Research on the relationship between the accuracy of spad value and sample number in growing period of cotton | 棉花生育时期SPAD值准确性与样本数的关系

chinois. 目的研究高产滴灌棉花观测样本数对SPAD诊断棉花氮素营养准确性的影响，为SPAD法指导棉花施肥提供理论依据。方法采用田间膜下滴灌方式，在390 kg#8226;hm-2施氮量下，设计5个氮肥运筹方案，监测棉花全生育期5个氮肥运筹方案下随机观测的60个SPAD值样本数，测定产量后分析高产施氮水平下不同样本数与标准偏差的变化关系。结果高产棉花各生育时期SPAD值的变异系数和标准偏差差异不大，但是存有明显不同；棉花蕾期、初花期、盛花期、花铃期、盛铃期、吐絮初期的SPAD值样本数（y）与标准偏差（x）的乘幂函数关系分别为：y=5632.2x-4.893（R2=0.7968），y=26456x-6.0437（R2=0.8732），y=50184x-6.5227（R2=0.9782），y=174890x-7.2601（R2=0.9655），y=27410x-5.743（R2=0.9182），y=16237x-5.3756（R2=0.942），并可通过y=ax-b（a，b均为参数）模型来预测SPAD值的变异程度及各生育时期诊断样本数量的多少；不同样本数下SPAD值和叶片含氮量线性相关性的高低排列顺序为：60个样本＞30个样本＞5个样本。结论用SPAD叶绿素仪诊断棉花各个生育时期氮素营养，测定样本数为26―60时样本数越大，SPAD均值准确性越高，SPAD均值与叶片含氮量的相关性就越好，SPAD值测定样本数与标准偏差呈乘幂函数关系。

anglais. Objective Research of the effects of observed samples of high yield drip irrigation cotton on N nutrition diagnosed by SPAD, to provide a theoretical basis for guiding fertilization of cotton with SPAD method. Method Adopting the method of drip irrigation under mulching in field, when N fertilizer rate was at 390 kg#8226;hm-2, 5 programs for N fertilizer application were designed in order to supervise the 60 samples’ SPAD values of 5 N fertilizer application randomly during the whole growing period of cotton. The changing relationship between sample number and standard deviation under N fertilizer application rate of high-yielding cotton was analyzed after the yield was measured, and the influence of sample number on the accuracy of SPAD value in the growing period of cotton was defined. Result Differences were observed between the coefficient of variation of SPAD value and standard deviation; the power function relations of sample number of SPAD value (y) and standard deviation (x) at the stages of bud, early blooming, fill blooming, blooming and boll-forming, full boll-forming and early wadding are as follows: y=5632.2x-4.893（R2=0.7968），y=26456x-6.0437（R2=0.8732），y=50184x-6.5227（R2=0.9782），y=174890x-7.2601（R2=0.9655），y=27410x-5.743（R2=0.9182），y=16237x-5.3756（R2=0.942）, and y=ax-b(a,b parameters) model could be used to predict the variation of SPAD value as well as sample number in every growing period. Linear dependence of SPAD value and N content of leaves is 60＞30＞5. Conclusion The sample number influences the accuracy of SPAD average value, the larger the sample number the more accurate the SPAD average value, and the pertinence between SPAD average value and N content of the leaves will be better. Sample number of SPAD value and standard deviation is a power functional relationship.