Research & Utilization Of The Relationship Among The Quantitative Traits Of High-yield Hybrid Cotton

Based on the regional experiment results on hybrid cotton varieties in Henan Province from 2007 to 2009, this paper discusses the development trend of hybrid cotton breeding (i.e. breeding direction) and the degree of variation of their traits and utilizes analyzing methods including phenotypic correlation, canonical correlation, path analysis, optimal regression analysis and principal component analysis for studying the relationship among various kinds of traits of hybrid cotton in a multi-perspective way, to provide useful guidance for favorable composition & filtering with high advantages, and to serve as the basis for proposing the main direction high-yield, high efficient hybrid cotton breeding. The results show that:1.The development trend of hybrid cotton is: the growth period is slightly extended; morphological traits such as plant height, number of fruit branches, location of the nodes of the first fruit branch, seed index etc are in increasing trend, while disease resistance is decreasing; among the factors on production yield, the highest rate of increase is in the number of bolls of a single plant; the overall level of fiber yield is relatively high, yet in decreasing trend, while the production level is increasing; cotton fiber length, Micronaire, breaking tenacity, uniformity, reflectance, spinning consistency index are all in a trend of increasing, but a Micronaire values above 5.0 indicates an unfavorable trend to thicker fiber.2. Analysis on traits variation shows that: among yield components, the largest variation coefficient is on the number of bolls per plant, followed by that of boll weight and lint percentage. All variation coefficients on indexes of reproductive morphological traits are relatively small, especially that of growth period is only 2.41%. On disease resistance traits, variation coefficients of both Fusarium vasinfectum index and Verticillium alboatrum index are bigger, especially that of Fusarium vasinfectum, which is 66.30%. Variation coefficients on quality traits are small; in descending order, they are: yellow degree>spinning index>Micronaire value>specific strength> fiber length> elongation rate>reflectance>uniformity.3. Phenotypic correlation and path analysis show that: the number of bolls of a single plant affects the yield most directly and overall significantly; secondary factors are lint percentage and boll weight. For reproductive morphological traits, the degrees of relevance with yield are as follows in descending order: growth period (0.615 **) > fruit branch number (0.601 **) > the location of first fruiting branch (-0.459 *) > plant height (0.391) > seed index (-0.137); the three yield components are positively related to each other with varying degrees, among which the boll number and lint percentage are most closely related (0.544 **), but the correlation between boll weight and effective bolls per plant and lint percentage are not significant; effective bolls per plant has a relatively greater positive correlation with growth period (0.602 **), plant height, and number of fruit branches, and a significant negative correlation with the location of first fruit branch (-0.427 **).The correlations between fiber quality and yield are in the order: elongation rate (-0.77 **) > specific strength (-0.370) > length (-0.360) > spinning index (-0.310) > uniformity of seed index (0.220) and Micronaire value (0.220), i.e., they are mostly bigger negative correlations.4. The correlations between lint yield and the 4 groups of traits are as follows in descending order: yield components (0.8729 **), quality characters (0.8119 *), morphological traits (0.7527 *), and disease resistance characters (the canonical correlation of the latter was significant). Correlation between yield and yield components is mainly due to the correlation between effective bolls per plant and yield; correlation between yield and morphological traits is mainly caused by that of the number of fruit branches and yield; correlation between yield and quality characters is mainly caused by that of the elongation rate and yield.The correlation between reproductive morphological traits and yield traits is very significant, and is mainly induced by the negative correlation between effective bolls per plant and seed index, and the positive correlation between lint percentage and reproductive period; quality traits and reproductive morphological traits correlate most significantly, and it is mainly caused by that of uniformity and plant height; yield characters and quality traits are significantly correlated, and the main factor is due to the correlation between effective bolls per plant and spinning index; correlation between disease resistance traits and yield characters or morphological traits is not significant.5. Optimal regression analysis on yield and agronomic traits showed that the yield is significantly correlated with boll number per plant, lint percentage and number of fruit branches; the quantitative forecast equation is: Y = 13.05 +0.51 X1 +1.15 X3 +1.73 X7.6. Principal components show that: the principal component of cotton's quantitative traits are prominent; a value of 5 principal component m will cumulatively contribute 83.02% of the variance, i.e., among all the 17 characteristic roots of principal components, the first 5 contribute 83.02 percent of total variance; so these 5 components can basically reflect the overall characteristic of hybrid cotton. According to the absolute value of their eigenvectors the 5 principal components are chosen in descending order as follows: boll number per plant (0.3470), lint yield (0.3141), elongation (-0.3102), growth period (0.3081), number of fruit branches (0.2945).