1. Introduction
Cowpea [
Vigna unguiculata (Linn.) Walp.] (2n = 2x = 22), which belongs to the family Leguminosae [
1], is one of the most important pulse crops because its seeds are rich in proteins, minerals, and vitamins, making them a valuable component of food and feed [
2,
3,
4]. Cowpea is native to Africa and was traditionally cultivated in tropical and subtropical regions. Its ability to fix atmospheric nitrogen makes cowpea useful for intercropping with other species and may help to explain its relatively broad distribution [
5,
6]. According to FAO data (
https://www.fao.org/faostat/en/#data/QCL (accessed on 1 September 2022)), the total cowpea planting area has fluctuated slightly and the total cowpea yield has increased somewhat in recent years, reaching approximately 10 million tons (from 15 million hectares) in 2021 [
7].
Frequent climate changes result in environmental stresses, especially drought and heat, that seriously affect plant growth [
8]. For example, previous studies showed that in the dry season, the yield of six main crops can decrease by 0.6–10% with every 1 °C increase in temperature [
9,
10]. Cowpea can grow in fertile [
11] or saline soil [
12], while also tolerating high temperatures and drought conditions [
13], making it a popular choice for crop cultivation in saline-alkaline regions as well as for improving soil conditions (e.g., soil adversely affected by natural disasters). Cowpea may help to address the major challenges to food and nutrition security and agricultural sustainability due to global climate changes [
14,
15]. However, being a typical short-day species, most cowpea cultivars are sensitive to the photoperiod and temperature. Additionally, the introduction of cowpea to new regions or its cultivation in different seasons may lead to a prostrate growth habit and abnormal or incomplete maturation before the next planting season or the onset of winter, resulting in a poor mechanical harvest and low yields.
In China, cowpea is a minor crop that has traditionally been grown in the mountainous regions, usually as a monoculture in terraced fields and gardens or intercropped with other crops and fruit trees, such as peanut [
16], cherry [
17], and apple trees [
18]. In 2021, the total cowpea cultivation area in China was approximately 15,000 hectares and the total yield was about 15,000 tons, which were 10% and 44% increases from the corresponding data in 2016 [
7].
The effects of the temperature and photoperiod on cowpea growth remain relatively unknown. Guangxi province, which is located in southern China, has light and temperature conditions that are suitable for cowpea growth from spring to late autumn. Accordingly, there has been a recent rapid increase in the production of this crop in Guangxi. However, the photoperiod and temperature vary greatly among seasons, which influences the adaptability of different varieties [
19,
20]. Hence, assessing the adaptability of cowpea lines sown on different dates may be useful for optimizing cowpea production and the associated economic benefits for local farmers.
In a previous study, we developed a set of new cowpea lines via hybridizations and evaluated these lines in Guangxi in the summer of 2019 and 2020. Most of the tested lines completed their life cycles and were relatively productive. In the current study, we selected 41 lines and examined their growth in different seasons (spring, summer, and autumn) in Wuming (Guangxi) to analyze their adaptability and utility across seasons. The findings of this study may provide the theoretical basis for improving cowpea germplasm and may also enable local farmers to select ideal varieties for southern China or other global regions with similar environmental conditions.
3. Results
3.1. Climate Conditions in Wuming
Wuming, which is a county in southern China, has a humid subtropical monsoon climate with lots of sunshine and rainfall and an annual average temperature of approximately 21.7 °C. The temperature, rainfall, and photoperiod varied among seasons. From April to October, the day length initially increased, peaking in mid-to-late June, and then decreased (
Figure 1C). Changes in the daily maximum and minimum temperatures were not obvious, but the temperatures were generally the highest between June and September (
Figure 1A). Rainfall was the greatest in June and October (
Figure 1B).
3.2. Analysis of Phenotypic Traits across Sowing Seasons
The analysis of the 41 cowpea lines revealed that the ranges in GP, PH, NBP, NPP, PL, and PY were wider for SpS than for the other two sowing seasons. In contrast, the ranges for NSP and 100-SW were widest for SuS. For most traits, the ranges were smallest for AuS (
Table 2).
The mean GP and PH values for the 41 cowpea lines were lower for SuS (GP: 51 days and PH: 57.8 cm) than for AuS (GP: 59.7 days and PH: 62.4 cm) and SpS (GP: 69.2 days and PH: 103.7 cm). Similar trends were observed for the mean NPP, NSP, PL, and PY values. The mean NBP value was highest for SuS (3.4), followed by SpS (3.1) and AuS (2.5). The mean 100-SW value was highest for AuS (14.0 g), followed by SpS (13.0 g) and SuS (12.5 g). There were no obvious differences in PW among the sowing seasons (
Table 2).
The coefficient of variation (CV) for GP, NSP, PL, PW, and 100-SW was low across the sowing seasons (CV < 20%), which was in contrast to the relatively high CV for NBP and PY (CV > 20%). Moreover, the CV for NBP was slightly higher for SpS (24.7%) than for SuS (21.4%) and AuS (23.5%), whereas the CV for PY was higher for SuS (27.0%) than for SpS (25.7%) and AuS (23.0%). The CV for PH was lowest for SuS (13.7%), while the CV for NPP was lowest for AuS (14.6%) (
Table 2).
The ANOVA results (
Table 3) indicated there was no significant block effect on the cowpea traits, with the exception of PH (
p < 0.001), PW (
p < 0.05), and 100-SW (
p < 0.05). However, the genotype (G) and sowing date (D) significantly affected all traits (
p < 0.001). The G × D interaction also significantly affected all traits (
p < 0.001 and
p < 0.01).
The correlation analysis showed GP, PH, PL, PW, and 100-SW were highly correlated across SpS, SuS, and AuS (
p < 0.01). In addition, NBP and PW were also highly correlated, but only for SpS and AuS (
p < 0.01). For NNP, NSP, and PY, there were no obvious correlations across the sowing seasons (
Table 4).
3.3. Variations among Lines across Sowing Seasons
The analysis of the data for each line (
Table S1) revealed GP was most consistent for line BJ–28, followed by BJ–29 and ZJ–1, whereas GP varied considerably for BJ–20, BJ–21, and BJ–-18. The CV for PH across three sowing seasons was less than 20% for BJ–3, BJ–38, BJ–25, BJ–-1, BJ–22, BJ–4, BJ–35, BJ–19, and BJ–13. In contrast, PH differed substantially across sowing seasons for BJ–33, BJ–15, BJ–10, BJ–27, and BJ–28. Among the examined lines, the NBP, NSP, PL, PW, and 100-SW values were stable among the sowing seasons for BJ–13, BJ–19, BJ–25, BJ–28, BJ–29, and BJ–30. Lines BJ–3 and BJ–4 had stable traits (except for NBP). Similarly, with the exception of PH, BJ–23 also had stable traits. All examined traits were stable for BJ–22. In addition, the lines with a highly variable PH also had unstable NBP and PY values (e.g., BJ–27, BJ–28, BJ–29, BJ–33, and GX18–21). Overall, the growth performance of BJ–3, BJ–4, BJ–22, and BJ–23 was relatively unchanged across the three sowing seasons.
3.4. Principal Component Analysis of Traits
According to the PCA, the first three principal components contributed 72.47% of the total variability in the examined traits (
Table 5). The first component (PC1) explained 38.06% of the variation and was positively correlated with GP (0.90), PH (0.80), PW (0.80), and PY (0.59). The second component (PC2) explained 21.93% of the variation and was positively correlated with NPP (0.65), NSP (0.59), PL (0.57), and PY (0.68). The third component (PC3) explained 12.48% of the variation and was correlated with NPP (−0.50), NSP (0.41), and PL (0.65).
3.5. AMMI Model Analysis
The
AMMI biplot of the average yield (
Figure 2) suggested the selected lines were high-yielding cowpea
varieties. The y-axis of the AMMI biplot presents the effect of the interaction between varieties and sowing seasons. A low absolute value for PCA1 reflects a small interaction effect (i.e., a relatively stable variety across sowing seasons). The average yield was highest for G41 (GX18–21), followed by G19 (BJ–19) and G27 (BJ–27). Moreover, the effect of the interaction between the variety and the sowing season was the greatest for G27 (BJ–27). Among the three sowing seasons, the average yield was highest and lowest for SpS (E1) and SuS (E2), respectively. In the AMMI biplot, most of the lines were concentrated at E3 (i.e., AuS), suggesting they were appropriate for autumn sowing.
3.6. Selection of Cowpea Lines Suitable for Local Production
According to the yields for the three sowing seasons, 27 lines with high yield potentials were selected for SpS (>250 kg), SuS (>120 kg), and AuS (>200 kg). Of these lines, BJ–8 and BJ–19 had high yield potentials across the sowing seasons. Lines BJ–20, BJ–25, BJ–26, BJ–32, and GX18–21 were suitable for both spring and summer sowing. Lines BJ–13, BJ–14, BJ–15, BJ–28, and BJ–33 were suitable for spring and autumn sowing. Only BJ-18 was suitable for summer and autumn sowing.
Some of the lines with high yield potentials exhibited a prostrate growth habit, which is not conducive for mechanical harvesting. These lines included BJ–11, BJ–13, BJ–15, BJ–18, BJ–19, BJ–-38, and GX18–21. Additionally, the high GP values for BJ–6, BJ–7, BJ–13, BJ–17, and BJ–20 might affect the growth of the intercropped species. Specifically, BJ–20 had a GP that exceeded 76 days in autumn, making it relatively susceptible to powdery mildew infections, which typically occur during the late cowpea growth stage.
Finally, we selected nine elite lines useful for cowpea production in three sowing seasons (
Table 6).
4. Discussion
Frequent climate changes might lead to crop yield losses. Cowpea can withstand various environmental stresses, such as drought and high temperatures, and it can grow in low-quality soils. Hence, it is considered to be an important crop for dealing with natural disasters [
5], especially in the mountainous regions of China, where the soil quality is low and droughts are common. There have recently been considerable advances in cowpea breeding in China [
23]. Additionally, a set of elite lines has been developed and evaluated in different eco-regions [
19]. The climate in Guangxi province is appropriate for cowpea cultivation from spring to autumn. Cowpea is often intercropped with young fruit trees or sugarcane in Guangxi. However, the seasonal variations in the photoperiod and rainfall in this part of China affect the growth, development, and productivity of cowpea, which is sensitive to light and temperature. Therefore, the adaptability of new cowpea lines should be investigated to determine which seasons are ideal for their cultivation; similar evaluations have been conducted for many other crops [
24,
25,
26].
In this study, there were substantial morphological variations in the cowpea lines across the three sowing seasons. Moreover, there were significant differences in the main phenological, morphological, and yield-related traits, such as GP, PH, NPP, NSP, and PY, among the sowing seasons, implying these traits were sensitive to changes in the photoperiod and temperature.
Flowering, which is an important event during the life cycle of most crops, is controlled by environmental conditions (e.g., temperature and day length) [
27]. The meteorological data for Wuming indicated the temperature is generally lower in spring and autumn than in summer. Low temperatures can delay flowering, thereby prolonging the vegetative phase. Accordingly, delayed sowing can shorten the crop growth period [
28,
29]. In the present study, cowpea plants flowered in mid-late-May (SpS), late July (SuS), and early September (AuS). Because the temperature was lower in May than in July and September, the flowering time was shorter for SpS than for SuS and AuS. Similarly, the GP values for the 41 examined cowpea lines were higher in spring than in summer and autumn, but the CV for GP was low among the sowing seasons. Previous studies confirmed that the vegetative phase is prolonged when crops are sown early [
29,
30]. In the current study, the delayed flowering and reproductive growth of cowpea explains why PH was highest for SpS. In addition, the differences in rainfall, which also influences vegetative growth, may help to explain why the GP value was lower for SuS than for AuS and the PH value was higher for SuS than for AuS.
Early sowing enables crops to fully utilize soil moisture and nutrients during relatively long growing seasons, which may lead to increased crop productivity [
29,
31,
32]. In our results, the PY value was higher for SpS than for SuS and AuS, which is consistent with the results of previous studies [
33,
34]. However, the PY value was lower for SuS than for AuS, likely because of the higher temperatures in summer. Several studies determined that high temperatures could accelerate the phenological period of crops, with detrimental effects on yield [
35,
36,
37]. Additionally, the rainfall in Wuming occurs mostly from June to October. Thus, the high temperatures and humidity levels during this period adversely affect cowpea pollen viability, causing flowers and pods to fall, which ultimately leads to yield losses [
38]. Although the cowpea yield was highest for SpS, if cowpea is sown too early, several critical environmental factors (e.g., abrupt temperature decreases) may negatively modulate germination and seedling growth.
The AMMI model may be combined with ANOVA and PCA [
39] to elucidate the effect of the interaction between the genotype and environment on specific traits [
40]. Because of the intuitiveness of the associated analysis, the AMMI model has been widely used to in rice [
41], wheat [
42], maize [
43], and cotton [
44] yields. According to the AMMI biplot presented herein, lines GX18–21, BJ–19, and BJ–27 may produce high yields in Guangxi. Furthermore, sowing in spring may optimize cowpea production in this part of China. However, the high PH values in spring are unfavorable for mechanical harvesting. The AMMI biplot also indicated that most of the analyzed cowpea lines are suitable for growth in autumn. Because of their sensitivity to the photoperiod and temperature, the suitability of several cowpea lines (e.g., BJ–20, BJ–36, BJ–37, and GX18–21) for specific environmental conditions must be assessed before they are introduced to new regions. The wide adaptability of some of the lines (e.g., BJ–22), which was reflected by the relatively stable traits across sowing seasons, should be exploited by cowpea breeding programs.
Finally, we selected nine lines that satisfied the following criteria: high yield, short growth period, and erect growth habit. These criteria were selected because yield is the key index for evaluating the production potential of crops, but an erect growth habit is critical for the mechanical harvesting systems used in modern agriculture. However, we analyzed the lines for only 1 year. Thus, additional tests will need to be conducted before these elite lines are cultivated over a wide region. Nevertheless, we believe these new cowpea lines will be useful for increasing cowpea production in Guangxi and in other regions with similar environmental conditions.