Screening cotton cultivars for low-phosphorus tolerance: a comparison of hydroponic and field methods

Abstract

Background

Soil available phosphorus (AP) deficiency significantly limits cotton production, particularly in arid and saline-alkaline regions. Screening cotton cultivars for low phosphorus (P) tolerance is crucial for the sustainable development of cotton production. However, the effect of different growth media on the screening outcomes remains unclear. To address this, we evaluated the low P tolerance of 25 cotton cultivars through hydroponic culture at two P levels (0.01 and 0.5 mmol•L⁻¹ KH₂PO₄) in 2018 and field culture with two P rates (0 and 90 kg•hm⁻², in P₂O₅) in 2019.

Results

In the hydroponic experiments, principal component analysis (PCA) showed that shoot dry weight (SDW) and P utilization efficiency in shoots (PUES) of cotton seedlings explained over 45% of the genetic variation in P nutrition. Cotton cultivars were subjected to comprehensive cluster analysis, utilizing agronomic traits (SDW and PUES) during the seedling stage (hydroponic) and yield and fiber quality traits during the mature stage (in field). These cultivars were grouped into four clusters: resistant, moderately resistant, moderately sensitive, and sensitive. In low P conditions (0.01 mmol•L⁻¹ KH₂PO₄ and 4.5 mg•kg⁻¹ AP), the low-P-resistant cluster showed significantly smaller reductions in SDW (54%), seed cotton yield (3%), lint yield (− 2%), fiber length (− 1%), and fiber strength (− 3%) compared with the low-P-sensitive cluster (75%, 13%, 17%, 7%, and 9%, respectively). The increase in PUES (299%) in the resistant cluster was also significantly higher than in the sensitive cluster (131%). Four of the eight low-P-tolerant cotton cultivars identified in the field and six in the hydroponic screening overlapped in both screenings. Two cultivars overlapped in both screening in the low-P-sensitive cluster.

Conclusion

Based on the screenings from both field and hydroponic cultures, ZM-9131, CCRI-79, JM-958, and J-228 were identified as low-P-tolerant cotton cultivars, while JM-169, XM-33B, SCRC-28, and LNM-18 were identified as low-P-sensitive cotton cultivars. The relationship between field and hydroponic screening results for low-P-tolerant cotton cultivars was strong, although field validation is still required. The low P tolerance of these cultivars was closely associated with SDW and PUES.

Background

China is a major producer of cotton (Gossypium hirsutum L.) and a key player in the global cotton textile industry. The cotton industry contributes significantly to the economic welfare of farmers and provides essential raw materials for the textile industry, playing a vital role in the national economy (Feng et al. 2024). In 2023, seed cotton yield in China reached 5.6 million tons (National Bureau of Statistics of China 2024). To ensure higher lint yield and better fiber quality, cotton plants require substantial phosphorus (P) nutrition throughout their growth period (Chen et al. 2020; Sun et al. 2023). Proper application of P fertilizer (76–151 kg•hm⁻², in P₂O₅) can increase dry matter accumulation and P uptake, thereby improving cotton yield (Chen et al. 2020). Cotton production in China is concentrated in Xinjiang, and most in saline-alkali lands, beaches, and sandy dry lands (Mao 2019). The soils in these regions are predominantly calcareous, which strongly fixates P (Ahmad et al. 2013; Schachtman et al. 1998). Combined with low rainfall and drought, soil-available P (AP) deficiency is a pressing issue (Peng et al. 2024).

P deficiency is a major abiotic stress factor limiting cotton yield (Sun et al. 2022) and fiber quality (Huang et al. 2024; Sun et al. 2023). Furthermore, global P fertilizer consumption is increasing by 2.2% annually (Nowaki et al. 2017), and the worldwide accessible P deposits are expected to be depleted in less than a century (Cordell et al. 2009). In response, the Chinese government has started promoting reducing chemical fertilizer use and enhancing P utilization efficiency (PUE) by excavating the potential of soils and plants (Zhou et al. 2016). Low P tolerance can be assessed by evaluating the impact of reduced P fertilizer on cotton growth, yield, and fiber quality (Li et al. 2020). Screening cotton germplasm resources with low P tolerance and efficient P uptake and utilization capacities can reduce P fertilizer application and promote sustainable development of the cotton industry.

Low P stress typically promotes the preferential distribution of photosynthetic products to the roots, accelerating root growth while inhibiting the growth of the above ground parts (Marschner et al. 1987). Prolonged exposure to low P stress induces changes in shoot morphology, root traits, physiological and biochemical indexes, and gene expression, allowing the plant to adapt to the stress and optimize its growth and development (Niu et al. 2012; Wang et al. 2008; Zhou et al. 2016). Iqbal et al. (2019) evaluated 12 cotton cultivars, and they found that Xinluzao 49 and Xinluzhong 48 exhibited low P tolerance under hydroponic conditions, using parameters such as root biomass, aboveground biomass, and leaf photosynthetic characteristics. After 2 years of hydroponic experiments, Li et al. (2020) classified 16 cotton cultivars based on indicators such as leaf biomass, individual plant biomass, and P accumulation (PA) in roots. They identified two cotton cultivars with significant differences in low P tolerance (Yuzaomian 9110, a tolerant cultivar, and Lu 54, a sensitive cultivar). However, due to a lack of standardized indicators and methods for evaluating low P tolerance in cotton (Iqbal et al. 2019; Li et al. 2020), assessing and selecting low-P-tolerant cotton cultivars remain challenging. Most studies focus on hydroponic screenings, primarily assessing seedling or bud stages (Ahmad et al. 2001; Iqbal et al. 2019; Li et al. 2020). However, both hydroponic and field cultures have their limitations. In hydroponics, roots lack mechanical support, and achieving precise P content in the soil during field cultivation can be challenging. Few studies have comprehensively analyzed the low P tolerance of cotton seedlings with their yield and fiber quality at the harvest stage. In this study, we performed phenotypic analysis on 25 cotton cultivars using both field and hydroponic cultures to screen for low-P-tolerant cotton cultivars. Additionally, we examined whether growth mediums affect the phenotypic differences in low P tolerance.

Materials and methods

Experimental site and materials

The field and hydroponic experiments were conducted at Institute of Cotton Research of Chinese Academy of Agricultural Sciences test station in Anyang, Henan Province, China (36°06’ N, 114°21’ E). A total of 25 upland cotton (Gossypium hirustum L.) cultivars (Table S1) widely planted in the cotton growing areas of the Yellow River Basin and Xinjiang in China were used. The key meteorological data during the cotton growing period in 2019 are shown in Fig. S1.

Hydroponic experiment

The hydroponic experiment was conducted in 2018 in a plant growth cabinet. A split-plot design was used, with the 25 cotton cultivars as the main plot and P levels as the subplot. The temperature was controlled at 28 ± 2℃ for 16 h during the day and 20 ± 2℃ for 8 h at night, with a relative humidity of 60%. Cotton seeds with fully intact embryos were sterilized with 2% hydrogen peroxide and washed with distilled water. These seeds were then placed in incubators for sand culture. Once the cotyledons flattened, cotton seedlings with similar growth were selected and transferred to 8 L incubators (32 cm × 25 cm × 14 cm) containing 1/2 Hoagland nutrient solution, which was replaced every 7 d. Oxygen was supplied continuously for 24 h every day. After 7 d, the nutrient solution was changed to a full Hoagland solution (comprising 0.02 mmol•L⁻¹ H₃BO₃, 1 μmol•L⁻¹ ZnSO₄•7H₂O, 0.2 μmol•L⁻¹ CuSO₄•5H₂O, 1 μmol•L⁻¹ MnSO₄•H₂O, 5 nmol•L⁻¹(NH₄)₆Mo₇O₂₄•4H₂O, 2.5 mmol•L⁻¹ Ca(NO₃)₂•4H₂O, 1 mmol•L⁻¹ MgSO₄•7H₂O, 0.1 mmol•L⁻¹ C₁₀H₁₂N₂NaFeO₈, and 2.5 mmol•L⁻¹ KCl), with low P (KH₂PO₄ 0.01mmol•L⁻¹) and suitable P (KH₂PO₄ 0.5 mmol•L⁻¹) treatments. Each treatment had three replicates, with 45 cotton plants per treatment and 15 cotton plants per incubator. The cotton plants were harvested 30 d after transplantation.

Field experiment

The field experiment, conducted in 2019, tested the same 25 cotton cultivars used in the hydroponic experiment with three replicates. The P fertilizer location test conducted in the experimental site from 2015 to 2018 showed significant differences in soil AP under different P fertilizer applications, facilitating the screening of low-P-tolerant cotton cultivars (Schulte et al. 2007; Seth et al. 2018). The type of soil in the experimentwas classified as inceptisols (USDA Soil Taxonomy) and soil was clay loam (Li et al. 2017). The soil nutrient content of the 0–20 cm soil layer is presented in Table 1. A split-plot design was employed with cultivars as the main plot and P rates (0 and 90 kg•hm⁻² P₂O₅) as the subplot. The total nitrogen (N) rate in each plot was 225 kg•hm⁻², applied as urea (46% N). Basal fertilizers included 50% of the total N rate, the total P rate as triple superphosphate (44% P₂O₅), and 150 kg•hm⁻² potassium sulfate (51% K₂O). Top application included the remaining 50% of the N fertilizer, was applied at the start of the flowering period (Li et al. 2017). All phosphate fertilizers were evenly applied manually. Cotton seeds were sown using a manual drilling machine after plastic film mulching on May 1, with a row spacing of 80 cm. Each plot contained four rows, covering an area of 28.8 m² with a plant density of 67 500 plants•hm⁻². All field managements were carried out following high-yield cultivation practices for cotton.

Table 1 Nutrition properties of 0–20 cm soil layer in the field experiment

Dry weight and P efficiency indexes

Five plants with uniform growth were selected from each replicate, retaining the entire root system. After washing off the nutrient solution, the plants were separated into roots, stems, and leaves. Each part was dried at 105 ℃ for 30 min and then at 70 ℃ to a stable weight before weighing (Fig. 1). The samples were ground using a disintegrator (FS-II, Zhejiang Top Yunnong Technology Co., Ltd, China) with a 0.5 mm sieve. P content was measured using the colorimetric method (Yang et al. 1990) with an ultraviolet spectrophotometer (SPECORD 40, JENA, Germany). The equations for key indicators are as follows:

• (1) P accumulation (PA) = P content × dry weight of the cotton plant (Sun et al. 2018);

• (2) P utilization efficiency (PUE) = (Cotton plant dry weight/PA of whole plant) × 100% (Sun et al. 2018);

• (3) Relative value = Measured value under low P level/Measured value under suitable P level.

Fig. 1

Schematic diagram of measurement indicators for hydroponic and field experiments

Morphological characteristics

Five plants with uniform growth were selected from each replicate to measure plant height (PH) using a ruler (Fig. 1). The number of leaves (NL) was counted manually, and leaf area (LA) was measured with a portable LA meter (CI-203, CID, USA) from the selected cotton plants.

Net photosynthetic rate and SPAD value

From 9:00 to 11:00 a.m., the net photosynthetic rate of the fourth cotton leaf was measured using a portable photosynthetic apparatus (LI-6400XT, LI-COR Biosciences, NE; Li-Cor, Inc., Lincoln, NE, USA) (Fig. 1). Chlorophyll content in the fourth cotton leaf was determined using a portable chlorophyll meter (SPAD 502, Minolta Corporation, Japan).

Yield and fiber quality

Fifty cotton bolls without boll walls were harvested manually from each experimental plot. The boll weight was calculated based on the number and weight of cotton bolls. Lint percentage was determined by weighing the seeds and fibers after ginning. Seed cotton yield was harvested manually during the mature stage, and lint yield was calculated based on the actual seed cotton yield and lint percentage. The fibers were dried at 35 ℃ to a stable weight and then taken to Cotton Quality Inspection and Testing Center of the Ministry of Agriculture and Rural Affairs, China (Anyang), for assessments of fiber length and strength, micronaire, uniformity, and elongation.

Statistical analysis

Data analysis was performed using Excel 2007 (Microsoft Corp., Redmond, USA) to determine the means, standard errors, and coefficient of variation (CVs, %). Variance analysis was conducted using Statistical Package for Social Sciences V17.0 (IBM Corp., New York, USA), with the least significant difference used to compare means at the 5% probability level. Principal component analysis (PCA) and cluster analysis were performed using Origin 2018 software.

Results

Evaluation of cotton cultivars for low phosphorus tolerance in hydroponic culture

Analysis of variance revealed that P treatment (P), cultivar (C), and the P × C interaction had significant effects (P < 0.01) on various traits of cotton cultivars (Table 2). Notably, P treatment did not significantly affect SPAD values of the cotton cultivars. Compared with optimal P levels, the root dry weight (RDW), shoot dry weight (SDW), PA in the root and shoot (PAR and PAS), PH, NL, LA, and net photosynthetic rate (P_n) of cotton cultivars under low P conditions were significantly decreased (P < 0.05). Conversely, the root-to-shoot ratio (RSR), and P utilization efficiency in the root and shoot (PUER and PUES) significantly increased (P < 0.05) under low P conditions. Specifically, the dry weight of roots and shoots decreased by 20.0% and 58.1%, respectively, and PAR and PAS decreased by 84.8% and 85.4%, respectively, under low P conditions compared with optimal conditions. However, P utilization efficiency in the root and shoot increased by 428.6% and 184.4%, respectively.

Table 2 The effects of phosphorus (P) level and cultivar (C) on agronomic traits of cotton seedlings in the hydroponic experiment

PCA was used to analyze the growth responses of cotton cultivars to varying P levels, considering 12 indicators across 25 cotton cultivars (Table S2). The analysis of cultivar (C) and trait interactions revealed that the first and the second principal components (PC1 and PC2) explained 32.5% and 19.4% of the variability, respectively (Fig. 2). In PC1, the weighted coefficients for SDW, RDW, and LA were 46.4%, 40.2%, and 38.5%, respectively. PC2 was primarily associated with PUES, PAR, and P_n, whose weighted coefficients were 45.8%, 37.3%, and 29.6%, respectively. Biplot analysis under P treatments showed that SDW and PUES accounted for leading genetic variability. Therefore, cluster analysis was performed using the relative values of SDW and PUES (Fig. 3).

Fig. 2

The biplot of the relationship between 12 agronomic traits and 25 cotton cultivars. The serial number of cotton cultivars are listed in Table S1. RDW, root dry weight; SDW, shoot dry weight; RSR, root-to-shoot ratio; PAR, P accumulation in root; PAS, P accumulation in shoot; PUER, P utilization efficiency in root; PUES, P utilization efficiency in root; PH, plant height; NL, the number of leaf plant⁻¹; LA, leaf area; SPAD, SPAD value; P_n, net photosynthetic rate

Fig. 3

Clustering of cotton cultivars based on the relative values of SDW and PUES. Cluster 1 represents resistant cultivars; Cluster 2 represents moderately resistant cultivars; Cluster 3 represents moderately sensitive cultivars; Cluster 4 represents sensitive cultivars

According to the cluster analysis, cotton cultivars were classified into four categories (Fig. 3): resistant (cluster 1, average relative SDW 0.46 and PUES 3.99), moderately resistant (cluster 2, average relative SDW 0.42 and PUES 3.43), moderately sensitive (cluster 3, average relative SDW 0.42 and PUES 2.52), and sensitive (cluster 4, average relative SDW 0.25 and PUES 2.31). Among the 25 cotton cultivars, the resistant, moderately resistant, moderately sensitive, and sensitive clusters comprised 8, 8, 7, and 2 cultivars, respectively. Under low P conditions, SDW for the resistant and moderately resistant clusters decreased by 54% and 58%, respectively, compared with optimal P. In comparison, SDW in the moderately sensitive and sensitive clusters decreased by 58% and 75%, respectively (Fig. 4). However, under low P, the PUES of the resistant, moderately resistant, moderately sensitive, and sensitive clusters increased by 299%, 243%, 152%, and 131%, respectively. Significant differences were observed across these groups in RDW, SDW, RSR, PAR, PAS, PUER, PUES, PH, NL, LA, SPAD, and P_n (Fig. 4).

Fig. 4

Agronomic traits of cotton seedlings as affected by P level (P) and cultivar (C)

Evaluation of cotton cultivars for low phosphorus tolerance in field culture

Analysis of variance indicated that both P treatment and cultivar significantly affected (P < 0.01) boll weight, boll number per plant, seed cotton yield, and lint yield, with cultivar (C) also significantly affecting lint percentage (Table 3). Only boll weight was significantly affected (P < 0.01) by the P × C interaction. Compared with treatment of 90 kg•hm⁻² P₂O₅, reductions in boll weight, boll number per plant, seed cotton yield, and lint yield at 0 kg•hm⁻² P₂O₅ were 8.5%, 7.6%, 8.4%, and 9.2%, respectively, while the reduction in lint percentage was only 0.7%.

Table 3 The effects of phosphorus (P) level and cultivar (C) on yield traits of cotton during mature stage

Cotton cultivars were classified into four groups (Fig. 5) through cluster analysis based on yield-related indicators such as boll weight, lint percentage, boll number per plant, seed cotton yield, and lint yield. These groups included resistant (cluster 1, average relative seed cotton yield 0.97 and lint yield 1.02), moderately resistant (cluster 2, average relative seed cotton yield 0.94 and lint yield 0.93), moderately sensitive (cluster 3, average relative seed cotton yield 0.90 and lint yield 0.90), and sensitive (cluster 4, average relative seed cotton yield 0.87 and lint yield 0.83). Among the 25 cotton cultivars, the resistant, moderately resistant, moderately sensitive, and sensitive clusters contained 1, 12, 8, and 4 cultivars, respectively. Seed cotton yield for the resistant and moderately resistant clusters decreased by 3% and 6%, respectively, at treatments of 0 kg•hm⁻² P₂O₅, compared with treatments of 90 kg•hm⁻² P₂O₅. In comparison, the moderately sensitive and sensitive clusters decreased by 10% and 13%, respectively (Fig. S2). Similarly, lint yield in the resistant, moderately resistant, moderately sensitive, and sensitive clusters decreased by − 2%, 7%, 10%, and 17%, respectively, under treatments of 0 kg•hm⁻² P₂O₅. Boll weight, lint percentage, and boll number per plant followed a similar trend (Fig. S2).

Fig. 5

Clustering of cotton cultivars based on the relative values of yield and yield components. Cluster 1 represents resistant cultivars; Cluster 2 represents moderately resistant cultivars; Cluster 3 represents moderately sensitive cultivars; Cluster 4 represents sensitive cultivars

The length, strength, and elongation of fiber were significantly affected (P < 0.05) by P treatment, cultivar, and the P × C interaction (Table 4). These traits decreased by 2.0%, 1.7%, and 1.5%, respectively, at treatments of 0 kg•hm⁻² P₂O₅ compared with treatments of 90 kg•hm⁻² P₂O₅. Cluster analysis based on fiber quality parameters (length, strength, uniformity, elongation, and micronaire) also classified cultivars into four categories (Fig. 6). These groups included resistant (cluster 1, average relative fiber length 1.01 and fiber strength 1.03), moderately resistant (cluster 2, average relative fiber length 0.98 and fiber strength 1.00), moderately sensitive (cluster 3, average relative fiber length 0.96 and fiber strength 0.94), and sensitive (cluster 4, average relative fiber length 0.93 and fiber strength 0.91). Among the 25 cotton cultivars, the four clusters contained 6, 4, 11, and 4 cultivars, respectively. Fiber length in the resistant and moderately resistant clusters decreased by − 1% and 2%, respectively, at treatments of 0 kg•hm⁻² P₂O₅ compared with treatments of 90 kg•hm⁻² P₂O₅, while the moderately sensitive and sensitive clusters showed decreases of 4% and 7%, respectively (Fig. S3). Similarly, fiber strength decreased by 3%, 0%, 6%, and 8% in the resistant, moderately resistant, moderately sensitive, and sensitive clusters, respectively, under treatments of 0 kg•hm⁻² P₂O₅. Micronaire, uniformity, and fiber elongation showed similar trends (Fig. S3).

Table 4 The effects of phosphorus (P) level and cultivar (C) on fiber quality traits of cotton during mature stage

Fig. 6

Clustering of cotton cultivars based on the relative values of fiber quality traits. Cluster 1 represents resistant cultivars; Cluster 2 represents moderately resistant cultivars; Cluster 3 represents moderately sensitive cultivars; Cluster 4 represents sensitive cultivars

Comprehensive evaluation of low P tolerant cotton cultivars

The clustering results of agronomic traits, yield traits, and fiber quality traits under both hydroponic and field cultures were analyzed comprehensively to classify the low P tolerance of cotton cultivars (Fig. 7). Among the 25 cotton cultivars, the low-P-resistant cluster was identified through screening across both mediums, including ZM-9131, CCRI-79, JM-958, and J-228. These overlapping cultivars represented 50% of the cultivars in both hydroponic and field experiments. Compared with suitable P levels or 90 kg•hm⁻² P₂O₅, the average SDW of these overlapping cultivars decreased by 51.7% under low P conditions in hydroponic experiments. In field experiments under 0 kg•hm⁻² P₂O₅, the average seed cotton yield, lint yield, fiber length, and fiber strength decreased by 9.0%, 6.5%, 0.6%, and 1.8%, respectively. In the low-P-sensitive cluster, the proportion of overlapping cultivars in both hydroponic and field experiments was 100%, consisting of JM-169, XM-33B, SCRC-28, and LNM-18. Compared with the low-P-resistant cultivars, low-P-sensitive cultivars exhibited greater reductions in average SDW, seed cotton yield, lint yield, fiber length, and fiber strength decreased by 89.5%, 13.2%, 16.8%, 6.9%, and 8.8%, respectively. Similarly, the overlapping cultivars in the moderately resistant and moderately sensitive clusters in hydroponic and field experiments accounted for 37.5% and 75.0%, respectively, with 42.9% and 75.0% in the field trials. Notably, some cultivars classified as moderately sensitive in hydroponic experiments, such as CCRI-12, CCRI-110, and CCRI-112, were identified as moderately resistant (yield traits) and resistant clusters (fiber quality traits) in field experiments.

Fig. 7

Comprehensive evaluation of low phosphorus tolerant cotton cultivars based on cluster analyses

Discussion

P is a critical component of many important compounds in cotton plants and plays a key role in various metabolic processes (Singh et al. 2014). P deficiency directly impacts the morphological development, biomass, and yield of plants (Chen et al. 2020; He et al. 2017; Singh et al. 2006). Different crop species and cultivars exhibit varying responses to various levels of P supply (Bilal et al. 2018; He et al. 2017; Rose et al. 2010). Evaluating cotton germplasm for low P tolerance is crucial for breeding cultivars with higher P efficiency and/or enhanced low P tolerance. The seedling stage is a critical period for establishing the cotton population. Furthermore, screening low-P-tolerant cultivars at the seedling stage is advantageous because it is faster and allows for a larger sample size (Ahmad et al. 2001; Iqbal et al. 2019). Dong (2007) reported that the peak of P absorption in cotton occurs during the flowering and boll stages, as P uptake accounts for over 60% of the total P absorbed during the entire growth period, much higher than during the seedling stage. Due to the short duration of the seedling stage, PA in cotton seeds may play a role, potentially influencing the screening results. Therefore, it is essential to assess the low P tolerance of cotton cultivars under various mediums and growth periods in future studies. Moreover, because of the complex physical and chemical properties of soil, the performance of low-P-tolerant cultivars in hydroponic experiments may not consistent with field trial outcomes (Hayes et al. 2004). Therefore, conducting field tests is necessary to identify low-P-tolerant cotton cultivars accurately.

This study demonstrated that SDW and PUES in hydroponic experiments accounted for over 45% of the genetic variation in P nutrition. Under P deficiency conditions, cotton cultivars in the low-P-resistant cluster showed smaller decreases in SDW but greater increases in PUES compared with the low-P-sensitive cluster (Fig. 4). These results are consistent with findings in other low-P-tolerant cultivars of cotton (Gossypium hirsutum L.) (Iqbal et al. 2019; Li et al. 2020), wheat (Triticum aestivum L.) (Soumya et al. 2021), and soybean (Glycine max L.) (Krishnapriya et al. 2016). Significant differences in PUE were observed among cotton cultivars (Iqbal et al. 2019). Cultivars with higher PUE and greater SDW accumulation under low P conditions benefit cotton production in P-deficient areas (Ahmad et al. 2001). For wheat, PUE has a more significant impact on low P tolerance than PA (Soumya et al. 2021). Therefore, SDW and PUES can be used as indicators for screening cotton cultivars with low P tolerance in hydroponic experiments (Ahmad et al. 2001; Iqbal et al. 2019; Li et al. 2020).

Under low P conditions, both seed cotton yield and lint yield decreased in cotton cultivars. The reduction in seed cotton yield and lint yield were more significantly pronounced in the low-P-sensitive cluster (13% and 17%) compared with the low-P-resistant cluster (9% and 6%) (Fig. 5). Li et al. (2022) suggested that the main reason for the reduction in lint yield under low P is the limitation in boll number. This finding was confirmed in our previous research (Sun et al. 2022). Previous studies have shown that low-P-resistant clusters in wheat (Bilal et al. 2018), soybean (Zhou et al. 2016), and Brassica oleracea (Hammond et al. 2009) exhibited smaller reductions in yield and PA compared with low-P-sensitive clusters under low P conditions. Similarly, fiber length and strength in the four clusters exhibited a similar response to low P, consistent with the findings of Li et al. (2020). Furthermore, cotton cultivars that maintain higher dry matter, yield, and fiber quality (Sun et al. 2018; Sun et al. 2022, 2023) and/or demonstrate a positive response to P deficiency (Iqbal et al. 2020) are well-suited for cultivation in low P areas (Chen et al. 2024; Pandey et al. 2024).

In hydroponic experiments, the PAR, PAS, and RDW of the four clusters were significantly reduced under low P conditions, particularly in the sensitive cluster. The SDW and PUES of cotton cultivars in the low-P-resistant cluster were strongly correlated with RDW and PAR (Iqbal et al. 2019). Fageria et al. (1999) found that, in wheat, the low-P-resistant cluster exhibited better performance in PA and RDW than the low-P-sensitive cluster under low P conditions. Similarly, the low-P-resistant cotton cultivars achieved higher yields under low P, likely due to their efficient root structure (Chen et al. 2019) and adequate P uptake (Chen et al. 2020). Thus, plants under low P stress have developed special adaptive mechanisms, such as increased root length and root hair number (He et al. 2017), root-secreted acid phosphatase (Wang et al. 2008), and enhanced leaf transpiration (Huang et al. 2017), to improve P acquisition (PA). The surface root length of fine and medium roots in low-P-tolerant cotton cultivars was a key factor influencing PA in low P conditions (Chen et al. 2019). Low P stress typically increases RDW and RSR, driven by the enlargement of root length, surface area, and volume (Zhao et al. 2004; Richardson et al. 2011; Sun et al. 2018). These physiological adaptations are critical for improving PA in response to low P. Therefore, it is essential to investigate the root characteristics and other P absorption and transport pathways in the low-P-resistant cluster of cotton cultivars (Balemi et al. 2012; Gemenet et al. 2015; Liu 2021; Wang et al. 2008).

Some of the cotton cultivars screened in this study have been previously reported (Iqbal et al. 2019; Li et al. 2020; Yuan et al. 2019). Iqbal et al. (2019) classified CCRI-64 and CCRI-12 as low-P-tolerant and medium low-P-tolerant cultivars, respectively, based on RDW, SDW, PUE, total plant P uptake efficiency, and photosynthetic characteristics. In this study, we categorized CCRI-64 and CCRI-12 as belonging to the moderately sensitive and moderately resistant clusters, respectively. Li et al. (2010) classified CCRI-35 as a moderately low-P-sensitive cotton cultivar, which is consistent with our results and those of Yuan et al. (2019). However, Li et al. (2020) classified YZM-9110 as a low-P-tolerant cultivar, while our study placed it in the moderately sensitive cluster. This discrepancy suggests that when cultivating low-P-tolerant cotton cultivars, it is crucial to select cultivars that consistently perform well across multiple regions (Irfan et al. 2019).

Previous studies have indicated a limited correlation between hydroponic and field screening of low-P-tolerant cultivars (Hayes et al. 2004; Irfan et al. 2019). However, other studies suggest that hydroponic screening provides valuable reference data for field cultivation in P-deficient areas (Meena et al. 2021; Soumya et al. 2021). For example, through hydroponic and field screening, Soumya et al. (2021) identified nine low-P-tolerant and three low-P-sensitive wheat cultivars. Similarly, Meena et al. (2021) screened 100 mungbean cultivars, finding that 33.3% of the low-P-tolerant cultivars and 25.0% of the low-P-sensitive cultivars from hydroponic screening were verified in the field. Under low P stress, the decrease in wheat biomass in hydroponic cultures was greater than in field cultures (Meena et al. 2021), likely because roots in field conditions can activate soil P by adjusting root structure and exudates. In contrast, such interactions are absent in hydroponic culture (Wang et al. 2008). Despite differences in P absorption mechanisms between hydroponic and field media, most of the low-P-resistant or low-P-sensitive cultivars in this study were also classified accordingly in field conditions.

Conclusion

This study demonstrates that hydroponic screening of low-P-tolerant cotton cultivars offers important insights for cultivation. Most of the low-P-tolerant cotton cultivars identified in hydroponic conditions also exhibited higher yields and better fiber quality in field under low P. SDW and PUES in hydroponic experiments emerged as valuable indicators for screening low-P-tolerant cotton cultivars, as they significantly impact yield and fiber quality at maturity. Moreover, the primary factor contributing to differences in low P tolerance among cotton cultivars appears to be PUE rather than P absorption efficiency. The low-P-resistant cluster (ZM-9131, CCRI-79, JM-958, and J-228) could serve as valuable germplasm resources for breeding low-P-tolerant cotton genotypes. These cultivars also provide useful insights for cotton farmers in selecting suitable genotypes, thus reducing dependence on P fertilizers and mitigating associated environmental issues.