Variation in morphological traits of accessions
Variation in germination time
The statistical outcomes presented in Table 3 demonstrate the variation in germination time among the diverse accessions. The variation in germination time was found to be significantly different between accessions (p = 0.000) and not between blocks (p = 0.42). Furthermore, the interaction between accessions and blocks was also found to be significant. Consequently, the blocks do not exert an influence on the germination time. This outcome is consistent with the hypothesis that the four blocks were exposed to identical conditions.
However, the accession V1 exhibited accelerated germination, occurring six days after sowing. The process of germination of the accession V4 is observed to occur at a later stage than that of the white accession, with a delay of three days in the onset of this process. The remaining two are considered intermediate. In all the accessions studied, radicle emergence occurred between six and nine days after sowing. The result obtained is analogous to that previously reported by
Wassouo et al. (2019), who documented a germination time range of 6-7 days for Bambara groundnut. The findings of this study demonstrate a congruence with those of
Touré et al. (2012), who reported that the germination of seeds in Bambara groundnut occurs within the timeframe of 6 to 15 days following the process of sowing.
Karikari (1999) observed significantly higher values (14 to 24). The observed variability can be attributed to the presence of diverse plant material, variable climatic conditions and the varying quality of the seeds.
Variation in germination rate
A subsequent analysis of the results indicates a significant difference (Fig 1) between the accessions in terms of germination rate (p = 0.006). However, accession V1 exhibited the lowest germination rate (44%), while accession V2 demonstrated the highest (88%). However, no significant differences were observed between the V1 and V4 accession, the V4 and V3, or the V3 and V2. This negligible discrepancy may be attributable to the substantial variation observed among individuals (
i.e. high standard deviations), given the limited understanding of the origins of these seeds and their parents. However, the mean germination rate was 66%. This value is higher than that reported by
Touré et al. (2013), who found that in Bambara peanut accessions, the average germination rate is 58% at 30 days after sowing. Accession V2 exhibited the highest (88%). Consequently, it can be deduced from this observation that accessions exhibiting high germination rates would possess sufficient plants to ensure high yields, thereby conferring them with a certain agronomic interest in enhancing Bambara groundnut cultivation in the Sahelian regions.
Variation in length
The statistical analysis performed demonstrated that leaf length exhibited significant variation between accessions (p = 0.000) (Fig 2 and 4). However, the leaves of accession V1 are notably longer (7 cm) than those of the other three accessions. The shortest of these are the accession V3, which measure 5.5 centimetres. These results are identical to those of (
Djè et al., 2005) who reported in Côte d’Ivoire that morphotypes of Bambara groundnut (White, Black and Red), the white variety gave the longest leaflets than the red and black varieties. Similar to the results of
Bonny et al. (2011) for white and brown varieties, which had the longest leaflets compared to the others.
Variation in width
The results of the analysis demonstrate the presence of variability in leaf widths (Fig 3). Indeed, a significant variation in leaf width was observed between accessions (p = 0.000), However, accession V1 exhibited the widest leaflets (2.8 cm), which was greater than the other accessions. These results differ from those reported by
Yao et al. (2005), who found that the in Côte d’Ivoire, the Rouge variety was found to have the widest leaflets. than the white and black varieties. The leaf width observed in accession V1 indicates that these plants possess extensive photosynthetic surfaces, a trait that is likely attributable to the substantial light absorption capacity of these leaves in the Sahelian region, where temperatures are notably high. Consequently, they accumulate more organic matter.
The results from the descriptive statistical analysis (Table 4) demonstrate that, the mean number of leaves was determined to be seven at 40 days. This finding is incongruent with the observations reported by
Wassouo et al. (2019), who documented that the mean number of leaves in Bambara groundnut is equivalent to 40 at 56 days post-sowing. The observed discrepancy can be attributed to the temporal increase in the average number of leaves. The mean plant height was found to be 15.08 centimeters, which is different to the findings of
Temegne et al. (2018) who reported that the mean height of Bambara groundnut plants was 31,18 centimeters at 56 days post-sowing. This slight difference can be attributed to the fact that the growth of Bambara groundnut plants stabilizes after a certain number of days following the initial sowing. Consequently, the height of the plant exhibits a range from 9 centimeters to 20 centimeters. The values obtained in this study are analogous to those reported by
Touré et al. (2013) who found values of 14 centimeters. However, these values are lower than those reported by
Ndiang et al., (2012) in Cameroon, who found values ranging from 20 to 38 centimeters. The observed differences can be attributed to the genotype of each plant material, which is susceptible to environmental influences. A study of Bambara groundnut accessions on quantitative traits was reported by
Khan et al., (2020), who noted a coefficient of variation (CV) ≥20% for quantitative traits studied. This finding aligns with the results of our study on variation coefficients, suggesting heterogeneity within the plant material for these variables. A similar, relevant observation of a high coefficient of variation was confirmed by
Goli et al. (1997).
Variation in yield components and dry matter (Table 5 and 6)
A thorough examination of the data reveals a statistically significant impact of water stress on the overall weight of 100 pods and seeds produced per accession. Accession V1 exhibited a significant response to water stress, resulting in the production of the smallest pods (22 g). These results are consistent with those reported by
Erskine et al. (1993), who found that abiotic stress significantly reduces lentil yield. In contrast, accession V4 exhibited a reduced impact, yielding an average pod weight of 24 grams. Accessions V2 and V3 exhibited the highest recorded weight of 26 g. These two accessions can be regarded as exhibiting reduced sensitivity to water stress. Consistent findings have been reported in related studies
(Tambal et al., 2000). On average, P100Gss exhibited a discrepancy of approximately 0.6 g in favor of the V3 accession. This phenomenon does not appear to exert a substantial influence on yield in our experimental setting, as evidenced by the observation that accession V2, characterized by a low pod weight of 26.14 g, exhibits the highest seed weight of 24.99 g. The total number of seeds is diminished; however, the reduction in the number of sheaths due to water stress is comparatively greater in accession V1, which has seeds of a lighter hue. The range of these values was from 6.16 in accession V1 to 9.47 in accession V2. Water stress has been demonstrated to result in pod abortion, which is characterized by a reduction in seed number due to the loss of young pods prior to seed formation and filling. These results are consistent with those reported by
Shrestha et al. (2006), who observed that water stress led to a 27% reduction in the number of pods, with a significantly higher decrease in empty pods when water was retained. However, a higher root biomass weight was observed in accession V1 compared to the other accessions. However, this variability among accessions can be attributed to the observation that early accessions consistently yield high levels of dry matter. The findings of this study are consistent with those reported by
Ouedraogo et al. (2008), who observed that late varieties yielded the least. The range of above-ground biomass values was from 5.34 grams in accession V1 to 4.19 grams in V2. Accession V1 exhibited the highest above-ground biomass value in comparison with accession V2. Despite its notably high above-ground dry weight, accession V1 exhibited the greatest sensitivity to water stress among the accessions examined, resulting in the lowest grain yield. This phenomenon could be attributed to two potential factors. Firstly, there is the possibility that vegetative organisms are being disproportionately assimilated under moderate conditions. Alternatively, the stressed V1 accession may have allocated a greater proportion of its resources to biomass growth, thereby compromising grain yield. The present findings contradict those of
Bouzerzour et al. (1998), who posited that in variable environments, ensuring sufficient above-ground biomass production is imperative to ensure an acceptable yield. However, similar results were observed in the study by
Surson et al. (2025), who reported that the rice M2 RD.43 variant 1 had more leaves but fewer whole seeds, as well as lower whole seed weight, seed width, seed length and seed formation percentage.
Variation of physiological traits of accessions
A statistical analysis of Table 7 reveals that the maximum error ratio (TRE) was observed at N1 (1L/5D), while the minimum was observed at N3 (0.25/5D). This finding suggests that the TRE declines with increasing water deficit. These results are consistent with those reported by
Albouchi et al. (2000), who observed that the TRE in durum wheat declines as soil water content decreases.
Lilya et al. (2002) observed that accessions capable of maintaining a high TRE in the presence of water stress are classified as tolerant accessions. A minimum water loss (TDE) of (-2.114± 0.067a) was observed in accession V2 at a severe stress level (N3, 0.25L 5D) compared with a high-water loss (-19.721±0.063c) recorded in accession V1. Additionally, accession V2 exhibited the smallest leaf area at (N3, 0.25L/5d). This suggests that the rate of water loss is inversely proportional to the surface area, with smaller surface areas resulting in reduced water loss. The results of this study are consistent with those reported by
Kelliher et al. (1980). The aforementioned researchers observed that, in durum wheat, the TDE increases as the leaf area expands. Accession V1 exhibited the largest surface area across all stress levels, while accession V2 demonstrated the smallest surface area in comparison to the other accessions. The smallest leaf areas were recorded at the severe stress level (N3, 0.25 L/5 d).
This finding suggests a negative correlation between leaf area and water deficit, indicating that as the water deficit increases, leaf area decreases. According to
Granier et al. (2000), the leaves of plants subjected to water deficit exhibited smaller apparent final sizes in comparison to the control group.
Lebon et al. (2006) confirmed that one of the earliest responses of plants to water deficit is the reduction of leaf area.
Lebon et al. (2006) demonstrated that the reduction in leaf area under water is an adaptive mechanism for plants to limit leaf transpiration when water conditions become unfavorable. Conversely, only accessions V3 and V2 exhibited a diminished number of stomata among the accessions examined. A substantial decrease in the quantity of stomata was detected under conditions of severe stress (N3, 0.25 L/5 d) across all accessions, with the water deficit leading to a reduction of approximately 10 to 20% in stomata, there by constraining transpiration. This is explained by the fact that the onset of water stress implies a reduction in water loss, through the closure of stomata, resulting in a reduction in photosynthetic activity and transpiration. Thus, accession V2 exhibited a reduced level of transpiration in comparison to the other accessions. The findings of this study demonstrate a congruence with the results previously reported by
Bourou et al. (2022) on the subject of Tamarindus, wherein a correlation was established between transpiration, stomatal conductance and net photosynthesis. A low water use efficiency (WUE) was observed (Fig 3) in accession V1 (0.80g/l) in contrast to accession V2, which has a higher WUE (0.89g/l). Accessions V3 and V4 demonstrate intermediate WUE values. The mean WUE values obtained from the accessions range from 0.80 g/l to 0.89 g/l. The findings of
Chetto et al. (2020) indicated an average value of 0.712 WUE for the farms under analysis, signifying an average of 71.2% of the WUE by citrus growers. This finding aligns with the conclusions drawn from the present study. These divergent outcomes are consistent with the findings reported by
Karam et al. (2007), who conducted a study on the water use efficiency (WUE) in corn. This phenomenon can be attributed to the higher WUE exhibited by plants with C4 photosynthetic metabolism in comparison to C3 species (
Rubino et al., 1999). The V2 model demonstrated a high level of water use efficiency, as illustrated in Fig 5. This phenomenon can be attributed to the established relationship between yield and WUE, wherein elevated levels of yield are accompanied by heightened WUE. Therefore, it can be concluded from this observation that increases in yield are associated with accessions that demonstrate effective water use efficiency.
Table 8 presents the characteristics of accessions on yield in real conditions and under stress In the second season, an experiment was conducted at two distinct locations with the objective of estimating the mean yield per accession. The accessions were evaluated under two distinct conditions: Water-limited and natural. An evaluation of the test results indicates that accession V2 exhibits a significantly higher grain yield (65.72±0.50b) under conditions of water limitation in comparison to the other accessions. These results are similar to those of
Karthiga et al. (2025), who observed an increase in cucumber productivity when genotypes were evaluated under water-deficient conditions. These results are also in line with observations of
(Dolinassou et al., 2016) on Arrachis hypogea, who reported that in the same location, the Gobo-55-437 line exhibited the highest average yield, while the NW-Red Esimbi genotype demonstrated a low yield.Water deficits have been shown to reduce yield by up to 10% in the V1 accession. Nevertheless, accession V3 demonstrated the highest yield (65.50±0.47 g) when the accessions were assessed in the field. These results are consistent with those reported by
Shrestha et al. (2006), who found that seed yield in Lens culinaris was significantly higher than in other genotypes under conditions of adequate irrigation. Among the accessions examined, only two (accession V3 and V2) exhibited the optimal yield under stress conditions, with values of 64.96 g and 65.72 g, respectively. The observed outcomes can be attributed to the combined effects of water shock and end-of-season drought in this agroecological zone. However, it is noteworthy that only the V1 accession exhibited a reduction in grain yield. The decline in grain yield under water stress is predominantly attributed to the reduction in the number of seeds. These results are consistent with the observations of
Shrestha et al. (2006), who noted that in Lens culinaris medikus, the 70% reduction in seed yield induced by water deficit was primarily due to a decrease in the number of pods and seeds. The utilization of the agronomic trait (number of pods and seeds) can facilitate the identification of accessions with superior traits
(Gahoonia et al., 2006) that can be integrated into the breeding program for the production of drought-tolerant varieties.
The highest stress tolerance indices were observed in accessions V2 and V3, while accession V1 was the most sensitive (Table 7). Similar results reported by
Mohammadi et al. (2011) indicated that the index (STI) is well suited for selecting the most productive RILs (Recombinant Inbred Lines) under both stressful and favorable conditions. It is therefore clear that the stress tolerance index (STI) is a strong discriminator between genotypes with high yields and potential tolerance to water stress and other genotypes. According to
Fernandez (1992), the STI can be used to identify genotypes that produce high yields under both favorable and stressful conditions.
Correlation matrix between the parameters studied
Table 9 presents the Pearson correlation matrix between the traits studied. In the genetic variability study of accessions by
Khan et al., 2020, it was concluded that a strong positive correlation (0.75 <r <1) was found for the seed weight and pod weight traits. A similar variation in these traits was reported by
Goli et al. (1995), which confirms our results on the observation of strong and significant correlations between the weight of 100 seeds and the weight of 100 pods (r=0.956; p<0.01). This measurement of the weight of 100 seeds was considered an essential tool for evaluating morpho-physiological traits related to yield (
Mohammed et al., 2014).
Kirham et al., 1980 noted that in durum wheat, the TDE increases as the surface area expands. A strong correlation was observed between germination time and leaf length (r = 0.736, p<0.01). Thus, the earliest accessions have the longest and widest leaves, indicating that the earlier they germinate, the more organic matter they accumulate. Additionally, the larger the leaf surface area, the higher the growth rate and yield.
Cherfia et al. (2010) demonstrate a significant relationship between leaf area and yield. Indeed, the greater the leaf area, the higher the yield.
Patel et al. (2025) found that larger leaflets can exploit more sunlight, resulting in greater biomass accumulation and higher yields.
Araus et al., (1998) report that selecting for large leaf area improves grain yields.
Belkharchouche et al. (2009) report that leaf area determines drought resistance since high leaf areas lose more water than low leaf areas.