Introduction

Pigeonpea (Cajunus cajan) is a diploid (2n = 2x = 22) legume (Maesen 1990) grown in tropical and subtropical regions of the world. It is underutilised crop species, despite its contribution to food and nutritional security (Lin-Qi 2014). The crop improves the fertility of the soil through atmospheric nitrogen fixation and increase production and productivity of the crop (Adebowale and Maliki 2011; Choudhary et al. 2013; Saidia et al. 2019). Pigeonpea can be intercropped with cereal-based cropping system (Lin-Qi 2014). It is sources of macro- and micronutrients, vitamins and phytochemical compositions (Saxena et al. 2010; Gerrano et al. 2022). It is considered as climate smart crop in tropical and sub-tropical regions. It has the ability to withstand drought and give good economic benefits when planted under dryland farming conditions and sustain the livelihood of resource poor rural populations in tropical and sub-tropical regions of the African continent including South Africa. Furthermore, the crop helps in protecting the environment from soil erosion and degradation, improve the fertility of the soil, increase crop production and productivity at marginal crop lands towards soil and food security strategies. It is climate smart crop that adapt to the current climate change, which is tolerant to heat, drought, diseases and insect pests (Odeny 2007).

The seed of the crop can be eaten as a green vegetable and dry pulse and is an important source of nutritional components (Faris et al. 1987; Choudhary et al. 2013). The green pods and foliage of the plant can be used as animal feed (Mallikarjuna et al. 2011). The crop is cultivated by the resource poor small scale farmers with the low input agriculture in South Africa. Hence, identification of potential candidate genotypes and development of improved cultivars for increased production and productivity of the crop is important.

For an efficient evaluation and utilisation of the plant genetic resources, understanding and knowledge of genetic diversity, genetic background information, collection and classification are important and the basis for crop improvement programs (Khan et al. 2014; Syafii et al. 2015), which is elucidated through different marker systems such as agro-morphological, biochemical and molecular markers. Among these, morphological characterisation is considered as the initial step for designing breeding programs (Smith and Smith 1989; Khan et al. 2014), it is influenced by the growing environmental condition unlike that of DNA-based markers. Yohane et al. (2020) reported the existence of widest variability among test pigeonpea based on their morphological performance in Malawi. Assessing genetic variability helps to study heterosis (Virk et al. 2003), selection of transgressive breeding segregants and genes of novelty, and has a role in collection and maintenance of germplasm for crop improvement (Duran et al. 2009) in the gene bank for future use.

Potential parental lines were identified for improvement (Malik et al. 2014; Syed et al. 2019) through using multivariate statistical analysis. It is effective statistical tool for studying the differences and similarities between and within the genotypes (Peeters and Martinelli 1989; Kovacic 1994; Rachovska et al. 2002; Mondal 2003; Mohammadi and Prasanna 2003; Ajmal et al. 2013; Immad et al. 2018), which can help to generate new breeding population in the breeding programme. The knowledge and understanding of the crop species and their respective descriptors are necessary for informed breeding strategy (Upadhyaya et al. 2007; Gbaguidi et al. 2018). The assessment of phenotypic plasticity using morphological characters in pigeonpea is importance in order to determine the existing variability in the population which will finally enable the identification and selection of potential and superior lines of the genotypes for production and breeding. Therefore, the objective of the current study was to assess the variability and their interrelationship in pigeonpea genotypes using morphological traits.

Materials and Methods

Plant material and trial sites

The 19 pigeonpea genotypes were evaluated on the field during the summer season of 2019–2020. The origins of the pigeonpea genotypes used in the study are listed in Table 1.

The experimental trials were conducted at the North West University research farm at Mafikeng (25°48ʹ S, 45°38ʹ E; 1012 m. a. s. l.) in North West Province and the Agricultural Research Council – Tropical and Subtropical research station in Nelspruit (25.49°89’ S, 31.35°37’ E; 670 m. a. s. l.) in Mpumalanga Provinces during 2019/2020 cropping season in South Africa. Pigeonpea is widely grown predominantly in this two Province in South Africa and have extreme variations in agro-climatic conditions. The soils on the North West University farm belongs to the Hutton series, with sandy loam and a yellow sand alternating (Molope 1987; Kasirivu et al. 2011), while the Nelspruit research station field consisted of sandy loam soil. During the season, Mafikeng received a summer total rainfall, with a mean of 571 mm during the cropping season. The mean maximum temperature is 37°C, while the mean minimum temperature is 9°C during this cropping season. The field in Nelspruit is characterised by mean maximum temperature of 28°C. The mean minimum temperature is12.5°C with a precipitation of about 796 mm during the cropping season in 2019/2020.

Trial design and management

The trials were laid out in a randomized complete block design replicated three times with a plot consisting of two rows of 4 m length. The inter- and intra-row spacing’s were 90 and 60 cm, respectively. The experiment was conducted during summer cropping season in rainfed condition based on the farmers practice. Weeding was done manually. No fertilizer was applied to simulate low input cropping system in the region (Gerrano et al. 2015).

Data collection

Data were recorded according to standard descriptor list for pigeonpea (IBPGR 1994). Data were recorded from three randomly selected plants in the middle of each row per replications. The qualitative data recorded included base flower colour, second flower colour, vigour at 50% flowering, pod form, seed colour pattern, seed shape, and pattern of streaks. The list of quantitative traits studied and their data collection method is presented in Table 2.

Statistical data analysis

The recorded quantitative data were analysed using analysis of variance (ANOVA), principal component analysis (PCA), and Pearson correlations. The qualitative data were analysed using frequencies, spearman correlations, and Shannon Weaver Table 1: Pigeonpea germplasm used in the study

Entry number	Genotype Name	Origin/source
1	ICEAP 01147	Kenya
2	ICEAP 01154-2	Kenya
3	ICEAP 01150-1	Kenya
4	ICEAP 01179	Kenya
5	ICEAP 00979-1	Kenya
6	ICEAP 01172-2-4	Kenya
7	ICEAP 01159	Kenya
8	ICEAP 01544-2	Kenya
9	ICEAP 00540	Kenya
10	ICEAP 00554	Kenya
11	ICEAP 00557	Kenya
12	ICEAP 00850	Kenya
13	Ilonga 14-M1	Tanzania
14	Mali	Tanzania
15	Ilonga 14-M2	Tanzania
16	Karatu-1	Tanzania
17	Kiboko	Tanzania
18	Komboa	Tanzania
19	Tumia	Tanzania

Table 2: A list of economical traits measured, abbreviations and definitions

Trait	Abbreviation	Measurement/definitions
Plant height (cm)	PHT	Height of a plant from the base of the stem to the tip of the plant at harvest
Days to 50% flowering	DFF	Number of days from planting until 50% of the plants have flowered in a plot
Pod bearing (cm)	PDB	Distance from lowest to the top most of the plant
Leaf length mm)	LFL	Length from the tip of the leaf to the leaf petiole
Leaf width (mm)	LFW	Length in the middle of the leaf from one tip to the other tip
Branch number	BRN	Number of branches per plant
Stem diameter (cm)	STD	Diameter of plant stem
Pod length (mm)	PDL	Length of the pod from bottom end to top end at harvest
Pod width (mm)	PDW	Length at the centre of the pod from one end to the other end/diameter
100 Seed weight (g)	HSW	Weight of 100 seed picked randomly for each genotype
Pod weight (g)	PWT	Weight of dry pods harvested from each genotype
Seed number per pod	SNT	Number of seeds in a pod (average of 10 pods)
Seed number per plant	SNP	A number of seeds produced by a single plant.
Seed yield (g)	SYD	Weight of seeds produced per plant

DFF = Days to 50% flowering; PHT = Plant height; BRN = Branch number; LLT = Leaf length; LWT = Leaf width; PDB = Pod bearing; 100SW = hundred seed weight; PDL= Pod length; PDW = Pod width; SNP = Seed number per pod; PWT = Pod weight; SEP = Seed number per plant; STD = Stem diameter; SYD = seed yield per plant

diversity index. The biplots were generated using principal coordinate analysis in SAS version 9.6 (SAS Institute 2021). A dendrogram was constructed using Genstat 18^th edition (VSN International, Hempstead, UK) (2020).

Results

Genetic variability and genotype by environment interaction

The univariate analysis of variance (ANOVA) result depicted that there were highly significant (P ≤ 0.01) genetic variability observed among the tested genotypes for pod length (PDL) and pod weight (PWT) (P ≤ 0.01) (Table 3). Furthermore, the ANOVA mean squares showed genotype (G), site (S), and genotype × site interaction (GEI) effects on quantitative traits is presented in Table 3. Furthermore, the agronomic characteristics of the pigeonpea genotypes showed variability in different environments due to the significant genotype × environment interaction (Table 3) for plant height (PHT) (P ≤ 0.01), pod bearing (PDB) (P ≤ 0.05) and seed number per pod (SNP) (P ≤ 0.05). There were significant (P ≤ 0.05) differences between sites based on days to flowering (DFF), plant height, branch number (BRN), stem diameter (STD), pod bearing, pod length, pod weight and significant differences for seed number per pod. There was a significant (P ≤ 0.05) site × genotype interaction effect based on plant height, pod bearing and seed number per pod (Table 3).

Pearson correlation analysis

Pearson’s correlations (r) of 14 quantitative traits measured in the study are shown in Table 4. Days to flowering was significantly and positively correlated with plant height, branch number, stem diameter, and hundred seed weight. Similarly, Table 3: Combined analysis of variance for quantitative traits among the studied pigeonpea genotypes in terms of means squares

SOV	d.f	DFF	PHT	BRN	STD	LLT	LWT	PDB	100 SW	PDL	PDW	SNP	PWT	SEP	SYD
Site (S)	1	85323.9**	175005.4**	1141.9**	11556.7**	0.2	722.6	41198.2**	83.4	1957.3**	10.9	138.3*	142.2**	1.3	45.9
Genotype(G)	18	76.2	1908.7	15.6	13.5	1.9	327.2	1417.3	163.3	507.8**	3.7	57.0	46.0**	8.0	20.0
S × G	18	72.9	2867.6**	16.7	14.7	2.1	325.4	2733.8*	165.9	322.2	4.0	61.2*	30.8	8.9	17.9

SOV = Sources of variation; d.f. = degree of freedom; DFF = Days to 50% flowering; PHT = plant height; BRN = Branch number; LLT = Leaf length; LWT = Leaf width; PDB = Pod bearing; 100SW = hundred seed weight; PDL= Pod length; PDW = Pod width; SNP = Seed number per pod; PWT = Pod weight; SEP = Seed number per plant; STD = Stem diameter; SYD = seed yield per plant; * = significantly different from zero at P ≤ 0.05; ** = significantly different from zero at P ≤ 0.01

Table 4: Pearson correlations for quantitative traits among the studied pigeonpea genotypes

Variable	DFF	PHT	BRN	STD	LLT	LWT	PDB	100 SW	PDL	PDW	SNP	PWT	SEP
DFF	1.00
PHT	0.701***	1.00
BRN	0.625***	0.751***	1.00
STD	-0.900***	-0.667***	-0.492***	1.00
LLT	-0.089	0.017	0.040	0.241	1.00
LWT	-0.034	-0.019	0.075	0.169	0.672***	1.00
PDB	-0.498***	-0.405***	-0.341***	0.504	0.190*	-0.056	1.00
100SW	0.525***	0.431***	0.296**	-0.574**	-0.053	-0.003	-0.353***	1.00
PDL	0.183	0.046	-0.011	-0.136	0.117	0.095	-0.010	0.159	1.00
PDW	0.063	0.083	0.114	-0.024	0.003	-0.110	0.014	-0.060	0.018	1.00
SNP	0.086	0.133	0.089	-0.076	-0.085	-0.202*	0.020	-0.063	0.436***	0.135	1.00
PWT	0.189*	0.068	0.013	-0.139	0.102	0.055	-0.006	0.135	0.986***	0.161	0.526***	1.00
SEP	0.183	0.107	0.064	-0.130	0.060	-0.037	0.005	0.072	0.858***	0.453***	0.669***	0.932***	1.00
SYD	0.183	0.096	0.042	-0.136	0.065	-0.013	0.001	0.092	0.928***	0.248**	0.694***	0.974***	0.976***

DFF = Days to 50% flowering, PHT = plant height, BRN = Branch number, LLT = Leaf length, LWT = Leaf width, PDB = Pod bearing, 100 SW = hundred seed weight, PDL = Pod length, PDW = Pod width, SNP Seed number per pod, PWT = Pod weight, SEP = Seed number per plant, STD = Stem diameter, SYD = seed weight per plant.

Bold value represent significant association * = P < 0.05, ** = P < 0.01, *** = P < 0.001

Table 5: Factor loadings of the most important PCs for agro-morphological traits among the studied pigeonpea genotypes

Traits	PC1	PC2	PC3	PC4	PC5
DFF	0.57	-0.72	0.05	-0.02	0.01
PHT	0.47	-0.70	0.12	0.24	-0.25
BRN	0.37	-0.64	0.19	0.37	-0.31
STD	-0.52	0.74	0.12	0.13	-0.07
LLT	-0.00	0.21	0.87	0.18	-0.04
LLW	-0.04	0.08	0.91	-0.02	0.07
PDB	-0.27	0.58	-0.01	0.17	-0.22
HSW	0.37	-0.54	0.09	-0.35	0.32
PDL	0.82	0.42	0.13	-0.30	0.03
PDW	0.27	0.11	-0.15	0.78	0.53
SNP	0.612	0.31	-0.26	0.15	-0.45
PWT	0.87	0.44	0.07	-0.17	0.06
SEP	0.88	0.44	-0.05	0.14	0.10
SYD	0.89	0.45	-0.02	-0.07	-0.02
Eigenvalue	4.62	3.50	1.78	1.163	0.82
Explained variance (%)	32.97	24.96	12.694	8.307	5.87
Cumulative variance (%)	32.968	57.931	70.625	78.932	84.80

DFF = Days to 50% flowering, PHT = plant height, BRN = Branch number, LLT = Leaf length, LWT = Leaf width, PDB = Pod bearing, 100 SW = hundred seed weight, PDL= Pod length, PDW = Pod width, SNP = Seed number per pod, PWT = Pod weight, SEP = Seed number per plant, STD = Stem diameter, SYD =– seed weight per plant

days to flowering was significantly and positively correlated with pod weight and negatively correlated with pod bearing. Plant height was highly significant and positively correlated with branch number per plant, stem diameter, and hundred seed weight, and negatively associated with pod bearing. Branch number had a negative and significant association with stem diameter and pod bearing, and a positive correlation with hundred seed weight. Stem diameter had a significant and positive correlation with leaf length, whereas pod bearing showed a negative association with hundred seed weight. Leaf length showed positive and significant correlations with leaf width and pod bearing. Leaf width had a negative association with seed number per pod. Pod bearing had a highly significant negative correlation with hundred seed weight. Pod length showed a positive association with seed number per pod, pod weight, seed number per plant, seed yield. Pod width showed a positive and highly significant correlation with seed number per plant and seed yield. Seed number per pod was positively correlated with pod weight, seed number per plant, and seed yield. Pod weight had positive correlations with seed number per plant and seed yield. Seed number per plant was highly significant and positively correlated with seed yield (Table 4).

Fig. 1: PC biplot for quantitative traits among the studied pigeonpea genotypes. PC1=first principal component; PC2=second principal component

Fig. 2: PCA biplot for qualitative traits among the studied pigeonpea genotypes.

Principal component analysis

Five most important PCs were identified contributing 32.9, 24.9, 12.7, 8.3 and 5.9%, to the total variation of 84.7%, respectively (Table 5). The first PC had pod length, pod weight, seed number per plant and seed yield contributing to this variation. In the second PC, days to flowering, plant height, branch number, stem diameter contributed the most to variation. Leaf length and leaf width contributed the most variation in third PC. In the fourth PC, pod width was the most contributors to variation whereas in the fifth PC, pod width and seed number per pod was the traits that contributed the most variation.

Principal coordinate analysis

The principal component (PC) biplot of the quantitative traits showing grouping of genotypes superimposed with traits is presented in Fig. 1, PC1 had 31.35% and PC2 had 0.26% variances with the total contributing variation of 51.61%. Gerrano et al. (2022) reported that the angles lesser than 45^o between the vector lines of the two respective variables indicate positive and high trait correlation and revealed the ability to discriminate the test genotypes for breeding. Genotypes ICEAPO1150-1, ICEAPO1154-2, ilonga14-M2, ICEAPO1172-2-4, ICEAPO1544-2, Mali, ICEAPO4459 and longa14-M1 were grouped Table 6: Frequency percentages of qualitative traits for medium duration pigeonpea

Trait	Score	Frequency (%)	Cumulative frequency (%)	H’
Vigour at 50% flowering	Low	5.36	5.36	0.99
	Intermediate	23.21	28.57
	High	71.43	100
Base flower colour	Light yellow	19.65	19.65	0.97
	Yellow	51.78	71.43
	Orange-yellow	28.57	100
Second flower colour	Red	71.43	71.43	0.96
Second flower colour	Purple	28.57	100
Pattern of streaks	Sparse	35.09	35.09	0.97
	Medium amount	15.79	50.88
	Dense	22.81	73.68
	Uniform coverage of second color	26.32	100
Flowering pattern	Determinate	100	100	1.00
Stem Thickness rating	Thick (>13 mm)	100	100	1.00
Growth habit	Erect and compact	22.81	22.81	0.98
	Semi spreading	1.75	24.56
	Spreading	75.44	100
Stem color	Green	63.16	63.16	0.98
Stem color	Sun Red	36.84	100
Pod form	Flat	3.64	3.64	0.99
Pod form	Cylindrical	96.36	100
Seed color pattern	Plain	3.57	3.57	0.99
	Mottled	7.14	10.71
	Speckled	71.43	82.14
	Mottled and speckled	17.86	100
Seed shape	Oval	21.43	21.43	0.98
	Globular	64.29	85.71
	Square	14.29	100

H’ = Shannon Weaver Index

together based on high SYD, HSW, SNT, PWT, PDW and PDL. Further, ICEAPO1179 and Tumia were identified as best genotypes for BRN, PHT and SNP. The genotype ICEAP01147, Kiboko, and ICEAP00850 were associated with the variables LLF, DFF, and LLW, while the genotypes ICEAP00557, Karatu-1, ICEAP00554 and Komboa revealed less association to the variables recorded indicating that the genotypes were less responsive to the variables. Genotype ICEAP00850 was associated to PDB. Genotype ICEAP00540 is peculiar genotypes that was found far from the rest of genotypes from the scatter biplot (Fig. 1), which can be considered for further evaluation in the breeding program. Stem diameter and pod bearing were negatively correlated with plant height, branch number, seed yield, and 100 seed weight. Seed number per pod, pod length, pod width, pod weight, seed yield, and seed number per plant were positively correlated with hundred seed weight, while branch number and plant height were highly positively correlated. The same traits were also correlated with stem diameter, pod bearing, leaf width and leaf length.

The biplot for the qualitative traits, the PC1 showed 40.27% and F2 had 26.41% (Fig. 2). The first quadrant showed base flower colour and vigour at 50% flowering, which are positively correlated in this quadrant and are associated with the genotypes Ilonga 144-M1, ICEAP 00850, and ICEAP 01159, while the second quadrant showed seed shape that was associated with the genotypes positioned in this quadrant. The third quadrant had pod form and seed colour pattern that are positively correlated to each other. The genotypes Kiboko and Mali had similar pod form and seed colour pattern in this quadrant. The 4th quadrant consists of only second flower colour. All the genotypes scattered in this quadrant were grouped together based on this trait (Fig. 2). In Fig. 2, the genotypes that are circled have similar values for PC1 and PC2 scores, which made them to be positioned on one dot.

Frequencies of qualitative traits

The frequencies of eleven qualitative traits measured are shown in Table 6. Vigorousness at flowering was high with 71.4% of plants being vigorous and intermediate was 23.2%. The base flower colour was dominated by yellow flowers followed by orange-yellow. The second flower colour was predominantly composed of red flowers (71.4%). The pattern of streaks was dominated by sparse streaks (35.1%), followed by uniform coverage of second colour and dense streaks. All plants of various genotypes had 100% stems thicker than 13 mm with green stems dominating (63.2%). The growth habit of the crop was predominantly composed of spreading types (75.4%) followed by erect and compact at 22.8%. The genotypes were dominated by cylindrical pods 96.40 with speckled seed colour pattern at 71.4% Chart, box and whisker chart

Description automatically generated

Fig. 3: A dendrogram showing interrelationships and divergence among nineteen pigeonpea genotypes based on quantitative traits

followed by mottled and speckled at 17.9%. The shape of the seed was predominantly globular (64.3%) with oval shape being 21.4%.

Shannon weaver diversity

Shannon weaver diversity indices are shown in Table 6. The diversity indices range from 0.96 (second flower colour) to 1.00 (flowering pattern and stem thickness). All traits showed significant variation except for flowering pattern and stem thickness.

Hierarchical clustering

A dendrogram was constructed using hierarchical clustering to present differences and interrelationships among the studied pigeonpea genotypes (Fig. 3). The dendrogram grouped genotypes into three clusters and four singletons. The first cluster was composed of six genotype, Longa14-M1, Mali, ICEAP00557, Ilonga14-M2, ICEAP01159, and ICEAP00850. The second cluster was composed of four genotypes, ICEAP0050-1, ICEAP01179, kmboa and ICEAP01147. The genotypes, Kiboko, ICEAP01154-2, Tumia, ICEAP01172-2-4 and ICEAP01544-2 were grouped in third cluster. Four genotypes were identified as most divergent and grouped as singletons (ICEAP00554, ICEAP000979-1, ICEAP00540, and Karatu-1). These genotypes were far and distantly related with the rest of the test genotypes.

Discussion

The significant difference observed among the pigeonpea genotypes showed the existence of genetic variation with respect to the measured morphological characteristics. The knowledge of morphological variation for a trait and trait correlations are important components of any breeding objective. There were highly significant differences for sites based on days to flowering, plant height, branch number, stem diameter, pod bearing, pod length, pod weight and significant differences for seed number per pod (Table 3). This indicates that the expression of the significant traits varied with the growing environmental conditions they were tested in. Their performance was not stable across sites. The presence of highly significant differences in genotypes based on pod length and pod weight highlights the presence of genotypic variation among the genotypes evaluated based on the two traits which can be exploited for cultivar improvement in future breeding programmes. The significant differences on genotype x site interaction could be attributed to the different reactions of the genotypes to sites or due to differences between the sites. In each environment, phenotypic manifestation is the result of the action of the genotype under the influence of the environment. However, when considering a series of growing environments, in addition to the genetic and environmental effects, an additional effect can be detected from their interaction (Marais et al. 2013; Nunes et al. 2014). Significant genotype × environment interaction on yield and yield components in this study concur with the results reported previously (Vales et al. 2012; Kimaro 2016; Gerrano et al. 2020).

The positive and significant correlation observed among the quantitative characters indicated that direct selection for any of these traits could lead to simultaneous improvement in the other characters of pigeonpea for increased production and productivity. Yohane et al. (2020) reported the existence of positive correlations for most of secondary traits that revealed multiple trait identification and selection for simultaneous trait improvement, while the weak correlations among the traits would result in an inefficient selection or low genetic gains that will take long time to fix the traits of interest. In this study seed yield was positively correlated with seed number per pod, seed number per plant and pod weight whereas pod bearing was negatively associated with hundred seed weight. The positive and significant correlations in the current study indicated the importance of simultaneous improvement for the traits of interest in the crop (Sodavadiya et al. 2009; Linge et al. 2010; Prasad et al. 2013; Saroj et al. 2013; Ojwang et al. 2016; Kinhoégbè et al. 2020; Yohane et al. 2020).

The Principal component analysis over sites revealed five most important PCs with pod length, pod weight, seed number per plant, seed yield, leaf length, leaf width, days to flowering, plant height, and stem diameter being the most contributing traits to the total variation observed. This suggests that these traits are useful for selection. Other reports have indicated that trait contribution to different PCs varies with genetic diversity within the tested germplasm and the number of traits evaluated (Upadhyaya et al. 2007). The biplot also showed the different grouping of pigeonpea genotypes based on specific traits. These findings suggested that both agro-morphological traits revealed variability among the tested genotypes but complementary information for breeding.

The most of pigeonpea genotypes in the current study showed a tendency to spreading growth habit, yellow based flower colour, with red second flower colour, sparse pattern of streaks, green stems, with globular and speckled seed color pattern. Similar results have been reported for qualitative traits (Rupika and Bapu 2014). Shannon Weaver indices also confirmed the presence of genetic divergence based on qualitative traits. Thus, in spite of the influence of prevalent environmental factors, qualitative variables can be used to characterize pigeonpea genetic resources.

The pigeonpea genotypes were clustered into three major groups, indicating that these genotypes in the three groups are distantly related. The genotypes in the same cluster group are closely related and they maybe of the same source or origin. Hence, selection of genotypes within these clusters may not be desirable to get higher yield and economic benefits (Muniswamy et al. 2014; Rupika and Bapu 2014). Therefore, for the crop hybridization programs, the choice of suitable diverse parents based on their genetic differentiation would be more fruitful than the choice based on the geographical distances. ICEAP 00540, ICEP00979-1, Karatu-1 and ICEAP00554 would be the ideal genotypes for use as parents in any pigeonpea breeding programme for agronomic improvement. The genotypes clustered in the same group showed their genetic similarities that might be due to free exchange of similar materials with different names that may have overlapped in the previous diversity distribution pattern of the domesticated species (Jaradat and Shahid 2006; Aghaee et al. 2010). Reddy and Jayamani (2019) as well as Niranjana et al. (2014) reported the existence of genetic diversity in pigeonpea using multivariate analysis. Singh et al. (2014); Qutadah et al. (2019); Kinhoégbè et al. (2020) further reported different grouping of genotypes for the agronomic traits in the assessment of genetic variability.

Conclusion

In conclusion, there was sufficient genetic variability existed among pigeonpea genotypes that would help the improvement through identification and selection of parental lines for the traits of interest with greater chances of success in pigeonpea breeding. The study revealed the presence of genetic diversity among the pigeonpea genotypes studied based on the analysis of variance and multivariate tools used for analyses. The results indicated that the higher level of genetic diversity observed within the acquired genotypes from ICRISAT collection in Kenya and Tanzania would enable efficient utilisation and pigeonpea improvement in breeding programs in South Africa and other countries. The variability among the genotypes also will help to identify and select the potential parents for hybridization. The selection for single trait and improvement for this trait would require more breeding work to fix the trait, therefore it is suggested that selection of genotypes for multiple traits as well as directly correlated traits would accelerate pigeonpea breeding for improvement of traits of interest simultaneously. Further characterization this crop using molecular techniques should be conducted to elucidate the environmental factor for utilization in the future breeding programs.

Acknowledgements

The first author would like to thank the Department of Agriculture, Land Redistribution and Rural Development for funding. Additionally, the authors would like to thank the technical assistance and trial management of Paul Rantso, Dinah Scott, Deon Du Toit, and Theodora Mathobisa.

Author Contributions

Dr. Maletsema Alina Mofokeng is a Researcher in Plant Breeding at the Agricultural Research Council-Grain Crops, Potchefstroom, South Africa. Contribution: Conceptualization, experimental design, planting, trial monitoring, data recording, analysis and drafting of the manuscript. Dr. Zaid Bello is a Researcher in Agronomy department of Agricultural Research Council-Grain Crops, Potchefstroom, South Africa. Contribution: trial management, and review the manuscript. Dr Kingstone Mashingaidze is a Senior Research Manager in the Agricultural Research Council-Grain Crops, Potchefstroom, South Africa. Contribution: review, supervision and edition inputs into the manuscript. Dr Abe Shegro Gerrano is a Senior Research Specialist in the Agricultural Research Council-Vegetables, Industrial and Medicinal Plants, Pretoria, South Africa. Contribution: Conceptualization, experimental design, data analysis and review the manuscript.

Conflicts of Interest

The authors have not declared any conflict of interests.

Funding Source

This work was supported by the Department of Agriculture, Land Redistribution and Rural Development.

In Loving Memory

This article is dedicated to our colleague Dr. Maletsema Alina Mofokeng who passed away on 21 April 2022.

Data Availability

All new research results were presented in this article.

Ethics Approval

Not applicable.

References

Adebowale OJ, K Maliki (2011). Effect of fermentation period on the chemical composition and functional properties of pigeonpea seed flour. Int Food Res J 18:1329‒1333

Aghaee M, R Mohammadi, S Nabovati (2010). Agro-morphological characterization of durum wheat accessions using pattern analysis. Aust J Crop Sci 4:505‒514

Ajmal SU, MN Minhas, A Hamdani, A Shakir, M Zubair, Z Ahmad (2013). Multivariate analysis of genetic divergence in wheat (Triticum aestivum) germplasm. Pak J Bot 45:1643‒1648

Choudhary AK, S Kumar, BS Patil, JS Bhat, M Sharma, S Kemal, TP Ontagodi, S Datta, P Patil, SK Chaturvedi, R Sultana, VS Hedge, S Choudhary, PY Kamannavar, AG Vijaykumar (2013). Narrowing yield gaps through genetic improvement for Fusarium wild resistance in three pulse crops of the semi-arid tropics. Sabrao J Breed Genet 45:341‒370

Duran C, N Appleby, D Edwards, J Batley (2009). Molecular genetic markers: Discovery, applications, data storage and visualisation. Curr Bioinform 4:16‒27

Faris DG, KB Saxena, S Mazumdar, U Singh (1987). Vegetable Pigeonpea: A promising crop for India. ICRISAT, Patancheru, India

Gbaguidi AA, A Dansi, I Dossou-Aminon, DSJC Gbemavo, A Orobiyi, F Sanoussi, H Yedomonhan (2018). Agromorphological diversity of local Bambara groundnut (Vigna subterranea (L.) Verdc.) collected in Benin. Genet Res Crop Evol 65:1159–1171

Gerrano AS, A Moalafi, HA Seepe, S Amoo, H Shimelis (2022). Nutritional and phytochemical ompositions and their interrelationship in succulent pods of pigeon pea (Cajanus cajan L.] Millsp.). Heliyon 8:1–7

Gerrano AS, WSJV Rensburg, I Mathew, AIT Shayanowako, MW Bairu, SL Venter, W Swart, A Mofokeng, JJ Mellem, M Labuschagne (2020). Genotype and genotype x environment interaction effects on the grain yield performance of cowpea genotypes in dryland farming system in South Africa. Euphytica 216:80–90

Gerrano AS, WSJV Rensburg, PO Adebola (2015). Genetic diversity of amaranthus species in South Africa. S Afr J Plant Soil 32:39‒46

IBPGR (1994). International Board for Plant Genetic Resources. Rome, Italy

Immad AS, KV Imran, AM Shakeel, MS Pukhta, AD Zahoor, L Ajaz (2018). Genetic Diversity by Multivariate Analysis Using R Software. Intl J Pure Appl Biosci 6:181‒190

Jaradat AA, MA Shahid (2006). Patterns of phenotypic variation in a germplasm collection of (Carthamus tinctorius L.) from the Middle East. Genet Res Crop Evol 53:225‒244

Kasirivu J, S Materechera, M Dire (2011). Composting ruminant animal manure reduces emergence and species diversity of weed seedlings in a semi-arid environment of South Africa. S Afr J Plant Soil 28:228‒235

Khan SA, J Iqbal, H Khurshid, N Saleem, MA Rabbani, M Zia, ZK Shinwari (2014). The extent of intra-specific genetic divergence in Brassica napus L. population estimated through various agro-morphological traits. Eur J Acad Res 2:2255‒2275

Kimaro D (2016). Genetic improvement of pigeonpea (Cajanus cajan (L.) Millsp.) for fusarium wilt resistance in Tanzania. Ph.D. Thesis University of Kwazulu-Natal, Pietermaritzburg, South Africa

Kinhoégbè G, G Djèdatin, LEY Loko, RI Agbo, RK Saxena, RK Varshney, C Agbangla, A Dansi (2020). Agro-morphological characterization of pigeonpea (Cajanus cajan L. Millspaugh) landraces grown in Benin: Implications for breeding and conservation. J Plant Breed Crop Sci 12:34‒49

Kovacic Z (1994). Multivariate analysis, p:293. Faculty of Economics, University of Belgrade, Serbian

Linge SS, HV Kalpande, SL Sawargaonlar, BV Hudge, HP Thanki (2010). Study of enetic variability and correlation in interspecific derivatives of pigeonpea (Cajanus cajan (L.) Millsp.). Electr J Plant Breed 1:929‒935

Lin-Qi X, TT Li, Z Wei, N Guo, M Luo, W Wang, Y Zu, Y Fu, X Peng (2014). Solvent-free microwave extraction of essential oil from pigeonpea leaves and evaluation of its antimicrobial activity. Ind Crops Prod J 58:322‒328

Maesen LGJVD (1990). Pigeonpea: Origin, history, evolution and taxonomy. In: The pigeonpea, pp:15–46. Nene L, SD Hall, VK Sheilla (Eds), pp:15‒46. C.A.B. International, Wallingford, UK

Malik R, H Sharma, I Sharma, S Kundu, A Verma, S Sheoran S, R Kumar, R Chatrath (2014). Genetic diversity of agro-morphological characters in Indian wheat varieties using GT biplot. Aust J Crop Sci 8:1266‒1271

Mallikarjuna N, KB Saxena, DR Jadhav (2011). Cajanus. In: Wild Crop Relatives: Genomic and Breeding Resources, Legume Crops and Forages, pp:21‒33. Kole C (Ed). Springer-Verlag, Berlin, Heidelberg, Germany

Marais LDD, KM Hernandez, E Juenger (2013). Genotype-by-environment interaction and plasticity: Exploring genomic responses of plants to the abiotic environment. Annl Rev Ecol Evol Syst 44:5‒29

Mohammadi SA, BM Prasanna (2003). Analysis of genetic diversity in crop plants-salient statistical tools and considerations. Crop Sci 43:1235‒1248

Molope M (1987). Soil aggregate stability. The contribution of biological and physiological processes. S Afr J Plant Soil 4:121‒126

Mondal MAA (2003). Improvement of Potato (Solanum Tuberosum L.) through hybridization and in vitro culture technique. Ph.D. Thesis. Rajshahi University, Rajshahi, Bangladesh

Muniswamy S, R Lokesha, P Dharmaraj, S Yamanura, JR Diwan (2014). Morphological characterization and assessment of genetic diversity in minicore collection of igeon pea (Cajanus Cajan (L.) Millsp). Eur J Pharm Biopharm 5:179‒186

Niranjana KB, PS Dharmaraj, VB Wali (2014). Genetic diversity and variability studies of advanced breeding lines of pigeonpea (Cajanus cajana. L). Intl J Adv Pharm Biol Sci 3:404‒409

Nunes HF, FFH Freire, VQ Ribeiro, RLF Gomes (2014). Grain yield adaptability and stability of blackeyed cowpea genotypes under rainfed agriculture in Brazil. Afr J Agric Res 9:255‒261

Odeny DA (2007). The potential of pigeonpea (Cajanus cajan (L.) Millsp.) in Africa. Natural resources forum. Wiley Online Library

Ojwang JD, RO Nyankanga, OM Olanya, DO Ukuku, J Imungi (2016). Yield components of vegetable pigeonpea cultivars. Trop Subtrop Agric Environ 67:1‒12

Peeters J, JA Martinelli (1989). Hierarchical cluster analysis as a tool to manage variation in germplasm collections. Theor Appl Genet 78:42‒48

Prasad Y, K Kumar, SB Mishra (2013). Studies on genetic parameters and interrelationships among yield and yield contributing traits in pigeonpea [Cajanus cajan (L.) Millsp.]. Bioscan 8:207‒211

Qutadah SM, S Mehandi, IP Singh, F Singh (2019). Assessment of Genetic Diversity for Polygenic Traits in pigeonpea [Cajanus cajan (L.) Millspaugh]. Intl J Curr Microbiol Appl Sci 8:1581‒1588

Rachovska G, D Dimova, B Bojinov (2002). Application of cluster analysis and principal component analysis for evaluation of common winter wheat genotypes. In: Proceedings of the Scientific Session of Jubilee, Vol. 3, pp:68‒72. Sadovo, Bulgaria

Reddy DSE, P Jayamani (2019). Genetic diversity in land races of pigeonpea (Cajanus cajan (L.) Millsp.). Electr J Plant Breed 10:667‒672

Rupika K, KJR Bapu (2014). Assessment of genetic diversity in pigeonpea germplasm collection using morphological characters. Eur J Pharm Biopharm 5:781‒785

Saidia PS, F Asch, AA Kimaro, J Germer, FC Kahimba, F Graef, JM Semoka, CL Rweyemamu (2019). Soil moisture management and fertilizer micro-dosing on yield and land utilization efficiency of inter-cropping maize-pigeon-pea in sub humid Tanzania. Agric Water Manage 223:105712

Saroj SK, MN Singh, R Kumar, T Singh, MK Singh (2013). Genetic variability, correlation and path analysis for yield attributes in pigeonpea. Bioscan 8:941‒944

SAS Institute (2021). SAS/STAT user’s Guide, version 9.2 SAS Institute, Cary, North Carolina, USA

Saxena KB, RV Kumar, R Sultana (2010). Quality Nutrition through Pigeonpea-Review, Vol. 2, pp:1335‒1344. ICRISAT, Hyderabad, India

Singh AK, S Swain, RK Gautam, PK Singh, AK Betal, T Bharathimeena, N Kumar, SD Roy (2014). Agro-morphological characterization of Bay Islands ppigeonpea (Cajanus cajan) landraces and advanced lines using under Islands conditions. In: 3^rdInternational Conference on Agriculture and Horticulture, October 27–29, 2014 Hyderabad International Convention Centre, India

Smith JSC, OS Smith (1989). The description and assessment of distance between inbred lines of maize. The utility of morphological, biochemical and genetic descriptors and a scheme for the testing of distinctiveness between inbred lines. Maydica 34:151‒161

Sodavadiya MS, JJ Pithia, AG Savaliya, A Pansuriya, VK Korat (2009). Studies on characters association and path analysis for seed yield and its components in pigeonpea (Cajanus cajan (L.) Millsp.). Leg Res 32:203‒205

Syafii M, I Cartika, D Ruswandi (2015). Multivariate analysis of genetic diversity among some maize genotypes under maize-albizia cropping system in Indonesia. Asian J Crop Sci 7:244‒255

Syed MQ, M Suhel, IP Singh, S Farindra (2019). Assessment of Genetic Diversity for Polygenic Traits in Pigeonpea [Cajanus cajan (L.) Millspaugh]. Intl J Curr Microbiol Appl Sci 8:1581‒1588

Upadhyaya HD, KN Reddy, CLL Gowda, S Singh (2007). Phenotypic diversity in the pigeonpea (Cajanus cajan) core collection. Genet Res Crop Evol 54:1167‒1184

Vales M, R Srivastava, R Sultana, S Singh, I Singh, G Singh, S Patil, K Saxena (2012). Breeding for earliness in pigeonpea: Development of new determinate and non-determinate lines. Crop Sci 52:2507‒2516

Virk PS, GS Khush, SS Virmani (2003). Breeding strategies to enhance heterosis in rice. In: Hybrid Rice for Food Security, Poverty Alleviation and Environmental Protection, pp:21‒29. Virman SS, CX Mao, B Hardy (Eds). International Rice Research Institute Los Banos, Philippines

VSN International (2020). Genstat for Windows, 20^th edn. VSN International, Hemel Hempstead, UK

Yohane EN, H Shimelis, M Laing, I Mathew, A Shayanowako (2020). Phenotypic Divergence analysis in pigeonpea [Cajanus cajan (L.) Millspaugh] germplasm accessions. Agronomy 10:1682–1699