JIA-2019-11

2493 NIE Jun-jun et al. Journal of Integrative Agriculture 2019, 18(11): 2492–2504 1. Introduction The lint yield of cotton is calculated by multiplying three yield components: the number of bolls per unit ground area, the average boll weight, and the lint percentage. Numerous studies have shown that the individual boll weight increased in the 1970s and 1980s, the boll number per plant and the lint percentage increased in the 1990s, and further improvement in the individual boll weight in this century resulted in an increase in cotton yield (Tang et al . 1993; Kong et al . 2000; Zhang et al . 2003; Mao 2010; Iqbal et al . 2013; Shakeel et al . 2015). Therefore, knowledge of how yield formation change is vital for high-yield and fine-quality cultivation and the selection of new varieties in cotton breeding. Cotton has an infinite flowering and boll-setting habit with a complex shape and living state of leaves, bolls, and branches, and the distribution of bolls in various positions and their share of economic production are intimately linked to the final yield (Constable 1991; Liu et al . 2015). Additionally, reproductive growth does not correspond with apical dominance in cotton, and the generative organs are distributed within the cotton canopy such that cotton plants have many reproductive growth centers distributed throughout the canopy. For this reason, cotton yield and quality are more susceptible to environmental conditions (e.g., light, temperature, rainfall, drought, and hail) than gramineous crops (Schurr et al . 2006). Simultaneously, lint yield and quality are a function of boll-setting, which is significantly affected by the boll-setting period, the boll spatial position, and the physiological state of the cotton plant. Previous research showed that the production of fruiting sites and fiber properties were spatially correlated (Wilkerson and Hart 1996; Johnson et al . 2002). These authors also noted that the micronaire exhibited a moderate degree of spatial variability, and strength showed the lowest degree of variability. Optimization of cotton fruiting can be realized through the formation of more bolls in the best boll-setting period and in the best spatial positions of a cotton plant with the healthiest physiological state (Dong et al . 2014). In general, the ideal spatial positions for the best cotton boll quality, the largest boll weight, and the finest fiber quality are the middle branches and inner node bolls (Tan 1992). Considerable research has shown that an increase in bolls on the middle fruit branches and internal parts resulted in increased cotton yield, but studies have consistently focused on the variation of Bt and non-Bt cotton cultivars (Blaise 2006; Hofs et al . 2006) and the effect of cultural practices, e.g., plant density (Mao et al . 2015; Wang et al . 2016; Khan et al . 2017), water and fertilizer application (Read et al . 2006; Papastylianou and Argyrokastritis 2014; Dai et al . 2015; Zhang et al . 2015; Chen et al . 2016), mepiquat chloride utilization (Biles and Cothren 2001; Mao et al . 2015; Zhao et al . 2017), and sowing times (Liu et al . 2015; Lu et al . 2017). However, few studies have been conducted on the relationship between yield composition (boll number, boll weight, and lint percentage) and boll spatial distribution among different Bt cotton varieties since insect-resistant cotton breeding was initiated in the 1990s. Thus, three insect-resistant cotton varieties (Daizimian 99B (99B), Lumianyan 21 (L21), and Jimian 169 (J169)), which are local and prevalent cultivars that have been recommended by cotton breeders in the Yellow River region in China during different periods, were selected in this study. The objectives of this research were to investigate the range in agronomic characteristics and boll spatial distribution related to cotton yield and fiber quality among the different Bt cotton cultivars and to reveal the reasons for the differences in lint yield and quality. 2. Materials and methods 2.1. Experimental site and cultivars Field experiments were conducted from 2013 to 2017 at the State Key Laboratory of Crop Biology and the experimental farm of Shandong Agricultural University (36°10´N, 117°09´E, 158 m a.s.l.), Tai’an, Shandong Province, China. The experimental soil type is a brown loam. The concentrations of organic matter, total N, rapidly available phosphate, and rapidly available potassium in the upper 20 cm of soil were 16.64–18.33 mg kg −1 , 1.07–1.21 mg kg −1 , 37.47–42.63 mg kg −1 , and 120.98–128.35 mg kg −1 , respectively. The climate is temperate and monsoonal with an average annual temperature of 13°C, rainfall of 697 mm, sunshine duration of 2627 h, and a frost-free period of 195 d. The average temperature (°C) and rainfall (mm) of the cotton growing seasons from 2013–2017 are shown in Fig. 1. The total rainfall was 591.7, 310.6, 343.1, 509.2, and 452.0 mm during the cotton growing seasons in the five respective years. Cotton is usually planted in late-April and harvested at the end of October, with a growth period of nearly six months. Three Bt ( Bacillus thuringiensis ) cotton varieties with Keywords: Bt cotton, yield formation, boll spatial distribution, lint yield, fiber quality