Four agronomically and genetically diverse winter wheat parents were utilized as the experimental organisms. Atlas 66 and NB 68513 were selected as cultivars with a high and stable protein content when grown under different environmental condidtions. They are intermediate for grain yield when grown in the Pacific Northwest. Yamhill and Hyslop represented low protein, high yielding cultivars adapted to the Pacific Northwest. Data were obtained from crosses between the two high protein cultivars and the two low protein cultivars based on the performance of the parents and the F1 and F2 generations. These experimental populations were grown in 1971 at the Pendleton Experiment Station and the Central Oregon Experimental site at Madras, Oregon. Measurements were made on an individual plant basis for protein content, grain yield, 50 kernel weight, kernels per spike, tillers per plant and plant height. Differences among and within crosses were determined by the analysis of variance. Information concerning the nature of inheritance was obtained by comparing the F1 and F2 means in relation to performance of the parents; the frequency distribution of the generations for protein content; and by determining broad and narrow sense heritability estimates for the six characters studied. The existence of possible phenotypic associations among the six characters studied was determined by using correlation coefficients. In order to evaluate the possible direct and indirect effects of grain yield and the components of yield on protein content, path coefficient analyses were employed. Significant differences were observed among and within crosses at both the Pendleton and Madras sites for most characters measured. The F1 and F2 mean values were found to be near the mid-parent of the two parents in all four crosses for plant height, 50 kernel weight and kernels per spike. There were several exceptions depending on the particular cross and specific character. Protein content mean values were also intermediate between the two parents for the F1 and F2 generations. In crosses involving Hyslop, the mean values tended to be near the highest parent. Little or no transgressive segregation was noted in the F2 generation. Evidence of non additive gene action was noted both for grain yield and tiller number in the F1 and F2 generations with the mean values exceeding the highest parent in all crosses for grain yield at the Pendleton site. Tillers per plant at Pendleton and both tillers per plant and grain yield at Madras also showed some degree of hybrid vigor, but the magnitude depended on the particular cross. The high broad and narrow sense heritability estimates obtained both at Pendleton and Madras for all traits suggested that there was a large amount of genetic variation present for the characters studied. The narrow sense estimates further suggested that a high percentage of the total genetic variation was due to genes which function in an additive manner. Significant negative correlations were noted between protein content and grain yield including some of the components of yield. In evaluating the direct and indirect effects with path coefficient analysis, these negative associations resulted from the large negative indirect effects of 50 kernel weight and kernels per spike on protein content via grain yield at the Madras site. At the Pendleton site, where moisture became a limiting factor, the negative association resulted largely as the indirect effect of 50 kernel weight on protein content through grain yield. The large environmental influence on protein content was particularly striking at the Pendleton site. With the spring application of nitrogen, a delay in maturity for Hyslop and Yamhill was noted and with the subsequent loss of moisture, shriveled grain resulted and hence a higher protein content with lower grain yield. This resulted in the grain protein of Hyslop and Yamhill being higher than that of Atlas 66 and NB 68513. The results of this study suggest that it may be necessary to compromise in attempting to develop high protein lines with maximum yield. However, it should be possible to increase the protein content two to three percent and still maintain the yielding ability of Hyslop and Yamhill.

Parental and segregating populations derived from four winter x spring wheat crosses were investigated to obtain information concerning the inheritance and association of earliness, grain yield and yieldrelated traits. Feasibility of selecting in early generations for these characteristics was also evaluated. Four winter wheat cultivars (Hyslop, Yamhill, Bezostaia 1, and Sprague) and one spring wheat cultivar (Inia 66) were chosen on the basis of their relative maturity and contrasting agronomic characteristics. Parents, F1 s, F2' s, and reciprocal backcrosses to both parents were planted in the fall in a space-planted randomized complete block design. The two environmentally diverse locations selected were the Hyslop Agronomy Farm, Corvallis, Oregon (1000 mm of rainfall) and Sherman Experimental Station, Moro, Oregon (250 mm of rainfall). The effectiveness of early generation selection for the measured characteristics was evaluated by growing F3 lines identified as the earliest 1% and the highest yielding 1% of F2 individuals in each cross. These were grown along with the parents, F1s, BC1 s, BC2 s and F2' s under space-planted conditions at Hyslop Agronomy Farm. A study with the same populations was conducted by vernalizing and planting in the spring to gain further information on earliness. Analyses of variance were conducted for all characteristics measured. Frequency distributions for days to heading of F1, F2, backcross generations and parents were examined. From the data collected, estimates of F 1 -midparent deviations, degree of dominance, heritability in the narrow sense and genetic advance under selection were determined for each cross. The data were further analyzed by parent-progeny regression, correlation and path-coefficient analyses, polynomial and multiple regressions. Partially dominant major genes, varying in number between one to five depending on the particular cross, appeared to influence heading date. Modifying factors also seemed to affect the date of heading. The gene action involved in the inheritance of earliness was primarily additive indicating that selection for earliness would be effective as early as the F2 generation under both high and low rainfall conditions. Estimates of additive and nonadditive gene action suggested both were equally important in determining the yield components. Higher heritability estimates for the components of yield indicated that there was more genetic variability associated with the yield components than yield per se. Occurrence of additive genetic variation by location interaction implied that selection should be practiced simultaneously under different environments if wide adaptability of potential lines is desired. Since pronounced additive effect by year interactions occurred for the yield components, delayed selection for these traits may not be productive. Positive correlations were obtained between yield and the number of days to heading when all generations were combined. However, in the F2 generations, it appeared possible to select for the desired earliness with high yields as indicated by the low association between these two traits. The path-coefficient analyses suggested that tiller number had the highest direct effect on grain yield. However, because of a negative association between tiller number and kernel weight, selection pressures would have to be balanced between these two components. In most cases, linear relationships existed between grain yield and seven measured traits, respectively. The result of regression analyses also showed that grain yield may be described best as a linear function of its components.

The nature of inheritance and possible associations for traits influencing earliness and grain yield were investigated using a four parent diallel of winter and spring wheat cultivars. More genetic variability was observed for the traits measured in segregating populations resulting from crosses between winter and spring type wheats in contrast to spring x spring or winter x winter crosses. The one exception was plant height where more genetic variability resulted from spring x spring crosses. Narrow sense heritability estimates were high for time and duration of heading, anthesis, grain filling and physiological maturity and for plant height. Smaller values were noted for rate of grain filling, kernel number, harvest index, tiller number, kernel weight, whole plant dry weight and grain yield. Estimates of the coefficient of heritability and the parent-offspring correlation coefficient were similar in magnitude except for the traits grain yield, tiller number, kernel weight and whole plant dry weight where large variations due to the environment were encountered. Using the Jinks-Hayman model, no maternal effects were noted nor were any nonallelic interactions observed for total duration of grain filling and lag period. The actual grain filling period was influenced to some degree by such interactions. The spring cultivars also appeared to have more dominant genes for longer total duration of grain filling and lag period. In contrast the winter parents had more dominant genes for the longer actual grain filling period. Estimates of general and specific combining ability provided similar evidence in terms of the nature of gene action. Both additive and nonadditive gene action was present for all traits, the relative magnitudes depending on the specific trait. Based on individual combining ability effects, the winter x spring cross Yamhill x Siete Cerros would appear to provide the highest proportion of desired segregates when combining earliness and acceptable grain yield. From the direct and indirect associations of grain yield, it would appear that a shorter duration of grain filling along with a shorter lag period from heading to anthesis are important for higher rates of grain filling if negative associations between earliness and grain yield are to be avoided.

The purpose of this study was to determine the nature and amount of genetic variation and possible associations between winterhardiness and earliness in winter x spring wheat crosses. Four winter wheat cultivars selected for differences in earliness and winterhardiness were crossed with a nonhardy, day length insensitive spring wheat cultivar. The following year, experiments containing parents, F1, BC, and F2 populations were planted at two environmentally diverse sites located at the Sherman Branch Experiment Station, Moro, Oregon (250 mm of moisture) and the Hyslop Agronomy Farm, Corvallis, Oregon (1000 mm of moisture). The amount and nature of genetic variation involved were determined by obtaining broad and narrow sense heritability estimates, evaluating the degree of dominance and estimating the number of genes influencing both earliness and winterhardiness. Also frequency distributions were developed for each of the populations. Both broad and narrow sense heritability estimates for earliness were higher than those observed for winterhardiness. Both winterhardiness and earliness appeared to be conditioned by both additive and nonadditive gene action. Degree of dominance estimates for the four wheat crosses grown at two locations differed for each cross and location. Earliness was influenced by one to six genes while winterhardiness appeared to be controlled by two genes. The estimation of genetic advance indicated that the crosses with high narrow sense heritability estimates and high phenotypic variance in F2 generation would result in greater gains under selection for both traits. Based on the results of this study, it seems that Moro is a proper site to select for winterhardiness and Corvallis for earliness. However, it might be better to select for both traits at the same time at another site such as Pendleton, Oregon, where a realistic selection pressure can be applied for winter survival and drought would not influence the selection procedure. Such a site could also provide an opportunity to evaluate earliness at the same time. Correlation coefficient estimates showed the presence of a positive association between earliness and winterhardiness. The possibility of using leaf damage readings to measure the winterhardiness levels in wheat populations also appears promising.