Seed from the parental, F1, and segregating (F2 and F3) soybean populations were evaluated for quality by visual ratings for wrinkled seedcoats, shriveled cotyledons, green cotyledon color, an arithmetic average of the three previous ratings, and an overall visual rating. These ratings were on a 1 to 5 scale, with 5 being the most extreme expression of the character.The variance among single plants in the commercial varieties for the visual ratings was generally high in relation to the variance in the Japanese strains and in the F1 populations. However, the variance in the F2 populations was generally large in comparison to the arithmetic mean of the variances within the respective parental and F1 populations resulting in broad sense heritability estimates ranging from 41 to 63% for wrinkled seedcoats, 14 to 57% for shriveled cotyledons, 52 to 82% for green cotyledons, 41 to 71% for the average visual rating, and 22 to 64% for the overall visual rating. Narrow sense heritability estimates for the visual ratings ranged from 6 to 30% for wrinkled seedcoats, 10 to 38% for shriveled cotyledons, O to 29% for green cotyledons, 1 to 28% for the average visual rating, and 7 to 28% for the overall visual rating.Low visual ratings for wrinkled seedcoats, shriveled cotyledons, green cotyledons, low average visual ratings, and low overall visual ratings were generally associated with high percentages of normal seedlings in field emergence tests. This was especially true in the F3 and ‘Chippewa’ parental populations in which more variation existed in each of the characters being tested.The average visual rating was somewhat more consistent in predicting relative field emergence than any of the ratings of separate seed characters. However, the overall visual rating appeared to be about equal in this respect. Considering both precision of measurement and time utilization, an overall visual rating followed by field emergence testing may be the most efficient means of evaluating segregating populations of soybeans for seed quality.
Read full abstract