JIA-2019-11

2530 ZHOU Tian-mei et al. Journal of Integrative Agriculture 2019, 18(11): 2521–2533 using a single index (Zhang et al . 1992, 2004; Huang et al . 2012). Comprehensive indexes can reflect seedling quality more objectively(Yang and Zhou 2010). Most of the indexes that have been developed to assess seedling quality are relatively simple to measure, specifically, morphological characteristics (Pinto et al . 2011). These indexes also have good relations with successful field establishment (Dey and Parker 1997; Rose et al . 1997). In tobacco, correlation analysis was made between seedling quality and 17 single seedling quality indexes, plus 21 comprehensive evaluation indexes. On the basis of this analysis, composite indexes that correspond to eld success were assessed. They screened out six composite indexes that showed significant correlations with field success, which can be used as the vigorous seedling index model for evaluating the quality of tobacco seedlings (Bai et al . 2014). Yang and Zhou (2010) considered that [(Stem diameter/Plant height)×Dry mass of plant] could objectively reflect pumpkin seedling quality. Huang et al . (2012) showed that ([(Stem diameter/Plant height)+(Dry mass of root/dry mass of shoot)]×Dry mass of plant) may provide a more effective indicator of seedling performance. Based on these studies, and according to the rules of accuracy and easier determination, we assessed five composite indexes in this experiment. These five composite indexes were used to establish a simulation model of seedling quality based on TEP. The results showed that the root-shoot ratio and TEP curve fitting were not very good in tomato and cabbage seedlings, as the R 2 were only 0.7584 and 0.7511, respectively. Many authors argue that root morphology (root length, root volume, root mass, etc.) may provide a more effective indicator for seedling quality, especially in seedlings of forest plants (Ritchie and Dunlap 1980; Rose et al . 1997; Tsakaldimi et al . 2013). This may be because the root volume is larger in forest seedlings than in tomato and cabbage. For tomato and cabbage seedlings, root biomass was very small in the early stages of seedling growth; even 35 days after sowing, the dry mass of the root only ranged from 31.46 to 84.53 mg (Table 3), which may be a reason for the inaccuracy of the models. The curve fitting degree of healthy index models in our study was high; all R 2 values exceeded 0.95. However, for HI2 and HI3, simulated values showed greater deviation compared to the measured values, although HI1 model showed a high fitting accuracy with higher R 2 and lower RMSE. This may be because dry matter corresponds better to TEP than fresh matter. Yuan et al . (2006) considered that HI1 contains more information, especially dry matter accumulation, which is linked to crop dry matter partitioning and yield formation. The HI3 model has a less reliable prediction, possibly because root biomass is relatively low in the early stage of seedling growth, which will affect the accuracy of the model. Bai et al . (2014) argued that [(Stem girth/Plant height)×Total fresh mass×10] can be used as a vigorous seedling index model for evaluating the quality of tobacco seedlings, but fresh mass may be more susceptible 0 0.003 0.006 0.009 0.012 0 0.003 0.006 0.009 0.012 Measured G value Simulated G value 0 0.003 0.006 0.009 0.012 0 0.003 0.006 0.009 0.012 Measured HI1 value Simulated HI1 value 0 0.003 0.006 0.009 0.012 0 0.003 0.006 0.009 0.012 Measured G value Simulated G value 0 0.003 0.006 0.009 0.012 0 0.003 0.006 0.009 0.012 Measured HI1 value Simulated HI1 value A B C D Fig. 7 Comparison between simulated and observed values of G value and healthy index 1 (HI1) of cabbage seedlings.