刈割高度对草丛群体特征的影响的研究—典型变量分析在草地生态系统研究中的应用

典型变量分析（CVA）在农业研究中应用的条件已经具备，本文对CVA在草地生态系统研究中的应用作了例示与讨论，在控制光温的人工草地上，高羊茅草丛群体的3个特征－牧草产量，叶面积和分薛密度与钊割高度（3，6，10，14，18和24cm)的关系研究中采用CVA，结果表明，第一典型相关极显著，且第一典型变量能反映刈割高度问98％的变异，在用CVA，结果表明，第一典型相关极显著，且第一典型变量能反映刈割高度问98％的变异，在3－24xm范围内，刈割高度的作用是渐进式的，任两刈割高度间的群体之综合差异与其高谍差呈正比例，叶面积和分薛刈割高度间差异的解释能力基本持平，但作用方向相反，两都都大于牧草产量的作用，建议在应用CVA时，应该适当增加重复次数，考虑附加线性变量，剔除分布不对称的指标变量等等。

The Study on Effects of Cutting Heights on Swards Features-Application of Canonical Variate Analysis as a Tool for Grassland Ecosystem Research

Application of CVA in grassland ecosystem research was illustrated and discussed. CVA was used to study the relationship between 6 cutting heights of 3, 6, 10, 14, 18 and 24cm, and 3 population traits of herbage yield, leaf area and tiller density, on manned swards of tall fescue grown under constant light and temperature regimes. The first canonical variate, which is very significant, explained 98% of separation among the heights. Any two pasture populations varied in 3 traits in proportion to their difference in height. The rank in. contribution to separation among heights was that leaf area>tiller density>herbage yield, while tiller density behaves in opposite direction to both of leaf area and the yield. Those were recommended that increase in replications and additional nonlinear variables, and deletion of non-symmetrical distributed variables when CVA is to be used.