Abstract:

The present study was carried out to screen comprehensive feeding evaluation indexes of peanut straw and to identify peanut germplasm resources with highly feeding value. Forty peanut varieties and breeding lines were used for detecting protein, ash, calcium, phosphorus, neutral detergent fiber (NDF), acid detergent fiber (ADF), acid detergent lignin (ADL) and in vitro dry matter digestion (IVDMD) of stalks and leaves by employing near infrared reflectance spectroscopy technology respectively, as well as calcium, phosphorus and protein of seeds. By comparing content of nutrients between peanut leaves and stalks, peanut leaves contained 5% higher content of protein, approximately equal content of ash, calcium and phosphorus, and lower content of NDF, ADF, ADL, reflecting that peanut leaves have higher feeding value than peanut stalks. There were significant and positive correlation between content of calcium and phosphorus in peanut stalks, leaves and seeds respectively, indicating that absolute synergistic absorption existed. Furthermore, both content of calcium and phosphorus were significantly negative with content of protein in stalks and leaves respcetively, while the both content of calcium and phosphorus were significantly positive with content of protein in peanut seeds. Five comprehensive indexs with 91.12% accumulative contribution of 18 indexes of peanut stalks and leaves were extracted by PCA analysis. Based on comprehensive score (namely D) of each peanut variety, 40 peanut varieties and lines were divided into 3 groups by cluster analysis, the first group contained 10 peanut varieties with high comprehensive feeding scores (CFS) , the second group contained 17 peanut varieties with medium CFS, and the third group contained 13 peanut varieties with low CFS. The optimal mathematical model of peanut feeding evaluation was established on account of D scores and 18 indexes of peanut stalks and leaves, namely D=0.014X₁+0.012X₂-0.003X₅+0.009X₇-0.023X₈+0.009X₁₀+0.011X₁₆-1.582 (R²=0.994, F=764.329, P=0.000), which showed high estimation accuracy above 88%. Hence, 7 indexes were screened out, which consisted of five indexes of stalks (protein, ash, NDF, DM and ADL) and 2 indexes of leaves (protein and DM).

Key words: peanut, straw, seeds, multivariate statistical methods, comprehensive feeding evaluation, nutrient