|
|
Sequential Path Analysis for Determining the Interrelationships between Yield and Its Components in Peanut
|
|
|
Abstract: The current work was carried out at the Agriculture Research Station of East Al-Eweinat, New Valley governorate to evaluate the yield potential of 16 peanut genotypes during 2016 and 2017 growing seasons. The used experimental design was a randomized complete block design with three replicates. Correlation coefficients were computed between pod yields and its related attributes as well as normal and sequential path analysis models were automated to obtain information on the direct and indirect effects of important traits affecting pod yield for using them as selection criteria in future peanut breeding programs. Results showed that genotypes 7, 11 and 16 produced the heaviest pod yield while genotypes 13 and 15 recorded the lowest pod yield. It is obvious that the high yielder genotype 7 was characterized by high pod weight per plant while elite genotype 11 occupied the first order for most yield components being numbers of branches, pods and seeds per plant and pod and seed weight per plant. Also, the superior genotype 16 had the highest values of number of pods per plant and pod weight per plant. Results appeared that pod yield was positively and highly significant associated with number of pods per plant, pod weight per plant, number of seeds per plant, and seed weight per plant. The associations among these yield attributes were positive and significant; therefore, the selection of any one of them would contain the others but the number of pods per plant was visually easier to select in the field. Concerning the normal path analysis model, several undesirable symptoms were obtained indicating the presence of multicollinearity problem. The highest variance inflation factor (VIF) values (above 10) for some basic yield characters are clear indicators that the assumption of orthogonality (independence) among the yield related characters is violated. The path coefficients were obviously fluctuated recording very small values (close to be zero) and other inflated values exceeded 1. Subsequently, the poor estimators of normal path analysis model, as a result of multicollinearity, enough to reject the normal form of path analysis. Statistically, more precise results were obtained using the sequential path analysis model. Results revealed that the pod yield depended primarily upon pod weight per plant and number of pods per plant as first-order variables accounted for nearly 98% of the variation in pod yield. The maximum positive direct effects were obtained by pod weight per plant (0.88) followed by number of pods per plant (0.13) indicting that the indirect selection for pod yield through these traits would be effective for peanut improvement. The second-order path analysis showed that seed weight per plant had the considerable positive direct and indirect effects toward each of number of pods per plant and pod weight per plant. In fact, the sequential path analysis gave a somewhat different picture from what the normal model path analysis did.
|
| Publication year |
2020
|
| Organization Name |
|
| serial title |
المجلة المصرية للزراعة
|
| Author(s) from ARC |
|
| External authors (outside ARC) |
|
| Publication Type |
Journal
|