Study On Crude Protein Content And Accumulation Of Different Genotypes Of Forage Sorghum

In this paper, 141 forage sorghum was used to study the crude protein content of stem, leaf and spike of different types forage sorghum at the filling stage. The difference of CP content of different parts of sorghum was compared. The correlations of stem, leaf and spike CP content with other traits were analyzed. The CP content and CP yield of two years was compared. Six sorghum cultivars (line) were chosen to study the growth and accumulation of CP. The NIRS method for the determination of moisture and CP of different parts of sorghum was preliminarily studied. The main results obtained are as follows:1. The sorghums chosen were varied in Mid-rid color, days from sowing to flowering and the thickness of wax. The average crude protein content of stem, leaf and spike of was 3.46%, 11.10%, 10.61%, with coefficients of variation of 87.83%, 54.64% and 14.04%, respectively. The distributions of stem, leaf and spike CP content are nearly normal. Crude protein contents of different parts of sorghum descend in the order of leaf> spike>stem. Protein yield of stem, leaf, spike, and whole-plant averaged 1.487g, 2.371g, 1.581g, 5.440g, respectively, and their coefficients of variation were all above 30%. CP yield of the three parts descended in the order of: leaf>stem, spike.2. The CP contents of stem, leaf and spike were positively correlated to each other very significantly. The CP content of whole plant has significant positive correlations with CP contents of stem, leaf and spike, dry to fresh ratio of leaf and spike, and negative correlations with height, ratio of stem dry to whole plant.3. Variance analysis showed that the differences of leaf and stem CP content among different types sorghum were very significant, and spike CP content was significantly different among sorghum types. According to the results of multiple comparisons, the stem CP content of different types of sorghum descended in the order: sudan grass, lines>grain sorghum>sweet sorghum. The leaf CP content of different types of sorghum descended in the order: grain sorghum, sudan>line>sweet sorghum. The spike CP content of different types of sorghum descended in the order: grain sorghum> sudan grass, lines>sweet sorghum. The whole-plant CP content of different types of sorghum descended in the order: grain sorghum, sudan grass and lines>sweet sorghum.4. The CP contents of stem, stem, spike and whole plant of two years were not significantly different. Except that spike crude protein yield was not significantly different between years, the crude protein yield of stem, leaf and whole plant between years were significantly different.5. The CP accumulation of leaf was different from that of stem, spike and whole-plant. The leaf CP content decreased linearly during the whole growth period, while the stalk, spike and whole-plant CP contents were presented L-shaped trend, that is the CP contents decreased rapidly at the early stage and slowly later. The dynamics of CP yield of different parts showed double peak trend, and the latter was higher than the former.6. The NIRS models for the determination of moisture of sorghum stem, leaf and spike were established. The coefficients of determination R2 and RMSEC of each validation model were 0.9405, 0.1698, 0.9785, 0.2182, 0.9445, 0.0868 for stem, leaf, and spike, respectively; and coefficients of determination R2 and RMSEP of each prediction model were 0.9038, 0.2047, 0.9754, 0.2348, 0.9093, 0.1118 respectively.7. With the samples without drying, the models for stem, leaf and spike CP content and CP content (dry basis) were established. All the coefficients of determination for validation and prediction were higher than 0.88 (0.8828-0.98), and the error is less than 0.4 (0.0665-0.3989). With the dry samples, the models for determination of stem, leaf and spike CP content (dry basis) were established. The coefficients of determination for validation were higher than 0.92 (0.9275-0.9785), RMSEC were less than 0.38 (0.2755-0.3749), and coefficients of determination for prediction were above 0.85 (0.8547-0.9731), RMSEP were less than 0.52 (0.3922-0.5140).We can see that models established with dry samples are not as good as the ones established with samples not dried.