Ten Cycles of Recurrent Selection for Fruit Yield, Earliness, and Quality Three Slicing Cucumber Populations

Fruit yield, earliness, and quality have low to moderate heritability, but are traits of major importance in cucumber (Cucumis sativus L.). The objective of this study was to determine the changes made in those traits using recurrent selection in three slicing cucumber populations (NCMBS, NCES1, and NCBA1). During population improvement, one or two replications of 200 to 335 half-sib families were evaluated in the spring season for five traits: total, early, and marketable fruit per plot, fruit shape rating, and a simple weighted index (SWI = 0.2(total yield)/2 + 0.3(early yield) + 0.2(% marketable)/10 + 0.3(fruit shape). Families from each population were intercrossed in an isolation block during the summer season using remnant seeds of the best 10% selected using the index. Response was evaluated using a splitplot treatment arrangement in a randomized complete block design with 32 replications in each of two seasons (spring and summer). Whole plots were the three populations, and subplots were the 11 cycles (cycles 0 to 9 plus checks). We measured improvement in performance of the populations in a selected (spring) and unselected environment (summer). Significant gains were made for all traits in all populations over the 9 to 10 cycles of recurrent selection. Greatest progress was made for the NCMBS population, with an average of 37% gain from cycle 0 to 9 over all five traits. The trait where most progress was made was early yield, with an average of 63% gain from cycle 0 to 9 over the three populations. Cucumber is the second most important horticultural crop in North Carolina (including both pickling and slicing types), and North Carolina is the fourth largest producer of slicing cucumbers in the United States (U.S. Dept. of Agriculture, 1990). Breeding efforts in the United States have concentrated on the incorporation of qualitatively-inherited traits (gynoecy, disease resistance, fruit color) into elite inbreds using backcross or pedigree selection (Wehner, 1988b). The inbreds are then made available directly to growers or used in single-cross hybrids. However, many traits of interest to growers (fruit yield, earliness, and quality) are quantitatively inherited, and have low heritability. Recurrent selection for population improvement is an effective breeding method for improving quantitative traits with low heritability in cucumber (Wehner, 1989). Recurrent selection involves systematic testing and selection of desirable individuals from a population followed by recombination of the selected individuals to form a new population (Fehr, 1991). The new population should be superior (although performance might also get worse) to the original population, both in average performance and in performance of the best individuals. The accumulation of small gains from each cycle results in significant long-term improvement of the population. In cucumber, early testing for yield traits is effective (Rubino and Wehner, 1986a), and inbreeding depression is minimal (Rubino and Wehner, 1986b). Thus, the development of advanced populations with improved fruit yield, earliness and quality should result in inbred lines and hybrids with improved performance. or publication 28 June 1995. Accepted for publication 4 Dec. 1995. The ported in this publication was funded in part by the North Carolina al Research Service. The use of trade names in this publication does not orsement by the North Carolina Agricultural Research Service, nor f similar ones not mentioned. We gratefully acknowledge the technical of Rufus R. Horton, Jr., and the advice on recurrent selection methods . Moll. The cost of publishing this paper was defrayed in part by the f page charges. Under postal regulations, this paper therefore must be rked advertisement solely to indicate this fact. research assistant. Recurrent selection has been used in cucumber to increase low temperature germination ability (Nienhuis et al., 1983; Staub et al., 1988), herbicide resistance (Staub et al., 1991), and disease resistance (Sloane et al., 1985). Recurrent selection methods such as S1 line, half-sib family and full-sib family selection have been effective for yield improvement in cucumber (Lertrat and Lower, 1983, 1984; Nienhuis, 1982; Nienhuis and Lower, 1988; Wehner, 1989). On the other hand, convergent-divergent selection was not effective in the Gynoecious Synthetic population for improving yield over several environments in midwestern and southeastern United States (Wehner et al., 1989). The objective of this study was to determine whether progress was made for fruit yield, earliness and quality using recurrent selection in three slicing cucumber populations. Materials and Methods Research on population improvement was done in three general stages: population formation (1981–84), recurrent selection (nine or more cycles), and evaluation of response (a replicated study in two environments in 1993). All work was done at the Horticultural Crops Research Station, Clinton, N.C. Population formation. Three slicing cucumber populations were developed at North Carolina State Univ. The North Carolina medium base slicer (NCMBS) population consisted of 152 cultigens (cultivars, breeding lines and plant introduction accessions) intercrossed in 1981 and 1982 (Strefeler and Wehner, 1986; Wehner, 1996a). The North Carolina elite slicer 1 (NCES1) population consisted of eight cultigens intercrossed in 1981 and 1982 (Strefeler and Wehner, 1986; Wehner, 1996a). The North Carolina Beit Alpha 1 (NCBA1) population consisted of eight cultigens intercrossed in 1982–84 (Wehner, 1996b). The populations were developed to improve fruit yield, earliness and shape through a program of modified half-sib family recurrent selection. Recurrent selection. Populations were improved by testing in the spring season followed by intercrossing the best families in isolation blocks in summer for 9 (NCBA1) to 10 (NCMBS, J. AMER. SOC. HORT. SCI. 121(3):362–366. 1996. NCES1) cycles. Half-sib families from each population were planted in 1.5-m-long plots (changed to 1.2 m after selection cycle 4 was completed) on raised, shaped beds 1.5 m apart during the spring season from 1983–92. That plot size was found to be optimum for yield measurement (Swallow and Wehner, 1986). Plot end borders were found to be unnecessary for efficient trialing (Wehner, 1988a), as were multiple-row plots (Wehner and Miller, 1990), so neither was used. Recommended cultural practices (summarized recently by Schultheis, 1990) were used throughout the experiments. ‘Poinsett 76’ was planted every eighth row, and in field border rows and end tiers as a pollenizer. Irrigation was applied when needed for a total of 25 to 40 mm per week. Fertilizer was incorporated at a rate of 90N–39P–74 K kg/ha before planting, with an additional 34 kg N/ ha applied at the vine-tip-over (four to six true leaf) stage. Herbicide [Curbit, ethalfluralin, N-ethyl-N-(2-methyl-2-propenyl)-2,6dinitro-4-(trifluoromethyl)benzenamine] and insecticide (Sevin, carbaryl, 1-naphthyl N-methylcarbamate) were applied at recommended rates (Schultheis, 1990). The 20-year normals for mean daily air temperature during spring were 23 to 27C day, and 9 to 14C night. The 20-year normals for mean daily air temperature during the summer season were 30 to 37C day, and 18 to 20C night (National Climatic Data Center, Asheville, N.C.). The experiment was a randomized complete-block design with one replication of 200 half-sib families during the first 6 years of selection, increased to 390 families for 1 year, then increased to two replications the following year, and finally changed to 335 families in 22 sets of 16 each in the last 2 years. In the first 7 years, 24 families were selected for intercrossing; in the last 3 years, 40 families were selected for intercrossing. Changes were made in the test method to reflect current information on testing efficiency (Wehner, 1987; Wehner and Miller, 1984). The selection plots were planted 26 Apr. to 18 May and harvested 12 June to 15 July depending on year. J. AMER. SOC. HORT. SCI. 121(3):362–366. 1996. Table 1. Mean squares from analysis of variance for simple weig fruits per plot, percentage marketable yield (marketable), populations (NCMBS, NCES1, and NCBA1) tested in two recurrent selection at Clinton, N.C.