MULTIPLE REGRESSION AND PATH ANALYSIS BETWEEN AGRONOMIC CHARACTERISTICS AND YIELD IN LATERALLY FRUITFUL INDIVIDUAL WALNUT

A study was conducted to compare the relations between tree characteristics such as tree height (X₁), tree stem diameter (X₂), vertical projection area of a tree crown (X₃), branching ability (X₄), blooming shoot percentage (X₅), the percentage of fruiting shoot (X₆), lateral bearing percentage (X₇), the average nuts number per fruiting shoot (X₈), the average length of fruiting shoot (X₉), the average diameter of fruiting shoot (X₁₀), the average mixed buds per fruiting shoot (X₁₁), average nut weight (X₁₂) and yield (Y) of 27 selected laterally fruitful individual walnut trees.
The correlation coefficients amongst the yield and its influencing factors were obtained according to the original datum.
There was very positive significant correlation between Y and X₃, positive significance between Y and X₆, Y and X₈. The optimal multiple linear regression equation was obtained after stepwise regression analysis adopting software of DPS: Y=-4.4299+0.1718X₃+0.0220X₆+1.5689X₈+ 0.1513X₁₂. The four factors which are the average nuts number per fruiting shoot, vertical projection area of a crown, the percentage of fruiting shoot and average nut weight are the main characteristics which affect a single tree yield dramatically.
Contribution for yield of the main four factors was illustrated by path analysis.
The main yield influencing factors are in order X₈>X₃>X₆>X₁₂ and the direct path coefficient is 0.60, 0.55, 0.47, 0.39 respectively.

VII International Walnut Symposium

B.J. Zhao, Y.H. Gong, F. Liu

walnut, agronomic characters, correlation analysis, multiple regression, path analysis

https://doi.org/10.17660/ActaHortic.2014.1050.31