Selection for lateral branch number in melon (Cucumis melo)

摘要

Manipulation of plant architecture in melon (Cucumis melo L.) may provide for increased fruit yield. Recurrent selection through pedigree breeding was practiced in a melon population derived from a cross between line USDA 846-1 (days to anthesis = 50;monoecious, 5 to 8 primary lateral branches; basal-concentrated fruiting habit) and 'Top Mark' (days to anthesis = 60; andromonoe-cious; 3 to 4 primary lateral branches; diffuse, distal fruiting habit). Thirteen low branching (LB; < 4 branches) and 13 high branching (HB; > 6 branches) plants were selected from 200 greenhouse-grown F, individuals. These selections were self-pollinated and then selected for two additional cycles using the same methodology. Parental lines, their F_1, F_2, and selected high and low F_3, F_4, and F_5 families (13 low and 13 high/cycle) and controls were evaluated for primary lateral branch number, and fruit weight and number in a replicated open-field trial at Hancock, Wise, during 2003. USDA 846-1 differed significantly (P_(0.05)) from 'Top Mark' for mean number of lateral branches (5.8 versus 4.4), fruit number per plant (3.2 versus 1.6), and fruit weight (kg) per plant (1.7 versus 0.9). Mean branch number for HB families (5.8) was significantly higher (P_(<0.05)) thanthe comparative means of USDA 846-1 (4.5 [4.4-see preceding statement]), commercial controls ('Hales Best Jumbo' = 4.7, 'Esteem' = 5.2, 'Sol Dorado' = 5.0), and the LB families (4.6). Major gains from selection for branching (low and high) were detectedin the first cycle of selection, i.e., F_2 to F_3, with a decrease and increase in branching in LB and HB populations over the three cycles of selection (P_(<0.05)). No consistent significant differences were observed over cycles of selection for the other traits examined. Differences were detected between HB and LB families for all traits (P_(<0.01)) and for cycles of selection (high versus low) for branch and fruit number (P_(<0.01)) and fruit number and fruit weight (P_(<0.05)). Positive and significant correlations were detected between the number of lateral branches and fruit number (r = 0.39, P <= 0.01) and weight (r = 0.23, P_(<0.05)). Fruit number and weight were significantly correlated (r = 0.54, P_(<0.01)). Selection was successful for developing HB and LB lines with differing fruit setting potential.