Effect Of Mass Selection On The Quality Improvement In Maize Populations

In china,the narrow genetic base has becomed already the limitative factors for the break throughing development in maize breeding.Germplasm enhancement,improvement and innovation is the only way to solve this problem .To study the quality improvement of maize population is significantly for the development of germplasm and quality improvement .The objectives of this study were to study the effect of high starch and protein improvement in maize populations with the combinative methods of mass seletion and NIRS .To improve four and three cycles selection for the Yuzong Number 5 Populations,Respectively.To evaluaate the feasibility for improvement and combing ability.using PLS methods,to vindicate the strach and the protein ration model for maize.The primary results were as follows:1.The results of measuing quality at different cycles showed that ,after 4 selections in high starch population,the conect of starch from 69.39% of basic population to 72.31% ,Which was raised 2.92%, And each reculrence gain was 0.73%.After 3 selections in high protein Population , the conect of pretein from 11.29% of basic population to 12.04% ,Which was raised 0.75%,And each reculrence gain was 0.25%.Indicate that method was effctive with the combinative methods of mass seletion and NIRS.2.About relativity qulatity ,gain strach content were significantly negative correlated with gain protein and lysine content, lysine content was significantly positive correlated with gain protein,gain fat content were positive correlated with gain protein and lysine content,but was lesser negative correlated with gain strach content.Four gain quality traits were positive or negative correlated to yiele traits,suggesting that it is possible to a certain extent to improve quality of hybrid without loss in yield.3.Coming ability indicate that,the GCA groups in most traits of quality ,economic and agronomic and were significant or extremely significant difference. The SCA of many quality and economic traits were detected that was not significant or extremely significant between groups and test combinations.The strach content of the cross 10 was the highest(72.23%) among high strach population, and The protein content of the cross 26 was the highest(11.91%) among high protein population. The SCA of grain yield per plant was extremely positively significant for the cross 26.and the SCA of protein content was positively significant .This crosses might have Poteintial utilization in production.4.Analysis of Quality at different Spike Position of the population.The content of the four main quality was obvious different in different parts of the spike . in high strach C4 population, Kernels at the basal part of a spike exhibited the greatest protein and lysine, followed by those at the middle part, and those at the apical part the least .The order of the starch content was the middle part﹤the apical part﹤the basal part.But the the basal part was greatest for the fat content,Secondly, the apical part,the middle part was the least.For high protein M3 population ,the mutative tide of protein and lysine content were the same as C4 population . The order of the starch content was the apical part﹤the middle part﹤the basal part.But the the middle part was greatest for the fat content,Secondly, the apical part,the part basal was the least.5.Different spectral data preprocessing methods and spectral regions have influences on the performance of model established.Calibration equations for quantitatively determination of crude starch and crude prorein content were developed by partial least square regression (PLS) of chemical values on NIRS dataand tested through external validation.The result showed that the data protreatment of first deravitive, first deravitive with straight line subtration achieved the best effects on starch content and protein content ,respectively ,and gave the ditermination coefficients of cross valadition 67.10%,90.15% and the root mean square error of cross valadition 1.06,0.36.In addition , respectively ,60 and 45 progenies of F1 were verified to further evalate their performance .Resulted indicated that the correlations between the predicted NIRS results and those obtained from traditional chemical methods of protein content was very strong,with the correlation coefficients generally 0.92,whereas lower coefficients (0.76)was obtained with models of starch content from reflectance spetra.