| Watermelon is a diverse crop in terms of qualitative genetic traits, with many different fruit types among the cultivars. We were interested in determining the optimum method for improving yield, and measuring the heritability and rate of natural outcrossing within elite populations. The rate of natural outcrossing is dependent upon the distance between plants, as well as other factors such as location and cultivar. The objective of this study was to (1) estimate narrow-sense heritability for yield using parent-offspring regression in two watermelon populations (NCHYWI and NCHYW2); (2) determine the inheritance of fruit yield, rind pattern, and fruit shape in six generations (Pa S1, PaS2, F1, F2, BC1Pa and BC1Pb) of three families; and (3) determine the rate of natural outcrossing effected by in-row spacing and cultivars. Field trials were conducted at two locations in North Carolina (Clinton and Kinston) to determine the narrow-sense heritability for yield in NCHYW1 and NCHYW2 watermelon populations using parent-offspring regression. Low estimates of narrow-sense heritability were recorded for total fruit weight (0.04-0.12), marketable fruit weight (0.06-0.15), total fruit number (0.04-0.16), fruit size (0.18-0.19), and percent culls (0.02-0.09) for NCHYW1 and NCHYW2 populations, respectively. Only low gain in yield can be made due to single-plant selection, based on the populations used. Strong positive genotypic correlations were observed between total fruit weight and marketable fruit weight; total fruit weight and marketable fruit weight with fruit size, and negative correlation was recorded between total fruit number and fruit size, and total fruit weight and marketable fruit weight with percent culls. In the second experiment consisting of three families and six generations, a low to intermediate level of heritability was reported for total fruit weight, total fruit number, and fruit size. Recurrent selection is recommended to improve populations for these traits to accumulate favorable genes.;The family 'Mountain Hoosier' x 'Calsweet' did not fit the expected segregation ratios for the single dominant gene for solid dark green rind pattern (G) (solid dark green vs. wide stripe), or for the single incompletely dominant gene for elongate fruit shape ( O) (elongate vs. round) in the F2 and backcross, different from previous reports. Deviation from expected ratios was also observed for the single dominant gene controlling solid dark green rind (G) and light green rind (g) in the families 'Mountain Hoosier' x 'Minilee' and 'Early Arizona' x 'Minilee'. However, segregation ratios showed that light green rind was controlled by two recessive genes, g-1 and g-2, by duplicate dominant gene action when crossed to a line having solid dark green rind.;The rate of natural outcrossing was determined by using split plot in randomized complete block design in two locations: Kinston and Clinton; eight in-row spacing as whole plots: 0.6, 1.2, 1.8, 2.4, 3.0, 3.7, 4.3, and 4.9 m; and two cultivars as sub plots: 'Allsweet' and 'Mickylee'. 'Moon and Stars' was used as the pollen donor of marker gene to track the rate of natural outcrossing. Location and cultivar, and all interactions were not significant. Closer in-row spacing (0.6 m and 1.2 m) had a significantly higher rate of natural outcrossing (11.0% and 16.9%, respectively) than wide in-row spacing (>4.3 m), which had a low rate of outcrossing (1.8%). In conclusion, watermelon appears to act more like a self-pollinated crops when spaced >5 m apart. The rate of natural outcrossing should also be taken in to account while estimating heritability and genetic variance in watermelon populations. |
