Path analysis is a statistical method for determining the magnitude and direction of multiple effects on a complex process. We used path analysis to determine the direct effects of nightshade density on yield components (number of green fruit per plant, rotted fruit per plant, total fruit per plant, and weight per fruit) of the processing tomato cv. Heinz 6004. In addition, the analysis indicated the direct and indirect effects of yield components on total yield per ha and marketable yield per ha. The greatest direct effects of eastern black nightshade and black nightshade were on green fruit per plant and total fruit per plant. Effects other than density (density-independent factors) were more important in determining the number of rotted fruit per plant and weight per fruit. Path analysis showed that the total number of fruit per plant was the most important yield component determining total yield and marketable yield per ha.
Milton E. McGiffen Jr. and Dan J. Pantone
Milton E. McGiffen Jr., Dan James Pantone, and John B. Masiunas
Path analysis is a statistical method for determining the magnitude and direction of multiple effects on a complex process. We used path analysis to assess 1) the impact of black nightshade(Solarium nigrum L.) or eastern black nightshade(Solarium ptycanthum Dun.) competition on the yield components of `Heinz 6004' processing tomato (Lycopersicon esculentum Mill.) and 2) the relationship between tomato yield components and total and marketable yield. Either black or eastern black nightshade was interplanted with tomatoes at population densities from 0 to 4.8/m2. Path analysis revealed that increasing weed population density led directly to fewer green and total fruit per plant, two components of marketable yield. However, the percentage of culls per plant and fruit weight were not affected by nightshade population density. Using correlation coefficients alone would have lead to the erroneous conclusion that the percentage of culls did not affect marketable yield; our path analysis demonstrated that decreasing the percentage of culls through breeding or cultural practices will strongly affect marketable yield. The total number of fruit was the most important yield component in determining total and marketable yields per plant. Breeding and management practices that maximize fruit set, increase maturity at harvest, and decrease the percentage of culls would be expected to increase marketable yield.
Mehdi Mohebodini, Naser Sabaghnia, and Mohsen Janmohammadi
its extracts for various medicinal uses and industrial uses for edible oil ( Rehman et al., 2012 ). The common path analysis approach might result in multicollinearity for variables, particularly when associations among some of the traits are high
Reza Amiri, Kourosh Vahdati, Somayeh Mohsenipoor, Mohammad Reza Mozaffari, and Charles Leslie
–and-effect relationships in the variables, because the association between two variables may depend on a third variable. The use of path analysis provides a plausible explanation of observed correlations by modeling the cause-and-effect relations between the variables
Peng Shi, Yong Wang, Dapeng Zhang, Yin Min Htwe, and Leonard Osayande Ihase
estimate the interrelation among the agronomic traits (Boo et al., 1990). Hierarchical cluster analysis shows the similarity of the samples ( Peeters and Martinelli, 1989 ). Path analysis serves as a great tool to evaluate the relationship among agronomic
Christopher S. Cramer and Todd C. Wehner
The relationships between fruit yield and yield components in several cucumber (Cucumis sativus L.) populations were investigated as well as how those relationships changed with selection for improved fruit yield. In addition, the correlations between fruit yield and yield components were partitioned into partial regression coefficients (path coefficients and indirect effects). Eight genetically distinct pickling and slicing cucumber populations, differing in fruit yield and quality, were previously subjected to modified half-sib family recurrent selection. Eight families from three selection cycles (early, intermediate, late) of each population were evaluated for yield components and fruit number per plant in four replications in each of two testing methods, seasons, and years. Since no statistical test for comparing the magnitudes of two correlations was available, a correlation (r) of 0.7 to 1.0 or –0.7 to –1.0 (r 2 ≥ 0.49) was considered strong, while a correlation of –0.69 to 0.69 was considered weak. The number of branches per plant had a direct positive effect on, and was correlated (r = 0.7) with the number of total fruit per plant over all populations, cycles, seasons, years, plant densities, and replications. The number of nodes per branch, the percentage of pistillate nodes, and the percentage of fruit set were less correlated (r < |0.7|) with total fruit number per plant (fruit yield) than the number of branches per plant. Weak correlations between yield components and fruit yield often resulted from weak correlations among yield components. The correlations among fruit number traits were generally strong and positive (r ≥ 0.7). Recurrent selection for improved fruit number per plant maintained weak path coefficients and correlations between yield components and total fruit number per plant. Selection also maintained weak correlations among yield components. However, the correlations and path coefficients of branch number per plant on the total fruit number became more positive (r = 0.67, 0.75, and 0.82 for early, intermediate, and late cycles, respectively) with selection. Future breeding should focus on selecting for the number of branches per plant to improve total fruit number per plant.
JoAnn Robbins and Patrick P. Moore
Fruit weight and morphological characteristics of `Meeker' red raspberry (Rubus idaeus L.) fruit, including drupelets (height, diameter, number), receptacle cavities (depth, diameter), and pits (individual weight) were measured five times in 1988. Fruit strength, as measured by compression, was recorded. The relationship of fro-it weight to fruit strength had linear and quadratic components. Fruit weight was correlated with fruit strength, drupelet height and number, receptacle cavity depth and diameter, and individual pit weight. Besides fruit weight, fruit strength was correlated with drupelet diameter and number, receptacle cavity depth, and individual pit weight. Drupelet number, receptacle cavity depth, and individual pit weight provided the largest component contribution to fruit strength, as determined by path analysis.
Yield components of 8- to 10-year-old trees were compared among `Kieffer' [Pyrus communis (L.) x Pyrus pyrifolia (Burro.)] and `Harrow Delight' and `Harvest Queen' [Pyrus communis (L.)]. `Kieffer' set more fruit than the other cultivars, even though flower density was similar. `Kieffer' also had similar size or larger fruit than `Harrow Delight' or `Harvest Queen'. Path analysis showed that the direct and indirect effect of fruit number on yield was important for all cultivars. Flower density only had a small direct effect on yield and this was at times negative. Fruit size had a small effect on yield when compared to fruit number.
Job Teixeira de Oliveira, Rubens Alves de Oliveira, Fernando França da Cunha, Isabela da Silva Ribeiro, Lucas Allan Almeida Oliveira, and Paulo Eduardo Teodoro
technique to establish how to increase yield in most crops. However, the correlation between two variables can be influenced by a third variable or a group of variables. Path analysis is the most appropriate technique to remove the effect of these other
Kenneth R. Tourjee, Diane M. Barrett, Marisa V. Romero, and Thomas M. Gradziel
The variability in fresh and processed fruit flesh color of six clingstone processing peach [Prunus persica (L.) Batsch] genotypes was measured using CIELAB color variables. The genotypes were selected based on the relative fruit concentrations of β-carotene and β-cryptoxanthin. Significant (p < 0.0001) differences were found among the genotypes for the L*, a*, and b* color variables of fresh and processed fruit. Mean color change during processing, as measured by ΔELAB, was greatest for `Ross' and least for `Hesse'. A plot of the first two principal components (PCs) obtained from PC analysis of the L*, a*, and b* variables for fresh and processed fruit revealed three clusters of genotypes that match groupings based on the relative concentrations in fresh fruit of carotenoid pigments. Path analysis showed that variation in β-cryptoxanthin concentration was more precisely determined from color data than β-carotene concentration. Chemical names used: β-β-carotene (β-carotene), (3R)-β-β-caroten-3-ol (β-cryptoxanthin).