High Planting Density Combined with Delayed Topping Improves Short Fruiting Branch Cotton Yield by Enhancing Biomass Accumulation, Canopy Light Interception and Delaying Leaf Senescence

Abstract

Short fruiting branch cotton (SFBC) has a compact plant architecture suitable for dense planting. Plant population density (PPD) and topping are important agronomic practices to achieve high yielding by optimizing cotton plant structure. However, their individual and interactive effects on SFBC growth and yield are poorly understood. This study aimed to explore cotton growth and yield responses to various combinations of PPD and topping time (TT) and the underlying physio-ecological mechanism. Four combinations were included in a two-year field experiment (2023–2024) involving two PPD levels (5.3 plants m⁻², low density LD; 8 plants m⁻², high density HD) and two TT levels (early topping for leaving ten sympodials per plant ET; late topping for leaving fifteen sympodials per plant LT), and compared in terms of biomass accumulation, photosynthetically active radiation capture, and leaf senescence during entire reproductive growth period. Compared to the other three combinations, the combination of HD and LT (HDLT) achieved a higher lint yield due to a greater biological yield, which was predominantly attributed to the higher average rate during the rapid biomass increasing period. Owing to delayed leaf senescence caused by the HD and the LT, the HDLT performed better in leaf senescence-related attributes at the late growth stage. Moreover, these improved attributes also contributed to a higher radiation interception rate and photosynthetic efficiency at the late growth stage. Taken together, combining high density with later topping tends to increase the lint yield of SFBC through increasing dry matter accumulation, delaying leaf senescence, and enhancing canopy radiation interception rate at the late growth.

1. Introduction

Cotton is widely grown worldwide as a primary source of natural fibers for the textile industry. Increasing plant density within a reasonable range has proved to be an effective strategy for achieving cotton high yield [1,2], which is largely attributed to increased boll density or biomass yield, but boll size and harvest index are decreased [3,4,5]. Enhanced biomass yield results from improved utilization of solar radiation as a result of elevated leaf index [3]. High plant density limits the vegetative branch (VB) growth of cotton, minimizing manual removal of VB which is a traditional cultural practice in Chinese cotton production, and thus leading to decreased labor cost [3]. However, excessive plant density typically aggravates intraspecific competition between plants, worsening light transmission and ventilation within plant population, and in turn leading to massive fruits shedding [6]. Cotton varieties with a compact plant architecture are suitable for high-density planting. In general, compact plant architecture can be developed through the application of mepiquat chloride (N, N-dimethylpiperidinium) [4,5,7]. However, the regulative effect of mepiquat chloride on plant architecture and yield is not easily predicted varying with genotype, weather, dose, etc. [8,9]. The fruiting branch is categorized into four groups based on its average internode length: 2–5 cm (I type, compact), 5–10 cm (II type, less compact), 10–15 cm (III type, less loose), and >15 cm (loose) [10]. Short fruiting branch cotton (SFBC) is commonly referred to as the I type that is caused by the suppressed cell elongation of the internode meristematic tissue [11], but the main stem does not show stunted growth. The phenotype of short fruiting branch allows more solar radiation transmission into the lower canopy and thus is fitted into a high-density planting system.

Cotton is a typical indeterminate growth crop where vegetative and reproductive growths are tangled with each other for a relatively long time. Topping can promote the formation of reproductive structures (squares, flowers, and bolls) as more assimilates are diverted to fruiting branches below due to apical dominance relief [12], which is widely adopted in China and other cotton-growing regions with abundant agricultural labor. Topping reduces total fruit sites but increases retained boll number, owing to a remarkable reduction in abscised fruit sites [13]. Decapitation does not necessarily bring yield benefits. For example, it is simply conducive to reducing bollworm infestations in Mali [14]; topping triggers leaf senescence as indicated by an obvious decline in the contents of nitrogen, chlorophyll, and cytokinin (CTK) as well as a concurrent increase in abscisic acid (ABA) to indoleacetic acid (IAA) ratio and ABA to CTK ratio in the fourth leaf from the apex [15]. The timing of plant topping plays an important role in improving cotton yield potential, which largely depends on the light and thermal availability for a given cotton-grown region, plant density, etc. Early topping facilitates reproductive growth but comprises vegetable growth, which tends to induce leaf senescence and earlier cutout [16]. Late topping is inclined to produce more ineffective reproductive forms in upper fruiting branches due to limited light and heat supplies during the late growth season [17]. Timely topping can regulate vegetative and reproductive growth by enabling reasonable partitioning of photoassimilates to enhance yields [17]. The effectiveness of topping varies depending on topping method (manual or chemical), topping agent (mepiquat chloride MC) dose, plant density, variety sensibility to MC, and environmental condition, etc. [6,18]. Compared with no topping, chemical topping at low plant density leads to decreased seed cotton yield as a result of insufficient biological yield, but as reversed at moderate and high plant densities as a result of higher assimilates allocation to reproductive forms [5]. For MC-insensitive varieties, chemical topping with MC can replace manual topping at moderate or high plant densities; for MC-sensitive varieties, no topping or continued low-dose chemical topping with MC is preferred [18]. Compared to manual topping, chemical topping delays the assimilate allocation to reproductive organs and boll development under high-density planting [6]. Presently, the cotton growth and yield responses to the interactive action of manual topping timing (TT) and plant population density (PPD) remain unclear, especially for SFBC with a different architecture development pattern from the common cotton cultivars. It was hypothesized that PPD and TT synergistically affect the yield formation of SFBC by the modulation of biomass production, maturation dynamic, and solar radiation capture.

2. Materials and Methods

2.1. Experimental Design and Treatments

A two-year field experiment was conducted in 2023 and 2024 in the agricultural experimental station of Jiangxi Agricultural University Nanchang, China (28°46′ N, 115°55′ E). The soil type was acid-red soil with a lower pH of 5.13. The upper 20 cm soil layer contained 26.34 g kg⁻¹ of organic matter, 259.29 mg kg⁻¹ of available potassium, 78.47 mg kg⁻¹ of available phosphorous, 64.69 mg kg⁻¹ of available nitrogen, and 1.48 g kg⁻¹ of total nitrogen. An SFBC line P4D was arranged in a randomized complete block design with three replications. The line P4D has a shorter average internode length at 4.5, 4.7, and 4.6 cm for the third (bottom), the sixth (middle), and the ninth (upper) fruiting branches, respectively. A 2 (TT) × 2 (PPD) factorial design was included in the experiment. PPD was set to be 5.3 plants m⁻² (lower plant density LD) and 8 plants m⁻² (higher plant density HD). Earlier topping for remaining ten sympodials (ET) single plant was on 23 July 2023, and 3 August 2024, and later topping for leaving fifteen sympodials (LT) was on 7 August 2023, and 23 August 2024, respectively. The traditional topping time is from late July to mid-August for the Yangtze River Valley cotton region. Each plot consisted of four rows of cotton with 9.5 m in length and 1.5 m in width. The row spacing was 0.5 m for the HD treatment, and 0.75 m for the LD treatment, and the interplant distance was 0.25 m for all treatments. Cotton planting was implemented following a seedling nursing and transplanting protocol with sowing dates on 10 April 2023, and 12 April 2024, and seedling transplanting on 19 May 2023, and 9 May 2024, respectively. The N, P₂O₅, and K₂O were incorporated into the field as urea, calcium-magnesium phosphate, and potassium chloride at rates of 276, 90, and 315 kg ha⁻¹, respectively. The P fertilizer was used as the base fertilizer before sowing. The K fertilizer was split applied before sowing and peak squaring in two equal rates, and N fertilizer was applied into a pre-plant application, a peak squaring, and a peak flowering in a 1:2:1 ratio.

2.2. Morphological Traits, Yield and Yield Characters

At the cotton maturation stage, ten randomly selected plants with uniform growth from each plot were selected to examine the following morphological traits, including plant height (PH), fruiting branch number (FBN), the third (bottom) fruiting branch angle (TFBA), the sixth (middle) fruiting branch angle (SFBA), the ninth (upper) fruiting branch angle (NFBA), the third (bottom) fruiting branch length (TFBL), the sixth (middle) fruiting branch length (SFBL), the ninth (upper) fruiting branch length (NFBL). Total bolls on the ten plants were counted and boll density (the number of bolls per unit ground area) was calculated. Fifty bolls were collected randomly from each plot to determine boll weight, seed index, and lint percentage. The total seed cotton yield for each plot was summed after multiple manual harvestings prior to frost, which was converted into lint yield by multiplying by the lint percentage and expressed in kilograms per hectare (kg ha⁻¹).

2.3. Biomass Determination

Five samplings were taken to determine the dry biomass amount of various plant parts at squaring, first flowering (FF), peak flowering (PF), peak boll-setting (PB), and boll-opening (BO) stages. Three growth-uniform plants randomly chosen from each plot were carefully uprooted, and dissected into roots, blades, main stems, branches (containing petioles), and reproductive organs. Samples were immediately killed in an electric fan-assisted oven at 105 °C for 30 min to stop cell metabolism, and then dried at 70 °C to a constant weight for the determination of dry biomass.

2.4. Photosynthetic and Leaf Senescence-Related Characteristics

Net photosynthetic rate (Pn) was measured between 9:00 and 11:00 am on cloudless days in the fourth leaf from the apex (the second leaf from the apex after topping) at the PF and BO stages. Measurements were taken using a CI-340 portable photosynthesis system (CID Bio-Science, Inc., Camas, WA, USA) when the ambient photosynthetic photon flux density exceeded 1500 µM m⁻² s⁻¹. The Pn was averaged across three plants in each plot with each plant replicated three times.

The SPAD value in the fourth leaf from the apex was examined at the PF, PB, and BO stages using a SPAD meter (SPAD-502, Konica Minolta, Tokyo, Japan), which was averaged across three plants in each plot with four measurements for each plant.

The fourth leaf from the apex was harvested at the PF, PB, and BO stages and oven-dried to constant weight, and then ground into fine powders until they can pass through a 32-mesh sieve (Huayin Boltingcloth Manufacturing Inc., Shanghai, China) for analysis of chlorophyll, carotenoid, soluble protein, and total nitrogen. Chlorophyll a, b, and carotenoid were quantified using the method as described by Murat and Yang [19]. Soluble protein was determined using the bicinchoninic acid (BCA) assay [20]. Total nitrogen content was measured using the Kieldahl method. Briefly, sample powders (0.5 g) were put in a decoction tube, and 0.5 g catalyst (4.50 g CuSO₄-5H₂O + 0.50 g K₂SO₄) as well as 10 mL H₂SO₄ were added followed by an even mixture. Later, they were transferred to a furnace for digestion at 420 °C for one hour. After being cooled to room temperature, the digested solution was submitted to nitrogen analysis using a FOSS Kjeldahl Nitrogen Analyzer (Kjeltec-TM 2300, FOSS, Hillerød, Sweden).

2.5. Light Interception

Solar radiation capture was examined and expressed as the radiation interception rate at the PF, PB, and BO stages. Photosynthetically active radiation (PAR) measurements were taken between 9:00 and 11:00 am on windless sunny days using a line quantum meter with ten sensors on it, producing an average PAR across the sensors (MQ-310, Apogee Instruments Inc., Logan, UT, USA). Instantaneous PARs above canopy and below canopy (10 cm above ground) were collected. A total of six light measurements were taken on the bottom surface of each observation point, each three times for row orientation and plant orientation, respectively. The radiation interception rate (%) = 100 × (PAR above canopy—PAR below canopy)/PAR above canopy.

2.6. Data Analysis

All data were subjected to a two-way analysis of variance where PPD and TT served as fixed effects and block (replicate) as a random factor using SPSS Statistics 26. The means were separated according to Duncan’s multiple range tests at p < 0.05.

A logistic regression model was fitted to the data of biomass accumulation in various plant parts including underground, aboveground, and the entire plant. The regression equation is as follows:

$$Y={\displaystyle \frac{a}{1+b{e}^{-kt}}}$$

(1)

where t(d) is the number of days after sowing (DAS), Y (g m⁻²) represents the biomass at a given time point (t), and a (g m⁻²) represents the maximum biomass, while b and k are two coefficients. Based on the regression equation, the following parameters can be calculated:

$$t_1=-{\displaystyle \frac{1}{k}}\ln ({\displaystyle \frac{2+\sqrt{3}}{b}})$$

(2)

$$t_2=-{\displaystyle \frac{1}{k}}\ln ({\displaystyle \frac{2-\sqrt{3}}{b}})$$

(3)

$$t_m=\frac{\ln b}{k}$$

(4)

$$V_{max}=\frac{ka}{4}$$

(5)

$$V_T={\displaystyle \frac{Y_2-Y_1}{t_2-t_1}}$$

(6)

where V_max (g d⁻¹) and V_T (g d⁻¹) represent the highest and average biomass accumulation rate during the exponential growth period, respectively. t_m (d) denotes the time point when the highest biomass accumulation rate occurs. t₁ and t₂ represent the onset and termination DAS of the exponential biomass growth period, respectively. Y₁ and Y₂ indicate the biomass at t₁ and t₂, respectively.

3. Results

3.1. Yield and Yield Components

Seed cotton, lint, and biological yields were significantly affected by the main effects of PPD and TT in both years and their interaction in 2023 (

Table 1. Effects of planting density and topping time on cotton yield and yield components in 2023 and 2024.

Year	Treatment	Boll Density (per m−2)	Boll Weight (g)	Lint Percentage (%)	Seed Index (g)	Seed Cotton Yield (kg ha−1)	Lint Yield (kg ha−1)	Biomass Yield (kg ha−1)	Harvest Index
2023	LDET	147.2 ± 21.6 a	4.89 ± 0.03 a	39.6 ± 0.6 a	9.8 ± 0.2 a	4142.1 ± 168.2 b	1638.7 ± 64.9 b	12,067.3 ± 1400.1 b	0.34 ± 0.01 a
	LDLT	152.9 ± 9.0 a	4.87 ± 0.05 a	38.6 ± 0.8 a	9.6 ± 0.1 a	4249.4 ± 87.7 b	1642.0 ± 66.05 b	13,135.1 ± 325.0 b	0.32 ± 0.01 ab
	HDLT	172.0 ± 29.2 a	4.73 ± 0.06 a	39.6 ± 0.7 a	9.6 ± 0.3 a	4696.8 ± 5.5 a	1859.0 ± 35.5 a	16,632.3 ± 662.5 a	0.28 ± 0.05 b
	HDET	164.8 ± 35.7 a	4.87 ± 0.31 a	40.2 ± 1.3 a	9.7 ± 0.5 a	4147.6 ± 151.7 b	1667.7 ± 29.2 b	12,001.9 ± 558.3 b	0.35 ± 0.01 a
	LD mean	150.0 ± 15.1 A	4.88 ± 0.04 A	39.1 ± 0.8 A	9.7 ± 0.1 A	4195.7 ± 133.6 B	1640.3 ± 58.6 B	12,601.2 ± 1081.0 B	0.33 ± 0.01 A
	HD mean	168.4 ± 29.4 A	4.80 ± 0.21 A	39.9 ± 1.0 A	9.7 ± 0.4 A	4422.2 ± 315.8 A	1763.3 ± 108.8 A	14,317.1 ± 2594.7 A	0.32 ± 0.05 A
	ET mean	156.0 ± 28.1 x	4.88 ± 0.20 x	39.9 ± 1.0 x	9.7 ± 0.32 x	4144.8 ± 143.3 y	1653.2 ± 47.8 y	12,034.6 ± 954.0 y	0.35 ± 0.01 x
	LT mean	162.4 ± 22.0 x	4.80 ± 0.09 x	39.1 ± 0.9 x	9.6 ± 0.2 x	4473.1 ± 251.3 x	1750.5 ± 128.0 x	14,883.7 ± 1971.5 x	0.30 ± 0.04 x
					Source of variation (F)
	PPD	1.51	0.68	2.47	0.67	10.43 *	16.98 **	12.55 **	10.84
	TT	0.19	0.68	2.37	0.45	21.91 **	10.63 *	34.59 **	13.22
	PPD × TT	0.03	0.37	0.07	0.07	9.93 *	9.91 *	13.52 **	50.44 *
2024	LDET	98.8 ± 14.2 b	2.87 ± 0.06 a	36.0 ± 0.1 a	12.0 ± 0.2 a	2831.0 ± 433.3 c	1023.0 ± 132.2 c	9807.2 ± 659.2 b	0.29 ± 0.06 b
	LDLT	145.3 ± 14.7 a	2.87 ± 0.23 a	35.7 ± 0.1 a	11.8 ± 0.1 a	4176.5 ± 758.1 ab	1486.7 ± 266.9 b	8333.2 ± 39.9 b	0.50 ± 0.10 a
	HDLT	159.9 ± 7.8 a	3.23 ± 0.06 a	36.0 ± 0.1 a	11.6 ± 0.1 a	5146.0 ± 306.3 a	1836.8 ± 44.1 a	16528.8 ± 1944.4 a	0.31 ± 0.03 b
	HDET	146.2 ± 10.0 a	2.80 ± 0.20 a	37.7 ± 0.1 a	11.9 ± 0.5 a	4052.0 ± 508.5 bc	1506.2 ± 125.9 b	10222.4 ± 730.8 b	0.40 ± 0.08 ab
	LD mean	122.0 ± 28.6 B	2.87 ± 0.15 A	35.8 ± 0.1 A	11.9 ± 0.2 A	3503.7 ± 921.0 B	1254.9 ± 316.2 B	9070.2 ± 909.0 B	0.40 ± 0.01 A
	HD mean	153.0 ± 10.9 A	3.02 ± 0.27 A	36.8 ± 0.1 A	11.7 ± 0.4 A	4599.0 ± 707.1 A	1671.5 ± 199.8 A	13375.6 ± 3695.6 A	0.36 ± 0.07 A
	ET mean	122.5 ± 28.2 y	2.83 ± 0.14 x	36.8 ± 0.1 x	11.9 ± 0.3 x	3441.5 ± 791.1 y	1264.6 ± 288.7 y	10014.8 ± 662.7 y	0.35 ± 0.08 x
	LT mean	152.6 ± 13.2 x	3.05 ± 0.25 x	35.8 ± 0.1 x	11.7 ± 0.2 x	4661.3 ± 741.2 x	1661.8 ± 257.0 x	12431.0 ± 465.4 x	0.41 ± 0.01 x
					Source of variation (F)
	PPD	20.15 **	2.70	0.39	0.78	12.91 **	19.56 **	46.82 **	0.92
	TT	18.91 **	5.63	0.39	2.67	16.02 **	17.78 **	14.75 *	2.35
	PPD × TT	5.64 *	5.63	0.17	0.20	0.17	0.50	38.23 **	13.43 **

Note: Data are represented as mean ± SD. Different lower-case and upper-case letters indicate significant differences at p < 0.05 among four individuals and between two plant density treatments across two topping time levels, respectively. Letters x and y indicate significant differences at p < 0.05 between two topping time treatments across two plant density levels. *, ** denote significance at the probability of 0.05 and 0.01 levels, respectively. LDET: low planting density combined with earlier topping; LDLT: low planting density combined with later topping; HDLT: high plant density combined with later topping; HDET: high planting density combined with earlier topping. The same as below.

). They were increased with increasing planting density and delaying topping time. Compared with the LD treatment, seed cotton, lint, and biological yields averaged across two years were increased by 18.33%, 20.35%, and 30.55% in the HD treatment, respectively. Compared to the ET treatment, the LT treatment presented an average increase of 21.68%, 18.65%, and 23.90% over two years in seed cotton, lint, and biological yields, respectively. Independent of years, the high density combined with later topping (HDLT) treatment produced higher seed cotton, lint, and biological yields relative to the other treatments. Lint yield increases of 11.47–13.44% in 2023 and 21.95–79.55% in 2024 were observed in HDLT than the three other treatments, respectively. Boll density was significantly affected by the main effects of PPD and TT and their interaction in 2024, which was enhanced with increasing planting density and delaying topping time.

3.2. Agronomic Traits

In both years, TT significantly affected PH and FBN (

Table 2. Effects of planting population density and topping time on agronomic traits of cotton in 2023 and 2024.

Year	Treatment	PH (cm)	FBN	TFBA (°)	SFBA (°)	NFBA (°)	TFBL (cm)	SFBL (cm)	NFBL (cm)
2023	LDET	81.3 ± 4.3 b	10.8 ± 0.6 c	53.2 ± 1.2 a	52.9 ± 2.1 a	53.6 ± 2.6 a	7.6 ± 1.6 a	11.3 ± 1.0 b	16.4 ± 1.5 a
	LDLT	117.8 ± 9.0 a	14.5 ± 0.8 ab	51.6 ± 1.2 a	51.4 ± 1.3 a	50.8 ± 0.7 a	8.6 ± 2.5 a	13.9 ± 1.4 a	14.7 ± 0.9 a
	HDLT	114.9 ± 3.5 a	14.9 ± 0.5 a	53.0 ± 4.2 a	52.5 ± 1.5 a	53.8 ± 3.1 a	8.1 ± 1.5 a	8.8 ± 1.4 c	9.0 ± 1.6 b
	HDET	79.7 ± 0.9 b	10.7 ± 1.0 c	49.6 ± 4.5 a	52.7 ± 2.6 a	52.0 ± 1.4 a	8.0 ± 1.5 a	8.8 ± 0.9 c	10.2 ± 1.9 b
	LD mean	99.5 ± 21.0 A	12.7 ± 2.1 A	52.6 ± 1.6 A	52.1 ± 1.8 A	52.2 ± 2.3 A	8.1 ± 1.9 A	12.6 ± 1.8 A	15.5 ± 1.4 A
	HD mean	97.3 ± 19.4 A	12.8 ± 2.4 A	51.3 ± 4.3 A	52.6 ± 1.9 A	52.9 ± 2.4 A	8.1 ± 1.4 A	8.8 ± 1.0 B	9.6 ± 1.7 B
	ET mean	80.5 ± 2.9 y	10.8 ± 0.7 y	51.6 ± 3.7 x	52.8 ± 2.1 x	52.8 ± 2.1 x	7.8 ± 1.4 x	10.1 ± 1.6 x	13.3 ± 3.7 x
	LT mean	116.4 ± 6.3 x	14.7 ± 0.7 x	52.3 ± 2.9 x	51.9 ± 1.4 x	52.3 ± 2.6 x	8.3 ± 1.8 x	11.4 ± 3.1 x	11.8 ± 3.3 x
		Source of variation (F)
	PPD	0.53	0.15	0.51	0.17	0.27	0.01	31.86 **	46.54 **
	TT	138.91 **	81.12 **	0.13	0.57	0.14	0.25	3.66	2.92
	PPD × TT	0.05	0.38	2.15	0.37	3.36	0.18	3.77	0.08
2024	LDET	63.1 ± 1.1 b	11.0 ± 1.5 b	56.7 ± 3.4 a	54.9 ± 2.5 a	55.7 ± 4.0 a	17.6 ± 2.0 a	15.2 ± 2.3 a	15.8 ± 5.7 a
	LDLT	105.7 ± 5.7 a	15.5 ± 0.8 a	52.7 ± 3.0 a	56.7 ± 2.3 a	55.0 ± 3.4 a	19.3 ± 4.6 a	19.4 ± 3.0 a	16.0 ± 1.4 a
	HDLT	104.2 ± 1.5 a	15.6 ± 0.1 a	50.6 ± 4.0 a	53.7 ± 4.0 a	50.8 ± 1.2 a	16.3 ± 1.3 a	15.9 ± 4.1 a	11.6 ± 4.1 a
	HDET	63.0 ± 3.8 b	11.4 ± 0.4 b	53.5 ± 3.5 a	53.2 ± 1.2 a	52.7 ± 2.4 a	16.4 ± 0.9 a	15.8 ± 3.4 a	11.3 ± 1.7 a
	LD mean	84.4 ± 23.6 A	13.2 ± 2.7 A	54.7 ± 3.6 A	55.8 ± 2.4 A	55.3 ± 3.3 A	18.4 ± 3.3 A	17. ± 3.3 A	15.9 ± 3.7 A
	HD mean	83.6 ± 22.7 A	13.5 ± 2.3 A	52.1 ± 3.7 A	53.4 ± 2.6 A	51.7 ± 2.0 A	16.3 ± 1.0 A	15.8 ± 3.4 A	11.4 ± 2.8 A
	ET mean	63.0 ± 2.5 y	11.2 ± 1.0 y	55.1 ± 3.5 x	54.0 ± 2.0 x	54.2 ± 3.4 x	17.0 ± 1.6 x	15.5 ± 2.6 x	13.5 ± 4.5 x
	LT mean	104.9 ± 3.8 x	15.6 ± 0.5 x	51.7 ± 3.4 x	55.2 ± 3.3 x	52.9 ± 3.2 x	17.8 ± 3.4 x	17.6 ± 3.8 x	13.8 ± 3.6 x
		Source of variation (F)
	PPD	0.16	0.33	1.67	2.27	4.49	1.92	0.61	4.46
	TT	419.97 **	77.85 **	2.92	0.57	0.58	0.26	1.35	0.01
	PPD × TT	0.1	0.14	0.07	0.19	0.11	0.34	1.23	0.01

Note: Data are represented as mean ± SD. Different lower-case and upper-case letters indicate significant differences at p < 0.05 among four individuals and between two plant density treatments across two topping time levels, respectively. Letters x and y indicate significant differences at p < 0.05 between two topping time treatments across two plant density levels. ** denotes significance at the probability of 0.01 level. PH: plant height; FBN: fruiting branch numbers; TFBA: the third fruiting branch angle; SFBA: the sixth fruiting branch angle; NFBA: the ninth fruiting branch angle; TFBL: the third fruiting branch length; SFBL: the sixth fruiting branch length; NFBL: the ninth fruiting branch length.

). The PH and FBN averaged across both years were significantly higher by 55.05% and 38.19% in the LT than the ET, respectively. PPD significantly influenced SFBL and NFBL in 2023, but not in 2024. The SFBL and NFBL were significantly higher by 44.01% and 61.60% in the LD than the HD, respectively. The fruiting branch angle was affected by neither PPD nor TT.

3.3. Biomass Accumulation

The dynamics of the underground, aboveground, and total biomass accumulation followed a logistic function with the determination coefficient significant at p < 0.05 (

Table 3. Logistic equation fitting of the dry matter accumulation in the underground part (roots) per unit ground area.

Year	Treatment	n	Regression Equation	R2	t1	t2	T	Vmax	VT	tm	Mmax	Mobs
					(d)	(d)	(d)	(g d−1)	(g d−1)	(d)	(g m−2)	(g m−2)
2023	LDET	5	Y = 65.50/(1 + 15,134.01e−0.090t)	0.9985 *	92.1	121.3	29.2	1.48	1.30	106.7	65.52	64.08
	LDLT	5	Y = 80.00/(1 + 52,698.42e−0.099t)	0.9430 *	97.1	123.8	26.8	1.97	1.73	110.4	80.00	69.54
	HDLT	5	Y = 107.00/(1 + 2853.55e−0.071t)	0.9879 *	93.1	130.0	36.9	1.91	1.68	111.5	107.29	101.44
	HDET	5	Y = 74.40/(1 + 67,291.91e−0.116t)	0.9722 *	84.8	107.6	22.8	2.15	1.89	96.2	74.45	81.67
2024	LDET	5	Y = 103.92/(1 + 936.16e−0.05t)	0.9889 *	110.6	163.3	52.7	1.30	1.14	136.9	103.92	83.84
	LDLT	5	Y = 67.71/(1 + 12,927.64e−0.082t)	0.9932 *	99.6	131.7	32.2	1.39	1.22	115.6	67.71	66.41
	HDLT	5	Y = 141.68/(1 + 366,508.3e−0.097t)	0.9784 *	118.8	146.0	27.2	3.43	3.01	132.4	141.68	133.01
	HDET	5	Y = 127.37/(1 + 959.97e−0.049t)	0.9980 *	113.6	167.5	53.9	1.56	1.36	140.5	127.37	98.03

Note: R²: coefficient of determination; n: number of samples; *: significant at the 0.05 level; t₁ and t₂ represent the initiation and termination point of the fast accumulation period (FAP), respectively; T is the duration between t₁ and t₂; V_T and V_max denote the average and maximum rates of biomass accumulation during FAP, respectively; t_m indicates the time point when the highest rate of dry matter accumulation occurs; M_max and M_obs represent the theoretical maximum and observed values of the final biomass, respectively.

,

Table 4. Logistic equation fitting of the dry matter accumulation in the aboveground part per unit ground area.

Year	Treatment	n	Regression Equation	R2	t1	t2	T	Vmax	VT	tm	Mmax	Mobs
					(d)	(d)	(d)	(g d−1)	(g d−1)	(d)	(g m−2)	(g m−2)
2023	LDET	5	Y = 1156.32/(1 + 5127.78e−0.073t)	0.9988 *	99.2	135.4	36.2	21.05	18.46	117.3	1156.32	1144.22
	LDLT	5	Y = 1582.64/(1 + 269.16e−0.04t)	0.9585 *	106.9	172.7	65.8	15.84	13.89	139.8	1582.64	1243.97
	HDLT	5	Y = 1754.42/(1 + 663.58e−0.05t)	0.9958 *	102.9	155.3	52.3	22.07	19.36	129.1	1754.42	1561.79
	HDET	5	Y = 1159.61/(1 + 2918.40e−0.071t)	0.9956 *	93.7	130.8	37.1	20.61	18.07	112.3	1159.61	1118.53
2024	LDET	5	Y = 1165.72/(1 + 1736.52e−0.053t)	0.9922 *	116.7	166.8	50.1	15.33	13.45	141.8	1165.72	896.89
	LDLT	5	Y = 828.59/(1 + 5675.79e−0.069t)	0.9926 *	106.3	144.6	38.3	14.27	12.52	125.4	828.59	766.90
	HDLT	5	Y = 1612.76/(1 + 426,138.79e−0.097t)	0.9874 *	119.9	147.0	27.1	39.16	34.34	133.4	1612.75	1519.87
	HDET	5	Y = 1030.43/(1 + 3658.43e−0.064t)	0.9978 *	93.7	145.0	41.5	16.36	14.34	129.2	1030.43	924.21

Note: R²: coefficient of determination; n: number of samples; *: significant at the 0.05 level; t₁ and t₂ represent the initiation and termination point of the fast accumulation period (FAP), respectively; T is the duration between t₁ and t₂; V_T and V_max denote the average and maximum rates of biomass accumulation during FAP, respectively; t_m indicates the time point when the highest rate of dry matter accumulation occurs; M_max and M_obs represent the theoretical maximum and observed values of the final biomass, respectively.

and

Table 5. Logistic equation fitting of total dry matter accumulation per unit ground area.

Year	Treatment	n	Regression Equation	R2	t1	t2	T	Vmax	VT	tm	Mmax	Mobs
					(d)	(d)	(d)	(g d−1)	(g d−1)	(d)	(g·m−2)	(g·m−2)
2023	LDET	5	Y = 1216.89/(1 + 5620.63e−0.07t)	0.9989 *	98.7	134.2	35.5	22.56	20.44	116.5	1216.89	1206.73
	LDLT	5	Y = 1506.08/(1 + 362.22e−0.05t)	0.9646 *	102.2	161.0	58.8	16.86	15.27	131.6	1506.08	1313.51
	HDLT	5	Y = 1834.76/(1 + 727.31e−0.05t)	0.9961 *	101.7	152.5	50.8	23.78	21.55	127.1	1834.76	1663.23
	HDET	5	Y = 1233.85/(1 + 2788.55e−0.07t)	0.9967 *	92.7	129.7	36.9	22.01	19.94	111.2	1233.85	1200.19
2024	LDET	5	Y = 1273.05/(1 + 1610.53e−0.05t)	0.9922 *	128.2	166.8	38.5	16.61	15.05	141.5	1273.05	980.72
	LDLT	5	Y = 895.66/(1 + 5637.45e−0.07t)	0.9929 *	114.6	143.6	29.0	15.52	14.06	124.6	895.66	833.32
	HDLT	5	Y = 1754.15/(1 + 4227.47e−0.10t)	0.9868 *	126.2	146.9	20.7	42.60	38.60	133.4	1754.15	1652.88
	HDET	5	Y = 1148.14/(1 + 3174.20e−0.06t)	0.9984 *	118.8	151.1	32.4	17.81	16.14	129.9	1148.14	1022.24

Note: R²: coefficient of determination; n: number of samples; *: significant at the 0.05 level; t₁ and t₂ represent the initiation and termination point of the fast accumulation period (FAP), respectively; T is the duration between t₁ and t₂; V_T and V_max denote the average and maximum rates of biomass accumulation during FAP, respectively; t_m indicates the time point when the highest rate of dry matter accumulation occurs; M_max and M_obs represent the theoretical maximum and observed values of the final biomass, respectively.

). Among the four combinations of PPD and TT, the HDLT exhibited the largest M_max (theoretical maximum value of the final biomass) and M_obs (observed value of the final biomass) irrespective of plant parts. There was an increase of 21.82–95.85% with M_max and an increase of 26.62–98.35% with M_obs for the whole plant biomass accumulation in the HDLT than the other treatments over both years (Table 5). Likewise, independent of plant parts, V_max (maximum rate of rapid biomass accumulation) and V_T (average rate of rapid biomass accumulation) were higher in the HDLT relative to the others except for the root biomass accumulation in 2023. For the total biomass accumulation, the V_T and V_max were greater by 5.41–174.48% and by 5.43–174.55% in the HDLT over two years (Table 5).

3.4. Photosynthetic Parameters

TT significantly affected Pn at the PF and BO stages over two years except the PF stage in 2023 (

Table 6. Effects of plant population density and topping time on the net photosynthetic rate in the main stem functional leaves (µmol m⁻² s⁻¹).

Treatment	2023		2024
	PF (DAS 113)	BO (DAS 169)	PF (DAS 121)	BO (DAS 165)
LDET	22.4 ± 1.3 a	12.6 ± 1.0 c	32.1 ± 2.6 a	17.2 ± 1.9 bc
LDLT	22.0 ± 4.6 a	18.5 ± 1.1 b	26.9 ± 0.6 b	20.6 ± 1.3 b
HDLT	22.5 ± 1.1 a	21.6 ± 2.1 a	26.3 ± 2.7 b	24.6 ± 4.2 a
HDET	24.0 ± 1.1 a	12.4 ± 2.7 c	32.8 ± 3.1 a	16.3 ± 2.9 c
LD mean	22.2 ± 2.9 A	16.5 ± 3.1 A	29.9 ± 3.3 A	18.7 ± 2.4 A
HD mean	23.3 ± 1.3 A	17.0 ± 5.4 A	29.4 ± 4.3 A	20.1 ± 5.5 A
ET mean	22.9 ± 1.4 x	12.5 ± 2.0 y	32.4 ± 2.8 x	16.7 ± 2.4 y
LT mean	22.2 ± 3.4 x	19.7 ± 2.2 x	32.1 ± 2.6 x	17.2 ± 1.9 x
	Source of variance (F)
PPD	0.65	2.43	0.01	1.45
TT	0.53	67.74 **	34.24 **	20.00 **
PPD × TT	0.16	3.12	0.44	3.65

Note: PF: peak flowering; BO: boll-opening; DAS: days after sowing; ** means significant at 0.01 level. Data are represented as mean ± SD. Different lower-case and upper-case letters indicate significant differences at p < 0.05 among four individuals and between two plant density treatments across two topping time levels, respectively. Letters x and y indicate significant differences at p < 0.05 between two topping time treatments across two plant density levels.

). However, PPD had no significant effect on Pn. Compared to the ET treatment, the LT treatment significantly reduced Pn by 19.17% at the PF stage in 2024, but which was reversed at the BO stage with an average increase of 47.32% across both years. At the BO stage, Pn was increased on average by 54.90%, 18.18%, and 60.95% in the HDLT treatment than the LDET, LDLT, and HDET treatments, respectively.

In both years, the SPAD value was significantly influenced by TT during the PF to BO stages. PPD had a significant effect on SPAD value during the PF to PB stages in 2023 (

Table 7. Effects of plant population density and topping time on SPAD value in main stem functional leaves.

Treatment	2023			2024
	PF (DAS 113)	PB (DAS 149)	BO (DAS 169)	PF (DAS 121)	PB (DAS 148)	BO (DAS 165)
LDET	40.8 ± 2.3 b	51.3 ± 2.6 b	45.6 ± 3.7 b	42.7 ± 2.7 a	53.6 ± 1.4 a	46.0 ± 2.2 b
LDLT	32.5 ± 1.3 c	48.5 ± 2.2 b	50.6 ± 1.8 a	37.9 ± 2.8 b	44.5 ± 2.1 b	47.1 ± 1.9 ab
HDLT	34.2 ± 2.0 c	50.3 ± 1.8 b	51.7 ± 2.0 a	37.4 ± 2.3 b	44.1 ± 2.3 b	49.7 ± 3.6 a
HDET	45.7 ± 2.8 a	55.6 ± 3.1 a	44.1 ± 2.6 b	41.7 ± 1.5 a	53.2 ± 3.0 a	44.6 ± 1.5 b
LD mean	36.6 ± 4.6 B	49.9 ± 2.6 B	48.1 ± 3.8 A	40.2 ± 3.7 A	49.0 ± 5.0 A	46.8 ± 2.0 A
HD mean	40.0 ± 6.4 A	53.0 ± 3.7 A	47.9 ± 4.5 A	39.3 ± 2.9 A	48.6 ± 5.3 A	47.7 ± 3.9 A
ET mean	43.2 ± 3.6 x	53.5 ± 3.6 x	44.9 ± 3.2 y	42.2 ± 2.2 x	53.4 ± 2.3 x	45.1 ± 1.8 y
LT mean	33.4 ± 1.8 y	49.4 ± 1.9 y	51.2 ± 1.9 x	37.6 ± 2.5 y	44.3 ± 2.1 y	48.6 ± 3.2 x
	Source of variance (F)
PPD	21.16 **	10.07 **	0.05	0.98	0.28	0.30
TT	188.30 **	17.64 **	52.91 **	35.66 **	144.36 **	7.33 *
PPD × TT	5.31 *	1.71	2.28	0.15	0.11	2.98

Note: PF: peak flowering; PB: peak boll-setting; BO: boll-opening; DAS: days after sowing. Data are represented as mean ± SD. Different lower-case and upper-case letters indicate significant differences at p < 0.05 among four individuals and between two plant density treatments across two topping time levels, respectively. Letters x and y indicate significant differences at p < 0.05 between two topping time treatments across two plant density levels. *, ** denote significance at the probability of 0.05 and 0.01 levels, respectively.

). Compared with the LD treatment, the HD treatment displayed significantly higher SPAD values by 9.06% and 6.20% at the PF and PB stages, respectively. Relative to the ET treatment, the LT treatment showed an average reduction of 16.89% and 12.04% across two years at the PF and PB stages, respectively. Conversely, the SPAD value with the LT treatment was enhanced by 14.09% in 2023 and by 7.81% in 2024 at the BO stage (p < 0.05), respectively. At the BO stage, the SPAD value in the HDLT treatment was higher than that in the LDET, LDLT, and HDET treatments with an average increase of 10.79% (p < 0.05), 3.88% (p > 0.05), and 14.39% (p < 0.05) across two years, respectively.

TT significantly affected the chlorophyll a concentration at the PF stage in both years and at the BO stage in 2024 (

Table 8. Effects of plant population density and topping time on the chlorophyll a concentration in main stem functional leaves (mg g⁻¹).

Treatment	2023			2024
	PF (DAS 113)	PB (DAS 149)	BO (DAS 169)	PF (DAS 121)	PB (DAS 148)	BO (DAS 165)
LDET	1.12 ± 0.01 b	1.17 ± 0.14 a	1.04 ± 0.11 ab	1.28 ± 0.19 ab	1.44 ± 0.03 a	0.92 ± 0.8 b
LDLT	0.85 ± 0.09 c	1.32 ± 0.014 b	0.92 ± 0.08 b	1.01 ± 0.13 b	1.50 ± 0.12 a	1.37 ± 0.17 a
HDLT	0.78 ± 0.10 c	1.43 ± 0.13 a	1.14 ± 0.12 a	1.17 ± 0.17 ab	1.57 ± 0.07 a	1.33 ± 0.06 a
HDET	1.33 ± 0.15 a	1.44 ± 0.02 a	0.98 ± 0.05 ab	1.36 ± 0.15 a	1.40 ± 0.14 a	0.87 ± 0.08 b
LD mean	0.98 ± 0.16 A	1.25 ± 0.12 B	0.98 ± 0.11 A	1.14 ± 0.21 A	1.47 ± 0.09 A	1.15 ± 0.27 A
HD mean	1.06 ± 0.32 A	1.43 ± 0.08 A	1.06 ± 0.12 A	1.26 ± 0.18 A	1.49 ± 0.14 A	1.10 ± 0.26 A
ET mean	1.22 ± 0.15 x	1.31 ± 0.17 x	1.01 ± 0.09 x	1.32 ± 0.16 x	1.42 ± 0.09 x	0.89 ± 0.08 y
LT mean	0.81 ± 0.09 y	1.38 ± 0.10 x	1.03 ± 0.15 x	1.08 ± 0.16 y	1.54 ± 0.10 x	1.35 ± 0.11 x
	Source of variance (F)
PPD	1.62	10.85 *	2.08	1.67	0.07	0.62
TT	51.38 **	1.57	0.16	6.40 *	4.04	57.05 **
PPD × TT	5.63 *	1.99	7.58 *	0.24	0.81	0.01

Note: PF: peak flowering; PB: peak boll-setting; BO: boll-opening; DAS: days after sowing. Data are represented as mean ± SD. Different lower-case and upper-case letters indicate significant differences at p < 0.05 among four individuals and between two plant density treatments across two topping time levels, respectively. Letters x and y indicate significant differences at p < 0.05 between two topping time treatments across two plant density levels. *, ** denote significance at the probability of 0.05 and 0.01 levels, respectively.

). PPD had a significant effect on the chlorophyll a concentration at the PB stage in 2023. The interaction of PPD and TT on the chlorophyll a concentration was observed at the PF and BO stages in 2023. At the PF stage, the chlorophyll a concentration in the ET treatment was significantly higher than that in the LT treatment by 50.61% and 22.22% in 2023 and 2024, respectively. Contrary to this, the LT treatment showed a greater chlorophyll a concentration by 51.69% at the BO stage in 2024, but no difference was detected in 2023. At the BO stage, the chlorophyll a concentration averaged across two years in the HDLT treatment was 26.53%,7.83%, and 33.33% higher than those in the LDET, LDLT, and HDET treatments, respectively.

The chlorophyll b concentration was significantly affected by PPD at the PB and BO stages in 2023, by TT at the PF stage in 2023 and at the BO stage in 2024, and by the interaction of PPD and PF at the PF and BO stages in 2023 (

Table 9. Effects of plant population density and topping time on the chlorophyll b concentration in main stem functional leaves (mg g⁻¹).

Treatment	2023			2024
	PF (DAS 113)	PB (DAS 149)	BO (DAS 169)	PF (DAS 121)	PB (DAS 148)	BO (DAS 165)
LDET	0.45 ± 0.01 b	0.59 ± 0.01 b	0.47 ± 0.01 a	0.56 ± 0.07 a	0.74 ± 0.06 a	0.48 ± 0.04 b
LDLT	0.34 ± 0.03 c	0.59 ± 0.05 b	0.39 ± 0.02 b	0.51 ± 0.04 a	0.69 ± 0.069 a	0.65 ± 0.08 a
HDLT	0.33 ± 0.03 c	0.65 ± 0.03 a	0.53 ± 0.01 a	0.52 ± 0.08 a	0.71 ± 0.01 a	0.64 ± 0.03 a
HDET	0.51 ± 0.03 a	0.65 ± 0.04 a	0.46 ± 0.06 a	0.58 ± 0.05 a	0.69 ± 0.06 a	0.52 ± 0.04 b
LD mean	0.40 ± 0.06 A	0.59 ± 0.01 B	0.43 ± 0.05 B	0.53 ± 0.06 A	0.71 ± 0.06 A	0.57 ± 0.11 A
HD mean	0.42 ± 0.10 A	0.65 ± 0.03 A	0.50 ± 0.05 A	0.55 ± 0.07 A	0.71 ± 0.04 A	0.58 ± 0.07 A
ET mean	0.48 ± 0.04 x	0.62 ± 0.04 x	0.47 ± 0.04 x	0.57 ± 0.06 x	0.72 ± 0.06 x	0.50 ± 0.04 y
LT mean	0.34 ± 0.03 y	0.62 ± 0.04 x	0.46 ± 0.08 x	0.51 ± 0.06 x	0.70 ± 0.05 x	0.65 ± 0.05 x
	Source of variance (F)
PPD	2.36	14.97 **	12.91 **	0.21	0.13	0.14
TT	92.38 **	0.01	0.22	2.44	0.32	25.73 **
PPD × TT	7.06 *	0.01	14.24 **	0.01	1.41	0.84

Note: PF: peak flowering; PB: peak boll-setting; BO: boll-opening; DAS: days after sowing. Data are represented as mean ± SD. Different lower-case and upper-case letters indicate significant differences at p < 0.05 among four individuals and between two plant density treatments across two topping time levels, respectively. Letters x and y indicate significant differences at p < 0.05 between two topping time treatments across two plant density levels. *, ** denote significance at the probability of 0.05 and 0.01 levels, respectively.

). Compared to the LD treatment, the HD treatment exhibited 10.17% and 16.28% greater chlorophyll b concentration at the PB and BO stages, respectively, in 2023 (p < 0.05). The chlorophyll b concentration in the ET treatment was increased by 41.18% at the PF stage in 2023, but decreased by 23.08% at the BO stage in 2024 relative to the LT treatment. At the BO stage, the HDLT exhibited a significantly higher chlorophyll b concentration than the LDLT in 2023, and the LDET and HDET in 2024.

The carotenoid concentration was significantly affected by PPD at the PF and BO stages in 2023, and by TT at the PF stage in 2023 and at the PB and BO stages in 2024 (

Table 10. Effects of plant population density and topping time on the carotenoid concentration in main stem functional leaves (mg g⁻¹).

Treatment	2023			2024
	PF (DAS 113)	PB (DAS 149)	BO (DAS 169)	PF (DAS 121)	PB (DAS 148)	BO (DAS 165)
LDET	0.21 ± 0.01 b	0.23 ± 0.02 a	0.20 ± 0.01 b	0.25 ± 0.04 a	0.23 ± 0.01 b	0.15 ± 0.01 b
LDLT	0.18 ± 0.01 c	0.22 ± 0.01 a	0.21 ± 0.01 ab	0.20 ± 0.02 a	0.25 ± 0.02 ab	0.23 ± 0.02 a
HDLT	0.18 ± 0.01 c	0.21 ± 0.01 a	0.22 ± 0.01 a	0.24 ± 0.04 a	0.26 ± 0.01 a	0.21 ± 0.01 a
HDET	0.24 ± 0.01 a	0.22 ± 0.01 a	0.23 ± 0.01 a	0.27 ± 0.02 a	0.22 ± 0.02 b	0.15 ± 0.02 b
LD mean	0.19 ± 0.02 B	0.22 ± 0.01 A	0.20 ± 0.01 B	0.23 ± 0.04 A	0.24 ± 0.02 A	0.19 ± 0.05 A
HD mean	0.21 ± 0.03 A	0.21 ± 0.01 A	0.22 ± 0.01 A	0.25 ± 0.03 A	0.24 ± 0.03 A	0.18 ± 0.04 A
ET mean	0.23 ± 0.02 x	0.22 ± 0.01 x	0.21 ± 0.02 x	0.26 ± 0.03 x	0.22 ± 0.01 y	0.15 ± 0.01 y
LT mean	0.18 ± 0.01 y	0.21 ± 0.01 x	0.21 ± 0.02 x	0.22 ± 0.04 x	0.26 ± 0.02 x	0.22 ± 0.02 x
	Source of variance (F)
PPD	12.70 *	1.45	11.23 *	2.59	0.81	0.54
TT	82.17 **	3.21	0.02	3.25	12.01 *	56.83 **
PPD × TT	8.72 *	0.23	1.41	0.29	1.12	0.24

Note: PF: peak flowering; PB: peak boll-setting; BO: boll-opening; DAS: days after sowing. Data are represented as mean ± SD. Different lower-case and upper-case letters indicate significant differences at p < 0.05 among four individuals and between two plant density treatments across two topping time levels, respectively. Letters x and y indicate significant differences at p < 0.05 between two topping time treatments across two plant density levels. *, ** denote significance at the probability of 0.05 and 0.01 levels, respectively.

). Compared to the LD treatment, the HD treatment significantly increased the carotenoid concentration by 10.53% and 10.0% at the PF and BO stages in 2023, respectively, (Table 10). The carotenoid concentration in the ET treatment was elevated by 27.78% at the PF stage in 2023, but declined by 23.08% and 31.82% at the PB and BO stages in 2024 than that in the LT treatment, respectively. At the BO stage, The HDLT treatment displayed a significantly higher carotenoid concentration than the LDET in both years and the HDET in 2024 (Table 10).

3.5. Soluble Protein and Total Nitrogen

The soluble protein concentration was significantly affected by TT during the PF to BO stages in both years except the BO stage in 2023, and by PPD at the PF stage over two years (

Table 11. Effects of plant population density and topping time on the soluble protein concentration in main stem functional leaves (mg g⁻¹).

Treatment	2023			2024
	PF (DAS 113)	PB (DAS 149)	BO (DAS 169)	PF (DAS 121)	PB (DAS 148)	BO (DAS 165)
LDET	9.01 ± 0.38 b	6.58 ± 0.61 a	6.22 ± 0.99 b	12.93 ± 0.41 a	11.07 ± 0.69 b	10.79 ± 1.91 ab
LDLT	8.02 ± 1.52 b	7.51 ± 0.44 a	6.92 ± 0.75 b	8.97 ± 0.29 c	13.34 ± 0.39 a	13.44 ± 1.51 a
HDLT	8.74 ± 0.62 b	7.27 ± 0.97 a	8.87 ± 0.71 a	10.66 ± 1.55 b	13.55 ± 0.66 a	13.30 ± 1.48 a
HDET	11.31 ± 1.24 a	5.15 ± 0.39 b	6.86 ± 0.95 b	13.71 ± 0.19 a	11.38 ± 1.77 b	8.32 ± 0.74 b
LD mean	8.52 ± 1.13 B	7.04 ± 0.70 A	6.57 ± 0.87 A	10.95 ± 2.19 B	12.20 ± 1.34 A	12.12 ± 2.12 A
HD mean	10.03 ± 1.66 A	6.21 ± 1.34 A	7.87 ± 1.33 A	12.18 ± 1.94 A	12.47 ± 1.68 A	10.81 ± 2.92 A
ET mean	10.16 ± 1.50 x	5.87 ± 0.91 y	6.54 ± 0.94 x	13.32 ± 0.52 x	11.23 ± 1.21 y	9.56 ± 1.88 y
LT mean	8.38 ± 1.11 y	7.39 ± 0.69 x	7.89 ± 1.25 x	9.81 ± 1.36 y	13.44 ± 0.50 x	13.37 ± 1.34 x
	Source of variance (F)
PPD	6.26 *	4.99	6.85	6.80 *	0.19	2.39
TT	8.67 *	16.84 **	7.42	55.03 **	14.06 **	20.16 **
PPD × TT	1.7	2.55	1.74	0.91	0.01	1.87

Note: PF: peak flowering; PB: peak boll-setting; BO: boll-opening; DAS: days after sowing. Data are represented as mean ± SD. Different lower-case and upper-case letters indicate significant differences at p < 0.05 among four individuals and between two plant density treatments across two topping time levels, respectively. Letters x and y indicate significant differences at p < 0.05 between two topping time treatments across two plant density levels. *, ** denote significance at the probability of 0.05 and 0.01 levels, respectively.

). The ET treatment upregulated the soluble protein concentration by an average of 28.81% at the PF stage, but downregulated it by an average of 18.51% at the PB stage over two years, and by 17.11% at the BO stage in 2024. The HD increased the soluble protein concentration by an average of 13.91% at the PF stage in both years compared to the LD treatment. The HDLT treatment exhibited a significantly higher soluble protein concentration than the HDET during the PB to BO stages in both years, the LDET at the PB stage in 2024 and at the BO stage in 202, and the LDLT at the BO stage in 2023.

Total nitrogen content was significantly affected by TT during the PF to PB stages in 2023 and at the PF and BO stages in 2024, and by PPD at the PF stage in 2024 and at the BO stage in 2023 (

Table 12. Effects of plant population density and topping time on the total nitrogen concentration in main stem functional leaves (%).

Treatment	2023			2024
	PF (DAS 113)	PB (DAS 149)	BO (DAS 169)	PF (DAS 121)	PB (DAS 148)	BO (DAS 165)
LDET	3.35 ± 0.18 ab	2.82 ± 0.25 a	2.46 ± 0.18 b	3.77 ± 0.21 ab	3.42 ± 0.27 a	2.60 ± 0.10 b
LDLT	3.10 ± 0.20 b	3.25 ± 0.17 a	2.55 ± 0.21 b	2.99 ± 0.13 c	3.87 ± 0.12 a	3.05 ± 0.06 a
HDLT	2.95 ± 0.31 b	3.24 ± 0.24 a	2.94 ± 0.19 a	3.53 ± 0.37 b	4.01 ± 0.56 a	3.05 ± 0.05 a
HDET	3.72 ± 0.17 a	2.89 ± 0.37 a	2.58 ± 0.15 b	4.07 ± 0.11 a	3.73 ± 0.21 a	2.67 ± 0.09 b
LD mean	3.23 ± 0.21 A	3.04 ± 0.30 A	2.50 ± 0.18 B	3.38 ± 0.45 B	3.64 ± 0.31 A	2.83 ± 0.25 A
HD mean	3.34 ± 0.48 A	3.07 ± 0.34 A	2.76 ± 0.25 A	3.80 ± 0.39 A	3.87 ± 0.41 A	2.86 ± 0.22 A
ET mean	3.54 ± 0.26 x	2.86 ± 0.29 y	2.52 ± 0.16 x	3.92 ± 0.22 x	3.58 ± 0.28 x	2.64 ± 0.09 y
LT mean	3.03 ± 0.25 y	3.25 ± 0.19 x	2.75 ± 0.28 x	3.26 ± 0.39 y	3.94 ± 0.37 x	3.05 ± 0.05 x
	Source of variance (F)
PPD	0.77	0.03	5.98 *	9.88 *	1.38	0.59
TT	16.03 **	6.26 *	4.41	24.72 **	3.57	90.98 **
PPD × TT	4.36	0.06	1.65	0.75	0.2	0.59

Note: PF: peak flowering; PB: peak boll-setting; BO: boll-opening; DAS: days after sowing. Data are represented as mean ± SD. Different lower-case and upper-case letters indicate significant differences at p < 0.05 among four individuals and between two plant density treatments across two topping time levels, respectively. Letters x and y indicate significant differences at p < 0.05 between two topping time treatments across two plant density levels. *, ** denote significance at the probability of 0.05 and 0.01 levels, respectively.

). Either at the PF stage in 2024 or at the BO stage in 2023, the HD treatment displayed a significantly higher total nitrogen content relative to the LD treatment. The ET treatment showed a greater total nitrogen content at the earlier PF stage over two years, but smaller at later stages such as the PB stage in 2023 and the BO stage in 2024 compared to the LT treatment. At the BO stage, the total nitrogen concentration in the HDLT treatment was 13.95–19.51% (p < 0.05) higher than the three others in 2023, and 14.23–17.31% (p < 0.05) than the two ET treatments (LDET and HDET) in 2024.

3.6. Solar Radiation Capture

PPD, TT, and their interaction significantly affected the radiation interception rate during the PF to BO stages in 2024 except PPD at the PB stage (

Table 13. Effects of planting density and topping time on radiation interception rate by cotton canopy (%).

Treatment	2023			2024
	PF (DAS 113)	PB (DAS 149)	BO (DAS 169)	PF (DAS 121)	PB (DAS 148)	BO (DAS 165)
LDET	96.95 ± 0.72 a	97.73 ± 0.62 b	90.29 ± 0.62 b	76.05 ± 1.19 c	95.58 ± 0.61 b	90.65 ± 1.21 a
LDLT	95.72 ± 2.82 a	96.27 ± 0.28 c	90.87 ± 1.03 b	69.85 ± 1.26 d	94.38 ± 1.20 c	93.76 ± 0.82 a
HDLT	97.33 ± 2.10 a	98.89 ± 0.24 a	96.24 ± 0.57 a	83.05 ± 1.41 b	94.17 ± 1.35 c	93.17 ± 4.76 a
HDET	97.61 ± 2.50 a	97.86 ± 0.29 b	81.01 ± 4.16 c	86.06 ± 1.43 a	97.07 ± 0.88 a	83.54 ± 3.51 b
LD mean	96.34 ± 2.07 A	97.00 ± 0.89 B	90.58 ± 0.87 A	72.95 ± 3.45 B	94.94 ± 1.13 A	91.93 ± 1.89 A
HD mean	97.47 ± 2.25 A	98.37 ± 0.60 A	88.62 ± 8.44 B	84.56 ± 2.07 A	95.91 ± 2.03 A	88.89 ± 6.43 B
ET mean	97.35 ± 1.97 x	97.80 ± 0.47 x	85.65 ± 5.61 y	81.06 ± 5.37 x	96.52 ± 1.10 x	87.49 ± 4.37 y
LT mean	96.69 ± 2.46 x	97.58 ± 1.39 x	93.55 ± 2.91 x	76.45 ± 7.01 y	93.63 ± 1.80 y	94.27 ± 0.74 x
	Source of variance (F)
PPD	2.14	95.00 **	5.11 *	459.18 **	0.06	21.99 **
TT	1.07	0.26	76.99 **	72.29 **	26.39 **	114.29 **
PPD × TT	0.79	59.21 **	67.77 **	8.75 *	7.97 *	36.16 **

Note: PF: peak flowering; PB: peak boll-setting; BO: boll-opening; DAS: days after sowing. Data are represented as mean ± SD. Different lower-case and upper-case letters indicate significant differences at p < 0.05 among four individuals and between two plant density treatments across two topping time levels, respectively. Letters x and y indicate significant differences at p < 0.05 between two topping time treatments across two plant density levels. *, ** denote significance at the probability of 0.05 and 0.01 levels, respectively.

). Additionally, all main effects and their interaction were significant at the PB and BO stages in 2023 except TT at the PB stage. The radiation interception rate at the BO stage was significantly greater in the LT than in the ET treatment in both years. The HDLT treatment exhibited 5.91–18.80% (p < 0.05) higher radiation interception rate during the PB to BO stages than the three others in 2023, and 2.78–11.53% higher at the BO stage than the LDET and HDET treatments in 2024, but 1.48–2.99% (p < 0.05) lower at the PB stage than the two ET treatments in 2024.

4. Discussion

4.1. Increasing PPD Combined with Later Topping Improves Seed Cotton and Lint Yields Through the Enhancement of the Biological Yield

Both increasing PPD to 8 plants m⁻² and delaying topping to remain 15 simpodials per plant increased seed cotton yield, thus the combination of HD and LT (HDLT) exhibited the highest seed cotton yield among the four combinations involving plant density and topping time both at two levels, as was for the lint yield since lint percentage was not changed by either PPD or TT (Table 1). In view of dry matter accumulation and partitioning, the higher seed cotton yield in the HDLT is the result of the enhanced biological yield compared to the others. By contrast, the harvest index in the HDLT was the lowest (Table 1). The accumulation rate (average and maximum) during the exponential biomass growth phase in the HDLT was the greatest by examining the dynamic of the dry matter production in the roots, aerial parts, and the total plant over the entire growth cycle (Table 3, Table 4 and Table 5), which should be responsible for the highest biological yield. Under deficit irrigation, increasing PPD can achieve a comparable yield to medium PPD under regular irrigation through enhancing biological yield and harvest index [21]. With respect to yield components, neither line percentage nor boll weight was affected by either PPD or TT; boll density increased with increasing PPD or delaying topping (Table 1). Thus, a higher boll density was detected in the HDLT than in the others (Table 1). Increasing PPD commonly promotes boll density but compromises boll weight, and lint percentage is less affected [3,4,5], which agrees with our results.

4.2. Increasing PPD Combined with Later Topping Delays Leaf Senescence as Indicated by Higher Photosynthetic Parameters, Soluble Protein and Total Nitrogen Contents from the Peak Boll-Setting Onwards

Cotton maturity performance can be assessed by the following indicators: Pn, SPAD value, photosynthetic pigments (Chlorophyll a, b, and carotenoid), soluble protein, and total nitrogen [22,23,24,25,26,27,28]. The TT effects on all leaf senescence-related attributes (including photosynthetic parameters) were detected during the reproductive growth period of at least one year. In general, those parameters were higher at the PF stage, but changed into lower values at the later BO stage in the ET treatment (Table 6, Table 7, Table 8, Table 9, Table 10, Table 11 and Table 12). More specifically, the SPAD value with the LT treatment was lower at the PF and PB stages but higher at the BO stage, and the soluble protein and total nitrogen contents were lower at the PF stage but higher at the PB and BO stages in the LT treatment (Table 7, Table 11 and Table 12). These results indicate that the earlier topping hastened cotton maturation and leaf senescence, which accords with the reports by Li et al. [15,29,30]. Cotton maturity varied with TT is due to the observation that topping leads to more photoassimilates partition to reproductive forms and less to vegetative structures; thus, vegetative growth is inhibited [12,17,31]. The effects of PPD on the photosynthetic parameters at certain growth stages were significant in 2023, including SPAD value at the PF to PB stages (Table 7), chlorophyll a concentration at the PB stage (Table 8), chlorophyll b concentration at the PB to BO stages (Table 9), and carotenoid concentration at the PF and BO stages (Table 10). In addition, the soluble protein content at the PF stage was significantly affected by the PPD in both years (Table 10), and so was the total nitrogen content at the PF stage in 2024 and at the BO stage in 2023 (Table 11). As expected, the attributes were higher in the HD than in the LD when significant differences were detected. Delayed leaf senescence due to increasing PPD has been reported [3,4,5,26,32]. This response may be associated with the source-sink relation that is changed by PPD. Higher PPD inclines to decrease cotton boll number more than leaf area per plant, thus leading to lowered bolls per unit leaf area [33], which is helpful to alleviate premature senescence. Conversely, increasing boll load per plant results in less photoassimilates partition to roots, limiting root development and triggering leaf senescence [23]. High density upregulates the contents of auxin and cytokinins promoting auxin polar transportation in cotton main-stem tips [3], which contributes to delaying senescence. The interaction of PPD and TT was significant only for SPAD value at the PF, chlorophyll a, b at the PF and BO, and carotenoid at the PF in 2023 (Table 7, Table 8, Table 9 and Table 10). Due to the integration of the effects of the HD along with the LT, the HDLT treatment performed better than the three others for all leaf senescence-related attributes at the BO stage (Table 6, Table 7, Table 8, Table 9 and Table 10).

4.3. Increasing PPD Combined with Later Topping Enhances Solar Radiation Capture and Photosynthetic Efficiency at the Cotton Late Growth Stage

There was a higher radiation interception rate with the HD treatment at the PF stage in 2024, and at the PB stage in 2023, which is consistent with the results reported by Brodrick et al. [34] and Gu et al. ([35], but the opposite was true of that at the BO stage in both years (Table 13). The plausible explanation is that the upper (the ninth) and the middle (the sixth) fruiting branches were elongated by the loose planting than the dense planting (Table 2), which contributes to a higher radiation interception rate in the LD treatment at the late growth phase. A lower light capture in late cotton growth may be beneficial to the high-density population where ventilation and light transmission in the lower canopy are supposed to be improved, thus minimizing boll deformities and decay. The ET significantly enhanced the radiation interception rate during PF to PB stages in 2024, but it was reversed at the BO stage, which can be explained that earlier decapitation of the main tips facilitates the earlier ontogeny of new phytomers on the branches below that helps to capture more solar radiation, but at the later growth stage, such as the BO stage, more leaf shedding due to premature senescence caused by the ET treatment may result in a smaller radiation interception rate (Table 13). PPD by TT interaction effects on radiation interception rate were detected during the PF to BO stages in both years (Table 13). Compared to the others, the HDLT treatment displayed a higher radiation interception rate at the BO stage over two years.

As mentioned above, the HDLT exhibited an increase in those leaf senescence-related characteristics at the BO stage over the three others (Table 6, Table 7, Table 8, Table 9 and Table 10) including Pn, SPAD value, chlorophyll a, b, soluble protein, and total nitrogen, which can also reflect a higher photosynthetic efficiency with the HDLT. Soluble protein and chlorophyll concentrations in cotton leaves are positively correlated with photosynthetic potential [27,36]. Approximately 50% of the leaf’s soluble protein is Rubisco (ribulose 1.5 bisphosphate carboxylase–oxygenase) which is a key rate-limiting enzyme for photosynthesis [27]. In addition, leaf nitrogen concentration is closely and positively associated with radiation use efficiency [34,37,38].

5. Conclusions

SFBC is a new plant-type cotton with shortened internodes in fruiting branches suitable for dense planting. PPD and TT are two essential agronomic practices to regulate cotton growth and development, and thus achieve a final yield. The effects of PPD and TT on the SFBC yield were explored over the entire reproductive growth period in terms of dry matter production, leaf senescence advancement, and canopy radiation interception. Compared to the other three combinations of PPD and TT at two levels, the combination of HD and LT (HDLT) displayed a greater biological yield but a lower harvest index, and equivalent lint percentage, which together accounted for a higher lint yield. The greater plant biomass is largely attributed to the higher average accumulation rate during the rapid biomass increasing period. Due to delayed leaf senescence caused by the HD and the LT, the HDLT performed better in those leaf senescence-related attributes at the PB and OB stages. Moreover, these improved attributes also contributed to a higher radiation interception rate and photosynthetic efficiency at the late growth stage such as the BO stage. Taken together, combining high density with later topping tends to increase the lint yield of SFBC through increasing dry matter accumulation, delaying leaf senescence, and enhancing canopy radiation interception rate at the late growth.