 Ogun State, Nigeria 12 NGB00059 Ondo State, Nigeria 13 NGB00060 Oyo State, Niigeria 14 NGB00061 Osun State, Nigeria 15 NGB00070 Ogun State, Nigeria 16 NGB00078 Oyo State, Niigeria 17 NGB00082 Oyo State, Niigeria 18 NGB00108 Oyo State, Niigeria 19 NGB00111 Oyo State, Niigeria 20 NGB00112 Oyo State, Niigeria 21 LASPO­COL­001 Lagis State, Nigeria Oyetunde et al. ‐ GYT biplot technique revealed trait profiles of okra accessions 279 The land was initially tilled mechanically by ploughing twice and then harrowing. Subsequently, raised beds; 2×1 m in diameter, were made manually. Cured poultry manure was applied on the bed at a rate of 10 tons/ha. The trial was laid out in randomized Complete Block Design with two replications. Each accession was grown in a 2­row plot fitted into a bed in each replicate. A seed rate of 1.5 kg/ha was used, and planted by drilling. Seeds were mixed with dry fine sand to enhance even distribution within the drills. The experiment was exclusively rain­fed. Weeds were controlled manually by rogueing, sub­ ject to field inspection, and chemical insecticides were not used throughout the experiment. At maturity, a random sample of five plants per row; making 10 plants per replicate per accession, were observed for plant height (PHT) (cm), number of leaves per plant (NOL), and stem girth (STG) (mm). All the plants in each plot were uprooted, and the roots washed carefully, to record data on total bio­ mass (TBM) (g), root weight (RWT) (g), stem weight (SWT), and harvest index (%). Observations were recorded according to the Amaranthus descriptors of IPGRI (1999). Data on TBM, RWT, and SWT were con­ verted to kg/ha. Harvest index was computed as the percent of the total biomass that is made up by the stem weight. Data analyses Test for homogeneity of variance was not signifi­ cant and thus analysis of variance (ANOVA) was per­ formed on combined data from the two­year evalua­ tion using ‘proc glm’ in SAS (SAS, 2011). Means of data collected for each accession were subjected to correlation (Pearson coefficients) analysis among all pair­wise combinations of measured traits while the standardized mean values were subjected to princi­ pal component analysis. To reveal the level of pheno­ typic divergence among pairs of accessions, esti­ mates of genetic distance (Euclidean) among all pos­ sible pairs of accessions were obtained using ‘proc distance’ while the accessions were distributed into clusters from dendrogram obtained using ‘proc tree’ (SAS, 2011). To investigate the trait profiles of the accessions, a accession × yield­trait combination (GYT) (Yan and Fregeau­Reid, 2018) biplot approach was employed where stem weight was taken as the yield. The GYT biplots were obtained using the GGEBiplotGUI package in R. To select accessions for the GYT biplots, a superiority index (SI) was comput­ ed for the accessions, based on standardized GYT estimates. The SI value of a accession was computed as the arithmetic mean of its standardized estimates. Ten accessions comprising seven and three acces­ sions with the highest and lowest SI values respec­ tively were selected for the GYT biplot. 3. Results Results of analysis of variance of amaranth acces­ sions are summarized in Table 2. Mean squares of accession were significant (P≤0.05/0.01) for all the measured traits except stem girth, while mean squares of year were significant (P≤0.05/0.01) for number of leaves per plant, root biomass, and har­ vest index. Accession × year mean squares were also found to be significant (P≤0.05/0.01) for plant height, number of leaves per plant, and harvest index. The first three principal components (PCs) jointly accounted for 88.18% of the total variation among accessions, with PCs 1, 2, and 3 having Eigen values of 2.64, 1.39, and 1.25, respectively, and explaining 44.06, 23.21, and 20.91% of the total variation in that order (Table 3). Plant height, root weight, and num­ ber of leaves per plant had high loadings (≥0.30) of Source of variation DF PHT NOL STG TBM RWT SWT HI Rep (Year) 2 4.160 7.043** 0.010 0.039 0.005 0.023 0.002 Year 1 0.481 7.346* 0.004 0.007 0.036** 0.061 0.264** Accession 20 10.989** 9.745** 0.014 0.057** 0.007** 0.031* 0.015** Accession × Year 20 5.800** 6.517** 0.011 0.019 0.004 0.011 0.022** Error 40 2.110 0.332 0.008 0.023 0.049 0.015 0.005 Table 2 ­ Mean squares of measured traits of Amaranthus evaluated in 2018 and 2019 Rep = Replicate; DF = degrees of freedom; PHT = Plant height; NOL = Number of leaves per plant; STG = Stem girth; TBM = Total biomass; RWT = Root weight; SWT = Stem weight; HI = Harvest index. * and ** significant at 5 and 1% probabilities respectively. Adv. Hort. Sci., 2021 35(3): 277­284 280 0.55, 0.50, and 0.44 respectively in PC 1. High load­ ings in PC 2 were 0.69 and ­0.66 observed for harvest index and stem weight respectively while PC 3 was characterized by stem girth, root weight, number of leaves per plant, harvest index, and stem weight with loadings of 0.57, ­0.49, 0.40, 0.37, and ­0.35, respec­ tively. The PCA biplot grouped the amaranthus acces­ sions into clusters over the four quadrants based on the contributions of the measured traits as explained by the PCs (Figs. 1 and 2). The accessions were scat­ tered on the score biplot with the local check dis­ tinctly placed. Accessions NGB00022, NGB00031, NGB00060, and NGB00070 in the top left quadrant were associated with harvest index. The top right quadrant, associated with number of leaves per plant and stem and root weights, had NGB00005, NGB00019, NGB00025, NGB00058, NGB00061, and NGB00112 as the corresponding accessions. The bot­ tom right quadrant, characterized by plant height and stem girth, was composed by NGB00001, NGB00029, and the Local check. Other accessions, clustered in the bottom left quadrant, were not asso­ ciated with any of the traits measured in this study. Estimates of correlation coefficient revealed sig­ nificant (P≤0.05/0.01) associations among pairs of measured traits (Table 4). Positive and significant (P≤0.05/0.01) correlation was observed in the associ­ ation of plant height with number of leaves per plant (r= 0.52), stem girth (r= 0.27), total biomass (r= 0.29), and root weight (r= 0.47). Number of leaves per plant showed positive and significant (P≤0.05) correlation with stem girth, total biomass, root weight and stem weight with a correlation coefficient of 0.27, 0.24, 0.22 and 0.22, respectively. Furthermore, total bio­ mass and root weight had a significant positive corre­ lation (r= 0.70; P≤0.01) while there was positive and significant correlation in the association of stem weight with total biomass, root weight and harvest index (r= 0.93, 0.43, and 0.27, respectively). Finally, harvest index had negative and significant correla­ Table 3 ­ Loadings from principal component axes of the com­ post used in this study Table 4 ­ Pearson correlation coefficients among measured traits of 21 Amaranthus accessions evaluated in 2018 and 2019 Re* and ** significant at 5 and 1% probabilities respectively. Trait Principal compo­ nent axis 1 Principal compo­ nent axis 2 Principal compo­ nent axis 3 Plant height 0.55 ­0.14 0.09 Number of leaves per plant 0.44 0.21 0.40 Stem girth 0.25 ­0.14 0.57 Root weight 0.50 0.02 ­0.49 Stem weight 0.27 0.66 ­0.35 Harvest index ­0.23 0.69 0.37 Eigen value 2.64 1.39 1.25 Proportion (%) of variation 44.06 23.21 20.91 Cumulative (%) variation 44.06 67.27 88.18 Fig. 1 ­ Principal component loading pattern of six traits of Amaranthus accessions. Fig. 2 ­ Principal components score plot for 21 Amaranthus accessions. NOL STG TBM RWT SWT HI PHT 0.52** 0.27* 0.29** 0.47** 0.14 ­0.27* NOL 0.27* 0.24* 0.22* 0.22* 0.10 STG 0.16 0.14 0.15 ­0.02 TBM 0.70** 0.93** ­0.04 RWT 0.43** ­0.68** SWT 0.27* Oyetunde et al. ‐ GYT biplot technique revealed trait profiles of okra accessions 281 tions (P≤0.05/0.01) with plant height (r= ­0.27) and root weight (r= ­0.68). The estimates of genetic distance among pair­ wise combination of accessions based on composited traits (Table 5) ranged from 0.99 between NGB00028 and NGB00111 to 7.90 between NGB00078 and the local check. In particular, high genetic distance esti­ mates were obtained between the local check and NGB00022 (7.45), NGB00059 (7.72), NGB00060 (7.02), and NGB00082 (7.21). At a genetic dissimilarity of 0.50, the grouping of the accessions by the dendrogram (Fig. 3) fairly com­ pares with the groupings on the PC scores biplot. For instance, cluster 1 was composed of eight accessions, six of which were NGB00060, NGB00058, NGB00112, NGB00025, NGB00061, and NGB00019 which were grouped together, and associated with total biomass and stem and root weights on the PC scores biplot. The other two accessions in cluster I, NGB00031 and NGB00070, were also grouped together on the PC score plot, and were associated with harvest index. Cluster 2 contained NGB00005 and the local check. The two accessions were also grouped together by the PC score biplot, and associated with plant height, number of leaves per plant, and stem girth. Similarly, all the accessions in cluster 3; NGB00001, NGB00029, Fig. 3 ­ Dendrogram of relatedness among Amaranthus acces­ sions, (X­axis) based on genetic dissimilarity (Y­axis) from Single­Linkage Cluster analysis. The red double­arrowed line delineates the accessions into clusters at approxima­ tely 0.50 level of dissimilarity; C1, C2, C3 are clusters 1, 2, and 3, respectively. Table 5 ­ Pair­wise genetic distance estimates based on observed phenotypes of 21 Amaranthus accessions ID ACCESSION 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 1 NGB00001 0.00 3.62 3.86 2.52 1.82 1.37 2.13 1.65 2.34 1.62 2.73 3.81 3.91 4.02 2.27 4.23 2.58 1.84 2.47 2.44 5.33 2 NGB00005 0.00 3.68 4.89 2.26 4.36 4.50 3.88 3.80 4.80 3.39 6.27 3.94 3.50 3.50 5.66 5.32 4.77 4.73 3.11 3.81 3 NGB00019 0.00 4.14 2.97 4.24 4.54 4.93 2.82 4.71 4.00 5.09 4.45 1.54 3.70 6.57 5.08 5.21 4.83 2.61 6.33 4 NGB00022 0.00 3.19 1.42 1.19 2.97 1.69 2.14 2.23 3.79 3.01 3.88 1.68 3.28 1.54 2.06 1.26 2.26 7.45 5 NGB00025 0.00 2.67 3.07 2.70 2.15 3.06 2.35 5.10 3.10 3.20 2.45 4.98 3.75 3.19 3.17 1.63 4.79 6 NGB00027 0.00 1.00 1.80 2.12 1.35 2.37 3.38 3.58 4.14 1.70 3.26 1.41 1.15 1.45 2.46 6.39 7 NGB00028 0.00 2.10 2.21 1.85 2.01 3.55 3.17 4.17 1.31 2.47 1.12 1.25 0.99 2.52 6.80 8 NGB00029 0.00 3.34 2.52 2.55 4.16 3.96 4.94 2.52 3.54 2.28 1.56 2.27 3.06 5.07 9 NGB00031 0.00 2.53 2.18 4.31 2.69 2.57 1.63 4.26 2.93 2.94 2.41 1.33 6.76 10 NGB00034 0.00 3.32 3.54 4.27 4.56 2.46 3.67 2.31 1.52 2.35 3.25 6.79 11 NGB00058 0.00 5.06 1.54 3.69 1.51 3.41 2.62 2.65 1.75 1.48 6.10 12 NGB00059 0.00 6.32 4.96 4.01 4.94 3.43 3.87 4.34 4.80 7.72 13 NGB00060 0.00 4.03 2.58 4.15 3.78 3.81 2.72 2.15 7.02 14 NGB00061 0.00 3.10 5.78 4.93 5.05 4.60 2.61 6.74 154. NGB00070 0.00 2.90 2.32 2.29 1.81 1.78 6.36 16 NGB00078 0.00 2.88 2.87 2.73 4.50 7.90 17 NGB00082 0.00 1.40 1.18 3.12 7.21 18 NGB00108 0.00 1.44 3.18 6.50 19 NGB00111 0.00 2.48 6.99 20 NGB00112 0.00 6.00 21 Local check 0.00 282 Adv. Hort. Sci., 2021 35(3): 277­284 NGB00027, NGB00108, NGB00034, NGB00022, NGB00028, NGB00111, NGB00082, NGB00078, and NGB00059, except NGB00022 and NGB00001 were also grouped together on the PC scores biplot, and were not associated with any of the traits measured in this study. The trait profiles of the Amaranthus accessions are displayed on the polygon view of the GYT biplot (Fig. 4). The polygon view revealed four sectors with the NGB00019, local check, NGB00060, and NGB00078 as the vertex accessions in their respec­ tive sectors. Accessions NGB00019 and NGB00061 belonged to sector 1 which was characterized by all the measured yield­trait (Y­T) combinations. This sec­ tor contained the highest ranked accessions. Sector 2 with the local check as vertex accession, also con­ tained NGB00005 and NGB00025, thus comprising accessions that were next in rank to the accessions in sector 1. Similarly, the accessions in sector 3; NGB00060 and NGB00112, ranked next to those in sector 2 while the­poorest ranked accessions; NGB00082, NGB00059, and NGB00078, constituted sector 4. Sectors 2, 3, and 4 were not associated with any of the Y­T combinations considered in this study. The names of the accessions 1 to 10 are available in Table 6. Fig. 4 ­ The accession × yield­trait biplot of ‘which won where’ of selected seven best and 3 worst Amaranthus accessions. Table 6 ­ Genotype × yield­trait combination data matrix for 21 Amaranthus accessions evaluated in 2018 and 2019 YxPHT, YxNOL, YxSTG, YxTBM, YxRWT, YxHI = Yield combination with plant height, number of leaves per plant, stem girth, root weight, and harvest index respectively. ID on GYT biplot Accession YxPHT YxNOL YxSTG YxTBM YxRWT YxHI Superiority index 1 NGB00019 1.779 1.183 1.386 2.337 2.299 1.774 1.793 2 NGB00061 1.291 1.564 0.793 1.903 1.701 1.655 1.485 3 NGB00005 1.426 1.923 1.337 1.200 1.470 0.943 1.383 4 Local check 1.564 1.081 1.098 0.553 1.476 ­0.230 0.924 5 NGB00025 0.791 0.833 1.272 0.783 0.856 0.683 0.870 6 NGB00060 0.831 1.186 1.313 0.595 ­0.367 1.576 0.856 7 NGB00112 0.886 0.676 1.005 0.766 0.462 1.087 0.814 Not selected NGB00031 0.070 0.244 0.441 0.561 0.156 0.576 0.341 Not selected NGB00058 0.427 0.394 0.423 0.005 ­0.326 0.560 0.247 Not selected NGB00070 ­0.188 0.060 ­0.251 ­0.244 ­0.245 0.017 ­0.142 Not selected NGB00001 ­0.294 ­0.365 ­0.041 ­0.295 0.003 ­0.453 ­0.241 Not selected NGB00022 ­0.499 ­0.537 ­0.340 ­0.438 ­0.531 ­0.101 ­0.408 Not selected NGB00111 ­0.584 ­0.635 ­0.466 ­0.710 ­0.826 ­0.388 ­0.602 Not selected NGB00028 ­0.685 ­0.633 ­0.707 ­0.726 ­0.740 ­0.576 ­0.678 Not selected NGB00027 ­0.758 ­0.802 ­0.692 ­0.694 ­0.490 ­0.707 ­0.690 Not selected NGB00034 ­0.978 ­0.748 ­0.593 ­0.682 ­0.499 ­0.822 ­0.720 Not selected NGB00029 ­0.638 ­0.784 ­0.737 ­0.825 ­0.614 ­0.950 ­0.758 Not selected NGB00108 ­0.990 ­0.959 ­0.879 ­0.946 ­0.881 ­1.005 ­0.943 8 NGB00082 ­0.883 ­1.093 ­1.010 ­0.955 ­0.972 ­0.879 ­0.965 9 NGB00059 ­0.952 ­1.269 ­1.518 ­0.874 ­0.612 ­1.251 ­1.079 10 NGB00078 ­1.617 ­1.321 ­1.833 ­1.317 ­1.318 ­1.509 ­1.486 Mean 0 0 0 0 0 0 Standard deviation 1 1 1 1 1 1 Oyetunde et al. ‐ GYT biplot technique revealed trait profiles of okra accessions 283 4. Discussion and Conclusions The significant difference observed among the accessions indicated the existence of variation with respect to the measured traits except for stem girth, and underscored the possibility of selection for improvement. Idehen et al. (2018) reported same result for stem girth among 10 accessions of Amaranthus spp. On the contrary, Mandal and Dhangrah (2012) studied 17 Amaranthus accessions and observed significant differences among the accessions for all the characters considered including stem girth. Plant height, number of leaves per plant, root biomass and harvest index can be relied upon as important tools in long term selection gain. Gerrano et al. (2015) found high phenotypic variability among 32 Amaranthus accessions using plant height, leaf length, leaf width, leaf area, leaf area index, number of leaves, stalk diameter, panicle or inflorescence length, number of primary branches, fresh biomass, dry biomass, harvest index, thousand seed weight and grain yield per plant. The loadings of plant height, number of leaves per plant, and root weight indicated that these charac­ ters chiefly accounted for most of the variation observed among the accessions. This suggests that these traits are crucial in maintaining variability with­ in the breeding population and they should be con­ sidered for selection in Amaranthus improvement programme. Gerrano et al. (2015) reported compara­ ble results on 32 Amaranthus species of South African origin. The positive significant correlation observed between plant height and number of leaves per plant, stem girth, total biomass, root weight and harvest index implied that direct selection for any of the trait could lead to improvement in the other. Gerrano et al. (2015) reported that plant height correlated positively with fresh biomass and dry biomass when evaluating the genetic diversity of Amaranthus species in South Africa. Thanapornpoonpong et al. (2007) also reported a significant and positive relationship between plant height and fresh biomass. Strong positive correlation of stem girth, total and root biomass and stem weight with number of leaves per plant, total biomass and root biomass, stem weight with total biomass, root biomass and harvest index indicated that the use of any of the characters can help to improve selection process in breeding programs. The significant negative correlation of harvest index with plant height and root weight suggested that an attempt to breed for high harvest index will lead to short plant height and less root weight in Amaranthus. The clustering pattern of accessions into groups showed the phenotypic diversity among the acces­ sions for the different characters studied. Genetic diversity, evidenced by phenotypic variability, is essential in the initiation of a breeding program because when absent, there cannot be meaningful selection and genetic advancement becomes impos­ sible (Govindaraj et al., 2014). The accessions within a cluster are closely associated and this suggests that the variability within group could be useful in the selection process for improvement of associated desired traits. Variability between groups could be explored in heterotic breeding where members of a cluster could serve as parents in crosses involving members from distinct clusters. For instance, acces­ sions in Cluster 1 were mostly associated with har­ vest index, number of leaves per plant, and stem and root weights suggesting the presence of favourable alleles for yield­related traits within the group and a potential to improve these traits via selection. Whereas, individuals in Cluster 2 were generally associated with plant height and stem girth. A cross between members of the two groups could give high­ er yielding progenies with good standability which would be invaluable to the vegetable industry. The GYT biplot is useful for multiple­trait­based evaluation of accessions, permitting a graphical rank­ ing of entries based on their levels in combining yield with related traits. Like in the GGE biplot, the GYT polygon is delineated into sectors, with their associ­ ated closely­related Y­T combinations, a vertex and other associated accessions (Yan and Frégeau­Reid, 2018). Thus, NGB00019 and NGB00061 were associ­ ated with all the Y­T combinations considered in this study which is an indication of superior trait profile. However, NGB00019 was the vertex accession and was thus identified as the best combiner of yield with other traits. There was sufficient genetic variability among the amaranthus accessions, to permit improvement through selection, with greater chances of success with plant height, number of leaves per plant, and root weight. Harvest index; which is the edible por­ tion of the plant, can be simultaneously improved with plant height, and stem and root weights. Information from principal component analysis, genetic distance estimates, and cluster analysis can Adv. Hort. Sci., 2021 35(3): 277­284 284 be utilized for parental selection in breeding pro­ grammes. Accessions NGB00019, NGB00025, NGB00058, NGB00060, NGB00061, and NGB00112 would be good genetic materials for total biomass, and root and stem weights while NGB00031 and NGB00070 have potentials to improve harvest index, and NGB00005 and the local check would be useful for plant height, number of leaves per plant, and stem girth. NGB00019 followed by NGB00061 had the most diverse trait profile among the evaluated accessions. Further studies involving more traits are required to determine the trait profiles of the remaining accessions used in the study. Acknowledgements We are grateful to the National Center for Genetic Resources and Biotechnology, NACGRAB, Ibadan, Nigeria, for providing the seeds used for the study. We thank the staff of the Teaching and Research Farms of the Department of Crop Production and Horticulture, Lagos State Polytechnic, Ikorodu, Nigeria, for assistance during field evaluation. References AKIN­IDOWU P.E., GBADEGESIN A.M., UTERDZUA O., IBITOYE D.O., ODUNOLA O.A., 2016 ­ Characterization of grain amaranth (Amaranthus spp.) germplasm in South West Nigeria using morphological, nutritional, and random amplified polymorphic DNA (RAPD) analy‐ sis. ­ Resources, 5(1): 6­15. EBERT A.W., WU T., WANG S., 2011 ­ Vegetable amaranth (Amaranthus L.). ­ AVRDC, The World Vegetable Center, AVRDC Publication no. 11­754, pp. 9. GERRANO A.S., VAN RENSBURG W.S.J., ADEBOLA P.O., 2015 ­ Genetic diversity of Amaranthus species in South Africa. ­ South Afr. J. Plant Soil, 32(1): 39­46. GOVINDARAJ M., VETRIVENTHAN M., SRINIVASAN M., 2014 ­ Importance of genetic diversity assessment in crop plants and its recent advances: An overview of its analytical perspectives. ­ Genetics Res. Int., pp. 1­14. IDEHEN E.O., ODUWAYE O.A., LATEEF L.A., IKEORA C.J., 2018 ­ Evaluation of genetic diversity in some Amaranthus spp. using morphological and Random Amplified Polymorphic DNA (RAPD) analysis. ­ Nig. J. Biotech., 35(2): 167­175. IPGRI, 1999 ­ Directory of germplasm collections. ­ IPGRI, International Plant Genetic Resources Institute, Rome, Italy. JANOVSKÁ D., CEPKOVA P.H., DZUNKOVA M., 2012 ­ Characterisation of the Amaranth genetic resources in the Czech Gene Bank, pp. 457­478. ­ In: CALISKAN M. (ed.) Genetic diversity in plants. InTech, Rijeka, Croatia pp. 498. KATIYAR R.S., SHUKLA S., RAI S., 2000 ­ Varietal perfor‐ mance of grain amaranths (A. hypochondriacus) on sodic soil. ­ Proc. National Academy of Sciences, India, Section B: Biological Sci., 70: 185­187. MANDAL J., DHANGRAH V.K., 2012 ­ Screening vegetable amaranth under summer condition in Red and Lateritic Belt of West Bengal. ­ Environment & Ecology, 30(4): 1430­1433. NEELESH K.M., PRATIBHA A., 2018 ­ Amaranthus grain nutritional benefits: A review. ­ J. Pharmacognosy Phytochem., 7(2): 2258­2262. SAS, 2011 ­ SAS System for Windows. ­ Release 9.4. SAS Institute Inc. Cary, NC, USA. SCHIPPERS R.R., 2000 ­ African indigenous vegetables: an overview of the cultivated species. ­ Natural Resources Institute, ACP­EU Technical Centre for Agricultural and Rural Cooperation, Chatham, UK. SHUKLA S., BHARGAVA A., CHATTERJEE A., PANDEY A.C., MISHRA B.K., 2010 ­ Diversity in phenotypic and nutri‐ tional traits in vegetable amaranth (Amaranthus tricol­ or), a nutritionally underutilised crop. ­ J. Sci. Food Agric., 90: 139­144. STALLKNECHT G.F., SCHULZ­SCHAEFFER J.R., 1993 ­ Amaranth rediscovered. New Crops. ­ Wiley, New York, USA, pp. 211­218. THANAPORNPOONPONG S., SOMSAK W., PAWELZIK E., VEARASILP S., 2007 ­ Yield component of amaranth (Amaranthus spp.) grown under northern Thailand irri‐ gated area. ­ Paper presented at Tropentag 2007, University of Kassel­Witzenhausen and University of Göttingen, 9­11 October 2007. YAN W., FRÉGEAU­REID J., 2018 ­ Genotype by yield x trait (GYT) biplot: a novel approach for genotype selection based on multiple traits. ­ Scientific Report, 8: 8242.