Assessment of Long-term Peach Evaluation Data Reveals Fruit Trait Distribution and Selection Tendencies

Conventional plant breeding has a long history of success. Although it is often a daunting job, it is still the most common approach used in today’s variety improvement programs. Efficient production of hybrids with potential for release as new varieties with desired traits relies on optimal selection of parents and long comprehensive evaluation on thousands of hybrids. Decades of efforts have led to the release of over a dozen peach (Prunus persica) varieties in the USDA-ARS stone fruit breeding program at Byron, GA, as well as producing many superior selections and much valuable evaluation data. Assessment of the data could facilitate further optimization of parents and cross combinations, and ensure continuous improvement of breeding efficiency. Fruit evaluation data collected over 30 years from a wide range of peach varieties and seedling selections was analyzed. A normal distribution was observed for fruit size (median diameter = 2.4 inches, average = 2.405), and ripening date (median = June 30). Other traits, including fruit set, external blush, attractiveness, firmness, freeness, shape, pubescence, and overall eating quality, showed an uneven distribution toward the desired directions, which was expected since data was rarely taken on discarded hybrids. The median value of fruit set was 7 and the average 6.590 on a scale of 0-9, suggesting this trait had been maintained in a majority of selections. Blush and attractiveness showed a similar distribution where 7 was the median for both and 6.806 and 6.945 were the averages respectively on a 0-9 scale, indicating the selection tendencies were toward a higher blush coverage but with some classic yellow background to indicate the ripening stage. Correlation analysis showed only blush (relatively objective) and attractiveness (subjective) were somewhat positively correlated (R = 0.413), consistent with the tendencies. But set and size were not negatively correlated (R = 0.004), inconsistent with a usual expectation and some reports. However set-size data might be confounded by mixing seedling selections with named varieties, as well as by the impact of ripening season. How to further use the evaluation data and the assessment summary to advance the future breeding objectives will be discussed.