Cowpea [Vigna unguiculata (L.) Walp.] being an excellent source of protein, contains all the essential amino acids (Vasconcelos  et  al.,  2010;  Oliveira  et  al.,  2016).  It  is also rich in carbohydrates, vitamins and minerals, besides having great fibre content and low-fat content, constituting an important food component in several countries including Ghana (Freire Filho et al., 2012; SARI, 2014). As such the crop is grown in all the Savanna ecological regions of Ghana where drought and heat stress compounded by poor soil fertility conditions limit the production of many other crops. Among the abiotic constraints to production of the crop, drought is one of the most important stresses because, the crop is typically grown in the Savanna and Sahel regions where rainfall amount and distribution are unreliable. The severity and occurrence of drought stress is expected to increase as a result of global environmental changes, causing major implications for food supply (Fan et al., 2015). Compounding this is an increasing world population that requires a rise in food production by more than 70% before 2050 (Godfray et al., 2010; Parfitt et al., 2010; Wallace, 2000).

Verbree et al. (2015) indicated that, the use of multiple traits for evaluation of drought tolerance affirms the complexity of drought tolerance in cowpeas compared to other crops. Also, Aliyu and Makinde (2016) and Swain et al. (2017) indicated that cowpea breeding is largely based on  selection of parents, followed by hybridization, in order to form a base population and generation advancement with simultaneous selection for more than one trait (Batieno et al., 2016). Therefore, the most common breeding method consists of screening under controlled drought stress; the offspring derived from populations, followed by the assessment of selected genotypes at a location where drought occurs frequently, and testing the most promising genotypes for yield potential and yield stability in multiple sites representing the target ecology (Batieno et al., 2016; Ortiz et al., 2002). Analysis of genetic divergence seeks to identify parents for creating populations with genetic variability and consequent genetic gain in successive selection cycles (Santos et al., 2016). Though various cowpea-breeding materials such as F2, F3 and backcross populations have been used for drought tolerance studies in cowpea, the empirical approach mainly relies on the use of recombinant inbred lines (RIL) to enable the consistent evaluation of performance and understanding of genotype-by-environment interaction, as the intensity and frequency of naturally occurring drought stress are not predictable. The RIL population, developed through single seed descent of several selfed generations consists of individual lines carrying dispersed homozygous segments of a parental chromosome. The objective of this study was to select drought tolerant cowpea recombinant lines developed from a drought tolerant and susceptible parents using the wooden box screening technique.

Population development

Four  hundred  and  fifty  Recombinant  Inbred  Lines  (RILS)   were developed through single seed decent and an F₂:₆ generation was obtained between 2010 and 2015. The parents for the developed population were IT-93k-503-1; a drought tolerant and a medium maturing, indeterminate line crossed with IT97k-279-3; an early maturing line with determinate character. These two lines were obtained from the International Institute of Tropical Agriculture (IITA) Kano, Nigeria.

Geographical location and experimental design for screen-house experiment

The study was carried out at Savanna Agricultural Research Institute (SARI), Nyankpala. Nyankpala is located in the Northern Guinea Savanna Zone with a mean annual rainfall of about 1000 mm. It is located on latitude 9°, 25″ N and longitude 0°, 58″ W with an altitude of 183 m above sea level. Wooden boxes of 130 cm length, 65 cm width, 15 cm depth, and 2.5 cm thick planks were arranged in a screen house as described by Singh et al. (1999a). The boxes were lined with perforated polyethylene sheets and filled with one: one mixture of top soil and sand which averaged 39% sand, 2% clay, and 59% silt by analysis of the composited soil used to fill the boxes. The composite soil had a pH value of 6 and an organic matter content of 6% with N, P, K of 0.05, 3, and 45% respectively. The soil had a bulk density of 1.33 g/cm³. Bulk density (g/cm³) was manually estimated as: Dry soil weight (g) / Soil volume (cm³)^. The boxes were filled to 12 cm depth leaving about 3 cm space on the top for watering. The polyethylene lining along the sides and bottom of the boxes ensured even distribution of water. A spirit level was used to ensure a flat soil surface on the boxes before and after watering.   Equidistant holes were made in straight rows 10 cm apart with a hill to hill distance of 5 cm within the rows. Each box was watered thoroughly and allowed to drain for two days before planting. The moisture content was then taken before planting using the WET sensor with the HH2 moisture meter (Plate 3). Two seeds were sown in each hole and were thinned to one plant per hill one week after germination. Each box contained one row each of 10 recombinant inbred lines plus the two parental checks making it 12 lines in each box (Table 1.). Seedling drought screening was done in two sessions using augmented design because of large size of recombinant inbred lines to be tested (200), as well as limited space for experiment arrangement in the screening house. Screen house experiment one (session one) was done between June and July 2015, while screen house experiment two (session two) was done between October and November 2015, as a repeat to confirm seedling screening for drought tolerance. The boxes were watered daily using a small watering can until the appearance of the first trifoliate leaf, after which watering was stopped (Plate 1 and 2).

Relative water content for screen-house experiment one and experiment two

The relative water contents ranged between 20 and 70% for both drought susceptible and drought tolerant inbred lines across the population for screening. The parental lines used as checks, had relative water contents of 60% for Screen-house. Experiment one and a relative water content range of 74 and 22% for Screen-house experiment two with the parental checks of 40 and 35% for IT93K-503-1 and IT97K-279-3 respectively. These results are based on the summarized and selected potential tolerant and susceptible lines used for field screening for drought (Tables 2 and 3).

Proportion of survival for screen-house experiments one and two

The proportion of recovery and survival for screen-house experiment one ranged between 93 and 5% for the potential tolerant  and  susceptible  inbred  lines  with  the parental checks scores of about 72 and 49% for IT93K- 503-1 and IT97K-279-3 respectively. The proportion of survival for screen-house two were relatively higher compared to screen-house one. Survival for some inbred lines were 100%, whereas the lowest scored 54% with the parental checks 91 and 80% for IT93K-503-1 and IT97k-279-3 respectively. Results for the selected potential drought tolerant and susceptible inbred lines for screen-house screening at the seedling stage for experiment one and two are summarized and presented in Tables 2 and 3 respectively.

Mean squares of measured traits for cowpea inbred lines evaluated in the screen -house for tolerance to drought

There were significant differences (p < 0.05) for chlorophyll contents for the 200 inbred lines used in the study. Chlorophyll contents taken during 7, 14, 21 and 28 days after water stress treatment significantly varied among the inbred lines, but no significant differences were observed for the parental checks used for the study (Table 4). Significant differences were observed for relative water content taken at 14 days of stress imposition, leaf wilting and recovery.

Pearson correlation coefficient between relative water content, chlorophyll and leaf wilting

Relative water contents at 14 days of water stress correlated negatively with chlorophyll at 7, 14, and 28 days of stress imposition. Relative water content also correlated negatively with leaf wilting at 7 days but positively correlated with leaf wilting at 14, 21 and 28 days of stress imposition. Relative water content at 28 days of stress however, correlated positively with chlorophyll at 7, 14, 21 and 28 days of stress and leaf wilting at 7 and 28 days but correlated negatively with leaf wilting at 14 and 21 days during the stress imposition (Table 5). Leaf wilting after 7 days of water stress correlated positively with chlorophyll for 7, 14, 21 and 28 days of water stress, however, leaf wilting 14, 21 and 28 days negatively correlated with chlorophyll at 7, 14, 21 and 28 days of stress imposition.

In the current study, water stress significantly reduced the chlorophyll contents for both the parental checks and the inbred lines used. This result corroborates with similar drought related studies report that, drought related traits such as leaf area index, leaf area duration and chlorophyll contents decreased as the water stress duration increased (Khan et al., 2015; Pirzad et al., 2011; Deblonde and Ledent, 2000).  Also, Yuan et al. (2016), who in their study on the effects of different levels of water stress on leaf photosynthetic characteristics and antioxidant enzyme activities of greenhouse tomato reported that water stress decreased stomatal conductance net photosynthetic rate, photosynthetic rate at light saturation, and chlorophyll content in all development stages of tomato resulting in yield reduction.

Water deficit affects the photosynthetic ability of plants by changing the content and components of chlorophyll, reducing the net CO₂ uptake by leaves, thereby decreasing activities of enzymes in the Calvin cycle (Cornic and Massacci, 1996; Gong et al., 2005; Lawlor and Tezara, 2009). In this current study, as water stress imposition progressed, there were reductions in the chlorophyll contents for all the inbred lines.

Similar study by Tuberosa (2012), on a three-week drought treatment on Katahdin-derived potato cultivars resulted in a decrease in the leaf water content of the cultivars in relative to the control. There was significant variation among the inbred lines for relative water content (RWC) in this study. This observation corroborates with Zegaoui et al. (2017) who reported that two cowpea land races originating from the arid area, maintained a higher RWC over the duration of the drought stress and transpired less than the landrace from the temperate area. Studies by Bogale et al. (2011) and Pirzad et al. (2011) on wheat genotypes and  Matricaria chamomilla respectively for drought tolerance reported that changes in the relative water content of leaves are considered as a sensitive indicator of drought stress and more useful indicator of plant water balance (Bogale et al., 2011; Clavel et al., 2005). Therefore, the inbred lines with higher relative water content may have a high potential for survival under field drought conditions and subsequently give good yields. Also, relative chlorophyll values for the screen-house experiment one and two in this study gradually reduced as the water stress advanced. This corroborates studies by Bogale et al. (2011) that water deficit has tremendous effects on chlorophyll fluorescence and leaf gas exchange parameters. Thus, photosynthesis rates decreased with decreases in stomatal conductance. Therefore, the relative water content, chlorophyll contents, leaf wilting, survival, and recovery from drought are very good indices for Screen-house selection for seedling drought tolerance of large populations of inbred lines for drought evaluation in the field (Pungulani et al., 2013; Muchero et al., 2008; Singh et al., 1999c).

There were variations with respect to seedling drought tolerance for the 200 inbred lines using the parents as checks in this experiment. The inbred lines that performed well for recovery were 84, 406, 325, 223, 75, 186, 131, 20, 38, 230, 398 and 353, the susceptible ones were, 142, 78, 55, 57, 408, 255, 396, 116, 189, 255, and 28. The inbred lines that maintained a higher relative water content during stress imposition avoided drought and recovered better, were inbred lines 406, 325, 84, 230, 38 and 75, which recorded higher percentages of relative water contents as well as chlorophyll content.

The inbred lines whose proportion of survival ranged between a score of 100- 60% consistently for screen-house 1 and 2 experiments were selected as the potential tolerant lines. Whereas the ones  that  had  poor survival and recovery after re-watering; whose recovery ranged between 0-40% were identified as potential susceptible inbred lines. The relative water content and chlorophyll for the potential seedling tolerant lines were in the range of 40-70% and 32-53%. The screen-house experiment was repeated to confirm the potential seedling tolerant and susceptible inbred lines that were subsequently selected for the field screening under managed drought conditions. Most of the potential susceptible inbred lines also had lower chlorophyll contents but the relative water content was most often within average (39 and 50%). This could be due to the environmental conditions prevailing at the time of the second experiment thus leading to the high proportions for survived seedlings across all the treatments for all the blocks. The use of physiological traits in scoring for seedling tolerance in this study has facilitated the classification of the inbred lines into drought tolerant and susceptible inbred lines. The genetic variability found for this morphological trait among these inbred lines in the screen-house study, suggest that opportunity exists for selecting superior genotypes under water limited conditions in the field.

The authors have not declared any conflict of interests.