Phenotyping cowpea accessions at the seedling stage for drought tolerance using the pot method

One of the most important screening techniques used in cowpea selection for drought tolerance is screening at the seedling stage. The objective of this study was to phenotype 60 cowpea genotypes for seedling drought tolerance in screen houses (glasshouse and greenhouse). A triplicated 6 × 10 alpha lattice design with four blocks was used for the experiments. After planting, pots were watered to field capacity, thereafter watering was completely withheld for 4 weeks after planting (WAP), when plants were at the three-leaf stage. Principal component analysis revealed that of the 14 variables, the first four expressed more than 1 eigenvalue. Data showed that PC1, PC2, and PC3 contributed 39.3%, 15.2%, and 10% respectively, with 64.68% total variation. Bartlett’s test of sphericity was significant at p<0.05, while the Kaiser-Meyer-Olkin measure of sampling adequacy was 77. A PCA plot and biplot showed that the number of pods (NP), seeds per pod (SP), survival count (SC), pod weight (PWT), and stem wilting in week one (WWK1) had the most significant contributions to genetic variability to drought tolerance and to yield after stress imposition Based on the PCA, biplot, and cluster plot, the accessions IT 07-292-10, IT 07-274-2-9, IT90K-59, 835-911, RV 343, and IT 95K-2017-15 had the maximum variability in terms of number of pods, seeds per pod, survival count, pod weight and wilting in week one after drought imposition. Cowpea accessions 835-911, IT 07-292-10, RV 344, Kangorongondo, and IT 90K-59 were the major individuals that contributed mainly to domain information model (DIM) 1 and 2. The accessions that contributed the least were IT 89KD288, Chibundi mavara, and TVU12746. Thirty-six cowpea accessions from both screen houses were tolerant to drought, 15 were moderately tolerant, while 23 were susceptible. The findings of the study provided a useful tool for screening and determining drought-tolerant and susceptible accessions at the seedling stage. Thirty-six cowpea accessions from both screen houses were tolerant to drought as well as those that showed great variability can be used as parents in future cowpea breeding programmes.


Introduction
Cowpea [Vigna unguiculata (L.) Walp.], Fabaceae, (2n = 2x = 22) is an important leguminous crop in developing countries, especially in sub-Saharan Africa, Asia, and Latin America, with a genome size of about 620 million base pairs (Boukar et al., 2018). The improvement of cowpea mainly dependent on breeding and selection from existing landraces according to the phenotypic variability, which is largely influenced by environmental conditions. According to the Food and Agriculture Organization of the United Nations (FAO), cowpea was grown on 1 million ha in Africa in 2014, with the bulk of production occurring in West Africa, particularly in Niger, Nigeria, Burkina Faso, Mali and Senegal (FAOSTAT, 2017). The global cowpea production was 5.59 million and the average yield 443.20 kg/ha (Gull et al., 2018). Africa leads in both cultivation area and production, accounting for about 95% of each. Niger and Nigeria are the leading producers of cowpea, together accounting for about 70% of the cowpea cultivated area and 67% of production worldwide. Most cowpea cultivars have relatively shorter growing period and maturation cycles of 60 to 80 days, which makes it suitable for drought-prone regions (Kyei-Boahen et al., 2017).
Drought is one of the most severe environmental stress, and it has a significant negative impact on crop yield. Gomes et al. (2019) recommend the use of water-efficient varieties in combination with good crop husbandry practices. Cowpea plants exposed to temperatures of 30 to 38 °C from eight days after emergence to maturity had very limited vegetative growth and reproductive potential (Singh et al., 2010). Hall et al. (2003) observed that there is a great need to screen and breed for drought-tolerant and water use efficient varieties in Africa, as cowpea is grown mostly under rain-fed conditions, with frequent exposure to intermittent droughts. Gomes et al. (2019) recommend the use of well-adapted, early maturing cultivars in the smallholder farming sector to escape losses from late season droughts. Araujo et al. (2018) studied germination of cowpea cultivars under osmotic stress, seeds of three cowpea cultivars (BRS Tumucumaque, BRS Aracê, and BRS Guariba) grown at five osmotic potentials after three pre-treatments: pre-soaking in deionised water, pre-soaking in salicylic acid, and without pre-soaking. It was observed that salicylic acid promoted a reduction in abiotic stress, and BRS Guariba was more tolerant to water deficits and adjusted its cellular electrolyte leakage to increase its proline content under induced water stress.
In a wooden box experiment to screen cowpea recombinant inbred lines (RILs) for seedling drought tolerance, Alidu et al. (2018) used 200 inbred lines. It was observed that 12 RILS performed well for recovery, 13 RILS were susceptible to drought stress, and 11 RILS had higher relative water and chlorophyll contents. Ajayi et al. (2018) analysed 10 cowpea accessions under screen house conditions and observed significant differences among accessions for percentage plant recovery, stem regrowth, and stem greenness. For the evaluation of four Mozambican cowpea landraces for drought tolerance, Martins et al. (2014) determined that variability exists among the landraces in terms of growth under drought conditions, with Timbawene moteado having considerably higher leaf dry biomass, leaf and nodule protein content, and symbiotic nitrogen fixation compared to those of other landraces, as well as the lowest increase in proteolytic activity.
In a screen house experiment to select drought-tolerant cowpea seedlings, Ismai'la et al. (2015) evaluated 23 cowpea accessions at the seedling stage in the 2013 and 2014 growing seasons.
They observed that plant height, number of leaves, and stem greenness were all affected by drought stress. It was found that five varieties, Kanannado, Danila, IT07K-297-13, IT03K-378-4, and Aloka local, were highly tolerant to drought. In addition, six varieties IT07K-322-40, IT07K-313-41, IT07K-291-92, IT06K-270, IT07K-244-1-1, and IT06K-275, were classified as highly susceptible to drought and the remaining 12 varieties were found to be neither tolerant nor susceptible to drought. Ismai'la et al. (2015) recommended the use of early maturing cowpea cultivars in order for farmers to escape the effects of a late season drought. Most cowpea plants exposed to moisture variation during the vegetative or reproductive stages perform poorly; hence, seedling-stage screening is ideal in this scenario. The objective of this study was to phenotype 60 cowpea genotypes for seedling drought tolerance in screen houses.

Plant material
A total of 60 cowpea accessions collected from three geographic origins were used in this study (Table 1)

Planting and data collection
Seeds of cowpea accessions were planted in 20 cm diameter pots filled with a mixture of topsoil and compost (3:1) in a greenhouse at the Agriculture Research Council -Grain Crops in Potchefstroom, South Africa, in January 2019 (environment 1). The experiment was repeated in a glasshouse in February 2019 (environment 2). An alpha lattice design with four blocks was used for both experiments. A total of 60 accessions were selected for drought tolerance at the seedling stage and were used in the experiments. A triplicated 10 × 6 alpha lattice design was used. The cowpea varieties were planted in 20 cm diameter pots in the screen houses. After planting, pots were watered to field capacity for establishment, thereafter watering was completely withheld for 4 weeks after planting (WAP), when plants were at the three-leaf stage.
Thereafter, wilted plants of each variety were counted daily until all the plants of the susceptible lines appeared dead. Stress was measured by observing all dead plants in the susceptible group.
Watering resumed at three weeks after stressing in both the greenhouse and glasshouse experiments until harvest. After the resumption of watering, numbers of recovered seedlings were rated for recovery. Based on the days to wilting and percentage recovery, the accessions were rated as either drought-tolerant or -susceptible.

Data collection
i. Temperature conditions of the screen houses The daily minimum and maximum temperatures of the screen houses were captured using temperature loggers. The loggers were placed in the screen houses and set to record the temperature at hourly intervals for the whole period of the experiment ( Figure 1).

Results
There were significant differences at the seedling stage of most genotypes tested on stem greenness in week one (SGWK1), stem greenness in week two (SGWK2), stem greenness in week three (SGWK3), wilting in week one (WWK1), wilting in week two (WWK2) and wilting in week three (WWK3) ( Table S1). Most of yield-related traits evaluated were affected by water stress, with the exception of the environment on the number of pods (NP), environment on number of seeds per pod (NS), replicates on number of pods (NP), replicates on pod length (PL), and replicates on pod width (PWDTH) ( Table S2).

Principal component analysis
Principal component analysis (PCA) revealed that out of the 14 variables, all four components expressed more than 1 eigenvalue (   A scree plot to show the relationship between eigenvalues and principal components was constructed to summarise the contribution of PCs ( Figure S1). The plot showed that maximum variation was present in variable 1 with the highest eigenvalue of 5.8 followed by variable 2 (2.1), variable 3 (1.4), and variable 4 (1). Variable 14 had the lowest eigenvalue (0).
A further PCA with VARIMAX rotation was conducted to assess how 14 variables were clustered. The plot showed how closely related the 14 parameters were and these results are in tandem with the PCA plot ( Figure S2). Three components were rotated based on the eigenvalues over 1 and the scree plot ( Figure 1). Bartlett's test of sphericity was significant at      The 60 cowpea accessions used varied in their response to drought imposition. Thirty-six cowpea accessions from both screen houses were tolerant to drought, 15 were moderately tolerant, while 23 were susceptible, based on the 14 traits measured (Table S4).
In the biplot, accessions IT 07-292-10, RV 343, and IT 95K-2017-15 had the maximum variability for the number of pods (NP), seeds per pod (SP), survival count (SC), pod weight (PW), and wilting in week one (WWK1) (Figure 6). The neighbour-joined cluster analysis generated by UPGMA divided the 60 cowpea accessions into two main clusters ( Figure S4). From the results, it was observed that there were two major clusters and other sub clusters whose accessions were closely related genetically. The cluster analysis showed that the 60 accessions were grouped into two major clusters and other sub clusters with their respective distances. The phenotypic distance index based on morphological traits ranged from one (IT 89KD-288 from IITA) to 50 (TVU 13004 and IT96D-610 from IITA). The phenotypic distance index of other accessions in other subclusters was less than 20.

Discussion
This study revealed that moisture is a very important component in plant growth and reproduction. According to Padi (2004), when moisture stress is imposed during the vegetative stage, it has the most effect on shoot and dry weight reduction in cowpeas. It is also during the vegetative stage that plants set up their architecture for reproduction. Alidu (2018) observed that moisture stress imposed after the pod-filling stage in determinate accessions has a limited reduction on the shoot and root biomass.
Most of the cowpea accessions showed differences in their response to drought imposition in their stem greenness from week one to week three after drought imposition. A similar variation was also observed when wilting was recorded from week 1 to week 3 after drought imposition.
In both environments, the temperature had a significant effect on the performance of the accessions. In the greenhouse experiment, the average daytime and night-time temperatures were 34.24 °C and 23.98 °C, respectively. In the glasshouse experiment, the mean daytime and The main traits that accounted for variability from PC1 to PC5 in the screen houses were pod weight (PWT), pod length (PL), seeds per pod (SP), seed weight (SWT), number of pods (NP), stem greenness in week two (SGWK2), and days to emergence (DTE). This implies that accessions that emerged earlier and withstood the imposition of drought had higher chances of podding and producing seeds. Thus, it is imperative to consider these traits in further enhancing cowpea accessions' tolerance to drought at the seedling stage. Ajayi et al (2018)

Conclusion
The findings of this study provided a useful tool for screening and determining drought-tolerant and -susceptible accessions at the seedling stage. The results of the investigation were also useful in selecting accessions especially for average seeds per pod (AVSPD), number of seeds