STUDY THE RELATIONSHIPS BETWEEN YIELD AND YIELD COMPONENTS OF POTATO VARIETIES USING CORRELATION ANALYSIS AND REGRESSION ANALYSIS AND CAUSALITY
Ghasem Rahimi Darabad
|Related article at Pubmed, Scholar Google|
In order to evaluate the correlation between attributes and their analysis to causal relationships in the potato crop, an experiment was performed in Agriculture and Natural Resources Research Station of Ardabil on six genotypes in a split plot test based on completely randomized blocks with three replications. In the statistical analysis of this study, 21 traits were studied and analyzed. To perform this experiment, water stress was applied by increasing the irrigation interval during the growth period of the plant. The irrigation period factor with three levels of irrigation without stress (irrigation every 6 days), the average stress (irrigation every 12 days) and severe stress (irrigation every 18 days) in the main plots and the potato varieties factor in 6 levels in the subplots were evaluated. ANOVA results showed that regarding different levels of irrigation and the studied varieties, apart from the traits of the number of days until tuber developing and the percentage of green plants, there were significant differences in all traits. The Sornad and Agria cultivars had the highest average in tuber yield, respectively, with 28.78 tons per ha and 28.28 tons per ha, and the Draga cultivar produced the lowest average tuber yield with 24.41 tons per acre. Potato cultivars with different irrigation periods showed mutual effects in terms of plant height, number of main stem, number of tubers per plant, tuber weight per plant, tuber yield, edible tubers percentage. This showed that the difference in cultivars has not been the same in different irrigation periods. The tuber yield had a significant positive correlation with leaf area index, edible tubers percentage, plant cover percentage, plant height, main stem diameter, weight of tubers per plant, number of tubers per plant and the percentage of 45-55 mm tubers, while it had a significant negative correlation with the percentage of 28-35 mm tubers, the percentage of 35-45 mm tubers, storage loss percentage and the percentage of non-seeding tubers. In fitting the best multivariate regression model by descending method for all studied traits, by inclusion the traits of green plant percentage, tubers weight per plant and the percentage of tubers between 35-45 mm, the best model was fitted. These were the most important traits as the independent variable to develop the tuber yield as the dependent variable.