大豆株高是由主效或微效基因控制的重要农艺性状。在已报道的株高QTL中，绝大部分定位区间较大，限制了大豆株高分子调控机制的解析。增加遗传图谱的标记密度会显著地提高QTL定位的效率和准确性。本研究利用双亲中黄13和中品03-5373及其衍生的241个重组自交系（RILs）全基因组重测序数据，构建一个包含4011个重组bin标记、总遗传距离为3139.15 cM的高密度遗传图谱，相邻bin标记间的平均距离为0.78 cM。比较基因组分析表明，所构建的遗传图谱与大豆参考基因组具有较高的共线性。基于此图谱，在6个环境中共检测到9个株高QTL，包括3个新位点（qPH-b_11，qPH-b_17和qPH-b_18）。其中，两个环境稳定主效QTL qPH-b_13和qPH-b_19-1可解释10.56%～32.7%的表型变异。qPH-b_13和qPH-b_19-1被精细定位到440.12 kb和237.06 kb的基因组区间，分别包含54和28个注释基因。进一步的拟南芥同源基因功能和候选基因表达分析表明，基因Glyma.13G292600和Glyma.19G194100分别为qPH-b_13和qPH-b_19-1的候选功能基因。

This paper estimates a stochastic frontier function using a panel data set that includes 4 961 farmer households for the period of 2005-2009 to decompose the growth of grain production and the total factor productivity (TFP) growth at the farmer level. The empirical results show that the major contributor to the grain output growth for farmers is input growth and that its average contribution accounts for 60.92% of farmer’s grain production growth in the period of 2006-2009, whereas the average contributions sourced from TFP growth and residuals are only 17.30 and 21.78%, respectively. The growth of intermediate inputs is a top contributor with an average contribution of 44.46%, followed by the planted area (18.16%), investment in fixed assets (1.05%), and labor input (-2.75%), indicating that the contribution from the farmer’s input growth is mainly due to the growth of intermediate inputs and that the decline in labor inputs has become an obstacle for farmers in seeking grain output growth. Among the elements consisting of TFP growth, the contribution of technical progress is the largest (32.04%), followed by grain subsidies (8.55%), the average monthly temperature (4.26%), the average monthly precipitation (-0.88%), the adjusted scale effect (-5.66%), and growth in technical efficiency (-21.01%). In general, the contribution of climate factors and agricultural policy factor are positive and significant.