摘要：
对秃杉分布区13个县(市)的纯林、混交林、散生木、孤立木进行调查,选出236株作为初选优树,并观测这些候选优树的立地因子、生长性状等指标.经相关性分析,选择单株材积作为优树复选的主要指标.采用数量化回归的方法,建立经度、纬度、海拔、坡度、黑土层厚度、树龄6个数量因子和坡向、坡位、坡形、基岩、土壤类型、起源、立木类型7个定性因子与秃杉单株材积的回归方程,其复相关系数为0.799.秃杉单株材积实测值与理论值的差值(Ii)代表基因型值,其频率分布成正态分布.以Ii与差值平均值I±标准差δ相比较作为划分优树等级的依据,秃杉以候选优树70％的入选率统计.差值Ii≥I+0.3δ,即Ii≥1.6446为Ⅰ级优树;I+0.3δ＞ Ii＞I-0.3δ,即1.6446＞Ii＞-1.6446为Ⅱ级优树;Ii≤I-0.3δ,即Ii≤-1.6446为Ⅲ级优树(一般林木),Ⅲ级优树淘汰不选.用此标准对秃杉236株野外初选优树进行复选,Ⅰ级优树54株,占候选优树的22.88％;Ⅱ级优树114株,占候选优树的48.31％;Ⅰ、Ⅱ级优树预估遗传增益达20.82％.%The site factors and growth traits of 236 candidate Taiwania flousiana trees of pure forest,mingled forest,scattered wood and isolated wood in 13 counties were measured.The correlation analysis of candidate trees showed that volume could be used as the main indexes for selection of superior trees of T.flousiana.The regression equation with individual volume was established by using 6 quantitative factors including longitude,latitude,altitude,slope,the depth of black soil layer,ages of trees and 7 qualitative factors including slope aspect,slope position,slope shape,bedrock,soil type,origin,tree type.Its multiple correlation coefficient was 0.799.There was a difference between theoretical and realistic volume,the frequency of difference was in normal distribution.The superior tree selection was based on the comparison of Ⅰi and the mean value of difference I ±-standard deviation δ.The superior tree selection standards with 70％ selective ratio were as follows:Class Ⅰ superior tree,Ii ≥ I + 0.3δ,that is Ii ≥ 1.644 6.Class Ⅱ superior tree,I + 0.3δ ＞ Ii ＞ I-0.3δ,that is 1.644 6 ＞ Ii ＞-1.644 6.Class Ⅲ superior tree (normal tree),Ii ≤ I-0.3δ,that is Ii ≤-1.644 6.54 strains of class Ⅰ superior trees and 114 strains of class Ⅱ superior trees were selected by the standards.They accounted for 22.88％ and 48.31％ in all candidate superior trees respectively.The genetic gain of the class Ⅰ and class Ⅱ superior trees reached 20.82％.
