Fruit Yield and Yield Components of Cucumber Populations Grown at Low Plant Density

Christopher S. Cramer

Department of Agronomy and Horticulture, MSC 3Q, Box 30003, New Mexico State University, Las Cruces, NM 88003-0003

Todd C. Wehner

Department of Horticultural Science, Box 7609, North Carolina State University,
Raleigh, NC 27695-7609

Additional index words. correlation, Cucumis sativus, Cucurbitaceae, indirect selection, path analysis, path coefficients

Abstract. Direct selection for improved yield in cucumber (Cucumis sativus L.) has been slower than expected, so selection for traits that are highly correlated with yield may be advantageous. Studies were conducted to determine the correlation between yield components and total, early, and marketable yield of eight cucumber populations as well as the change in correlation with selection. Four pickling and four slicing cucumber populations differing in their genetic diversity (wide, medium, elite, special) were developed using recurrent selection for improved fruit yield and quality. Four families were chosen at random from each of three selection cycles (early, intermediate, late) from each population, and families were self-pollinated. Three plants from each S1 family were evaluated in the spring and summer seasons of 1995 and 1996. At once-over harvest, plots were evaluated for number of branches, leaves, pistillate flowers, total fruit, early fruit, and cull fruit. Most yield components were weakly correlated with fruit yield in each population. For several populations, selection for an increased number of branches per plant would increase the number of fruit produced per plant. In most instances, selection had no effect on yield component means and correlations with yield. When selection did alter the strength of correlations, a majority of the correlations weakened with selection. Yield component heritability and better selection methods need to be developed before indirect selection for yield would be useful for cucumber.

We gratefully acknowledge the technical assistance of Tammy L. Ellington, Rufus R. Horton, Jr., Jinsheng Liu, Nischit V. Shetty, Joel L. Shuman, and S. Alan Walters.

 

Cucumbers (Cucumis sativus L.) are the second most planted vegetable crop in North Carolina, with a production area comprising 10,000 hectares (U.S. Department of Agriculture, 1998). Nationally, North Carolina is the third leading state in the production of pickling cucumbers and the fifth leading state in the production of slicing cucumbers (U.S. Department of Agriculture, 1998). Increased yield has been a breeding objective of many cucumber breeding programs. However, yield in cucumber is quantitatively inherited and has low heritability, making genetic gain difficult. Through the accumulation of small gains, recurrent selection has been successful in improving cucumber fruit yield and quality (Lertrat and Lower, 1983, 1984; Nien

huis and Lower, 1988; Wehner, 1989; Wehner and Cramer, 1996a, 1996b).

An alternative to direct selection for yield is to select for traits that are highly correlated with yield, but have higher heritability. For example, if fruit yield had a heritability of 0.25, and the indirectly selected trait had a correlation with yield of 0.70, then the heritability of the indirectly selected trait would have to be at least 0.36 for indirect selection to be effective. In addition, the selection methods used in indirect selection must be less expensive than direct selection. Indirect traits, often referred to as yield components, have been used widely to study yield in horticultural crops such as blueberry (Vaccinium corymbosum L.) (Siefker and Hancock, 1986), cucumber (AbuSaleha and Dutta, 1988; Solanki and Shah, 1989; Prasad and Singh, 1994a, 1994b; Yin and Cui, 1994; Zhang and Cui, 1994), strawberry (Fragaria ¥ananassa

 

Cucurbitaceae '98


Duchesne) (Hancock et al. 1984), and tomato (Lycopersicon esculentum Mill.) (McGiffen et al., 1994).

Some examples of yield components include harvests per plant, plants per hectare, branches per plant, nodes per branch, fruit per node, percentage pistillate flowers, percentage early fruit, and percentage marketable fruit (or culls). In tomato, McGiffen et al. (1994) concluded that factors such as increased number of flowers, reduced flower abortion, and greater fruit load could increase marketable yield by enhancing total fruit set, increasing earliness (decreasing the percentage of green fruit), and decreasing the number of culls.

Recently, Cramer and Wehner (1998a, 1998b) identified yield components that were correlated with fruit yield in pickling and slicing cucumber populations. The potential exists to use yield components to improve yield in those populations. The objectives of this study were to determine the correlation between pairs of yield components, and between yield and its components for eight cucumber populations tested at a low plant density; and to measure the change in the correlation with selection in each population.

Materials and methods

Plant materials. The four pickling cucumber populations used were the North Carolina wide base pickle (NCWBP), medium base pickle (NCMBP), elite pickle 1 (NCEP1), and hardwickii 1 (NCH1) (Wehner, 1997, 1998b). The four slicing cucumber populations used were the North Carolina wide base slicer (NCWBS), medium base slicer (NCMBS), elite slicer 1 (NCES1), and Beit Alpha 1 (NCBA1) (Wehner, 1998a, 1998b). The populations differed in their genetic diversity and mean yield performance, and were developed using modified half-sib recurrent selection to improve fruit yield, earliness, and shape of the population (Wehner and Cramer, 1996a, 1996b). Half-sib families were selected in the spring season based on a simple weighted index (SWI), that was weighted 70% toward fruit yield and 30% toward quality traits (Wehner and Cramer, 1996a, 1996b). Three cycles were chosen from each population (0, 3, 5 for NCWBP; 0, 5, 10 for NCMBP and NCH1; 0, 5, 9 for NCEP1; 0, 3, 6 for NCWBS; 2, 6, 10 for

NCMBS; 1, 5, 10 for NCES1; 0, 4, 8 for NCBA1) depending on what was available (Cramer and Wehner, 1998a, 1998b). Four families were chosen at random from each population­cycle combination in 1995 and 1996 and were self-pollinated in the greenhouse.

Statistical design. The experiment was a randomized complete block design with four replications in each of two seasons (spring, summer) and in each of 2 years (1995, 1996) at a planting density of 6,450 plants per hectare in a split-plot treatment arrangement. Whole plots were the eight cucumber populations and subplots were three cycles of recurrent selection (early, intermediate, late). Three test plots each of 'Dasher II', Gy 14, 'Poinsett 76' and 'Sumter' were planted in each environment as controls. Seasons were used instead of locations because they provide more information in North Carolina trials (Swallow and Wehner, 1989).

Nine seeds were planted in 3.1-m-long plots, which were found to be the optimum size (Swallow and Wehner, 1986), on raised, shaped beds at the Horticultural Crops Research Station, Clinton, N.C. The spring tests were planted 27 Apr. 1995 and 29 Apr. 1996, and the summer tests on 13 July 1995 and 8 July 1996. The soil type was a mixture (throughout the fields used) of Norfolk, Orangeburg and Rains (fine-loamy, siliceous, thermic, Typic Kandiudults) with some Goldsboro (fine-loamy, siliceous, thermic, Aquic Paleudults). Recommended cultural practices for North Carolina were used throughout the experiment (Schultheis, 1990).

Plots were thinned to three plants at the first true leaf stage. The test plots were harvested once-over by hand at the 10% oversize (>51 mm in diameter for pickles and >60 mm in diameter for slicers) stage as recommended by Miller and Hughes (1969) for optimum yield. Plots were harvested 30 June or 3 July in 1995, and 19 to 21 and 25 June in 1996 for the spring season. Plots were harvested 5 Sept. in 1995, and 19, 22, 23 or 27 Aug. in 1996 for the summer season. Each plot was evaluated for fruit shape, and number of branches, nodes, pistillate flowers, total, early, and marketable fruit. Fruit shape rating reflected how straight, uniform and cylindrical the fruit were, on a scale of 1 to 9, where 1­3 = poor, 4­6 = intermediate, 7­

Cucurbitaceae '98


9 = excellent (Strefeler and Wehner, 1986). The percentage of pistillate nodes was determined by dividing the number of fruit and pistillate flowers by the total number of nodes and multiplying by 100. In addition, the percentage of fruit set was calculated as the percentage of pistillate nodes that developed into harvestable fruit. Plants that had less than five leaves, with no flowers, and <0.4 m vine length were considered weak and were eliminated from consideration in the number of plants per plot and any data collection.

Data analysis. Plots with fewer than two plants were considered missing. Data were analyzed using the GLM procedure of the SAS statistical package (SAS Institute, Cary, N.C.) to determine differences in yield and yield component means for years, seasons, populations, and cycles. In the model, seasons and years were random effects, and populations and cycles were fixed effects. Mean separation tests (lsd) were conducted on those effects with more than two classes. Correlations were measured using PathSAS (Cramer et al., 1998), and values of 0.70 (positive or negative) or greater were considered to be strong while correlations of ­0.69 to 0.69 were considered weak. Those classifications were made based on the fact that a correlation of 0.70 would have a coefficient of determination of 0.49 and the correlation between the two traits would account for approximately 50% of the variation observed for the two traits. In addition, significant correlations with a probability of 95% or greater were also determined and presented.

Results

Yield and yield component analysis. When means were calculated for fruit yield averaged over years, seasons, populations, cycles, and replications, the total fruit yield per plot was similar between populations (Table 1). The NCEP1 and NCH1 populations produced on average two to three more early fruit per plot than the other populations. In addition, pickle populations produced a higher percentage of marketable fruit per plot than slicer populations (Table 1). Among the pickle populations, the NCWBP population produced a lower percentage of marketable fruit than other pickle populations. Fruit shape rating of the

wide-based populations was lower than the rating of the other populations (Table 1). The pickle populations except for the NCWBP populations had more branches and marketable fruit per plant than slicer populations. In addition, the NCEP1 population produced the fewest number of nodes per branch than any other population. The two wide-based populations and the special slicer population (NCBA1) had the highest number of nodes per branch (Table 1). The NCEP1 population had the highest percentage of nodes with pistillate flowers (Table 1). The NCWBS population had the highest percentage of pistillate nodes that developed into fruit by harvest time (Table 1).

When means were calculated for each selection cycle averaged over populations, total, early and percentage marketable yield per plot, fruit set, and marketable fruit yield per plant increased with selection from early to late cycles (Table 1). Selection for increased fruit yield and quality had no apparent change on fruit shape rating, number of branches per plant, number of nodes per branch, and the percentage of pistillate nodes (Table 1). A number of traits exhibited differences between years and/or between seasons. Total and early yield per plot, marketable yield per plant and the number of branches per plant were higher in 1996 than in 1995. Those traits were higher in the summer season than in the spring season when averaged over all populations, cycles and replications (Table 1). Conversely, the number of nodes per branch was higher in 1995 than in 1996, and also was higher in the spring season than in the summer season (Table 1). The percentage of marketable fruit per plot was higher in 1995 than in 1996 and the percentage remained unchanged between seasons. Fruit shape rating, percentage of pistillate nodes, and fruit set remained unchanged between years and between seasons (Table 1).

Latest cycle correlations. In order to determine the relationships between traits, correlations between yield components, and correlations between yield components and fruit yield were calculated for each cycle of each population (Table 2). Because the latest cycle of selection for each population is currently being used in our breeding program, the correlations at that cycle will be discussed for each population. In addition, the

Cucurbitaceae '98


effect of selection on correlations will be discussed for each population. A majority of the correlations between yield components, and the correlations between yield and its components were not significant (82%) or were weak (92%) (<0.70, positive or negative) for the latest cycles of selection (Table 2). However, several correlations were significant and strong. For the NCES1 population, the number of branches per plant was negatively correlated with the number of nodes per branch. In addition, the number of nodes per branch of the NCH1 population was negatively correlated with the proportion of pistillate nodes (Table 2). For both the NCWBP and NCEP1 populations, the percentage of pistillate nodes was negatively correlated with fruit set (Table 2).

With regards to the correlations between yield components and fruit yield, the number of branches per plant was positively correlated with total and marketable yield of the NCMBP population and all three yield traits of the NCH1 population (Table 2). In addition, selection for an increased number of branches per plant would

increase the total yield per plant of the NCWBS, NCMBS, and NCBA1 populations. For the NCWBP population, the number of nodes per branch was positively correlated with total yield per plant, while the same traits were negatively correlated for the NCH1 population (Table 2). Selection for an increased percentage of pistillate nodes in the NCES1 population would increase the total number of fruit per plant. In each population, selection for increased fruit set would not change fruit yield per plant.

Regarding correlations between pairs of yield traits, selection for an increased number of fruit per plant would increase the number of marketable fruit per plant in each population except for the NCWBS population (Table 2). In addition, selection for increased total yield per plant would also increased early yield per plant for the NCEP1, NCH1, and NCMBS populations (Table 2). For those same populations and the NCBA1 population, marketable fruit yield per plant was positively correlated with early fruit yield per plant (Table 2).

Table 1. Mean valuesz of fruit yield and yield component traits for three plants per plot.

Branch Nodes/ Pistill Fruit Market/

Effect Total Early Marketable Shapey /plant branch node (%) set (%) plant

Grand 14.9 4.2 75.7 5.9 8.76 10.30 17.3 39 3.60

Year

1995 11.1 3.4 86.3 5.8 6.25 10.58 16.8 40 3.16

1996 18.4 4.9 66.2 6.0 11.09 10.04 17.7 38 4.01

F ratio 129.2*** 30.2*** 159.0*** 3.7 52.3*** 8.1* 2.5 2 35.5***

Season

Spring 11.3 3.7 75.3 5.9 5.26 10.68 18.1 40 2.90

Summer 18.4 4.7 76.1 5.9 12.25 9.92 16.5 38 4.31

F ratio 114.0*** 17.9** 0.1 0.4 93.1*** 13.0** 4.2 2 79.7***

Cycle

Early 14.4 3.5 72.3 5.9 8.44 10.01 19.1 36 3.38

Intermediate 13.8 4.4 77.4 5.9 8.74 10.40 16.0 38 3.33

Late 16.4 4.8 77.4 5.9 9.10 10.49 16.6 42 4.09

lsd 5% 1.6** 0.8** 3.7* 0.2 1.14 0.83 2.4 4* 0.46**

Population

NCWBP 14.2 3.5 80.4 5.3 6.64 12.74 17.8 39 3.63

NCMBP 14.8 3.6 89.8 6.2 9.53 9.45 14.7 39 4.39

NCEP1 14.3 6.6 90.5 6.4 10.83 7.45 22.7 32 4.23

NCH1 15.7 5.4 89.1 6.0 11.12 8.70 18.6 35 4.42

NCWBS 14.6 3.2 59.3 4.8 7.34 12.44 12.5 47 2.69

NCMBS 16.0 3.9 57.5 6.1 8.41 10.88 16.4 39 3.06

NCES1 13.0 3.3 75.1 6.3 6.93 9.96 16.4 44 3.32

NCBA1 15.0 3.6 65.7 6.0 8.14 11.51 18.1 33 2.84

lsd 5% 2.6 1.2*** 6.0*** 0.4*** 1.87*** 1.36*** 4.0** 6*** 0.75***

zData are means of 384 (grand), 192 (year, season), 128 (cycle), or 48 (population) replications of three plants per plot.

yFruit shape ratings (shape) are 1­3 = poor, 4­6 = intermediate, 7­9 = excellent.

*,**,***Significant at P = 0.05, 0.01, 0.001, respectively.

Cucurbitaceae '98


Changes in correlations with selection. Correlations between pairs of yield components, and correlations between yield and its components were calculated for each cycle of selection for each

population, in order to determine the effect of selection on correlations. Most (82%) of the correlations among populations were weak initially and remained weak through the latest cycle of

Table 2. Correlation coefficientsz between branches per plant, nodes per branch, percentage of pistillate nodes, percentage fruit set and total, marketable and early fruit yield per plant for the latest cycle in each population.

Yield Nodes/ Pistillate Fruit Fruit/plant (no.)

Population component branch nodes (%) set (%) Total Marketable Early

NCWBP Branches/plant ­0.37 ­0.41 0.31 ­0.12 ­0.25 ­0.08

Nodes/branch ­0.07 ­0.22 0.66* 0.25 ­0.38

Pistillate nodes ­0.61* 0.30 0.34 0.55

Fruit set ­0.10 0.10 ­0.16

Total fruit 0.63* ­0.02

Marketable fruit 0.39

NCMBP Branches/plant ­0.19 ­0.28 ­0.02 0.72** 0.67** ­0.18

Nodes/branch 0.30 ­0.28 0.27 0.40 0.19

Pistillate nodes ­0.20 0.21 0.21 0.18

Fruit set 0.15 0.10 0.41

Total fruit 0.95*** 0.41

Marketable fruit 0.13

NCEP1 Branches/plant ­0.30 ­0.11 ­0.03 0.42 0.39 0.48

Nodes/branch 0.23 ­0.16 0.40 0.43 0.23

Pistillate nodes ­0.59* 0.35 0.43 ­0.01

Fruit set 0.19 0.11 0.12

Total fruit 0.99*** 0.61*

Marketable fruit 0.57*

NCH1 Branches/plant ­0.48 0.24 0.13 0.95*** 0.92*** 0.61*

Nodes/branch ­0.71** 0.00 ­0.52* ­0.47 ­0.25

Pistillate nodes ­0.46 0.31 0.26 0.11

Fruit set 0.25 0.29 0.08

Total fruit 0.99*** 0.68**

Marketable fruit 0.72**

NCWBS Branches/plant ­0.38 ­0.38 ­0.32 0.68** ­0.34 ­0.23

Nodes/branch 0.05 ­0.02 0.05 0.44 0.12

Pistillate nodes 0.12 0.12 0.25 0.12

Fruit set ­0.18 0.13 ­0.50

Total fruit ­0.20 ­0.25

Marketable fruit 0.45

NCMBS Branches/plant ­0.50 ­0.20 0.02 0.71** 0.48 0.47

Nodes/branch ­0.19 ­0.12 ­0.24 ­0.15 0.16

Pistillate nodes ­0.32 ­0.02 0.28 0.09

Fruit set 0.43 0.25 0.01

Total fruit 0.78*** 0.58*

Marketable fruit 0.71**

NCES1 Branches/plant ­0.55* ­0.13 ­0.36 0.41 0.43 ­0.07

Nodes/branch 0.39 0.04 0.20 0.16 0.61*

Pistillate nodes ­0.15 0.52* 0.36 0.65**

Fruit set 0.21 0.34 ­0.18

Total fruit 0.95*** 0.48

Marketable fruit 0.42

NCBA1 Branches/plant ­0.32 ­0.01 ­0.23 0.87*** 0.44 0.30

Nodes/branch ­0.38 ­0.22 ­0.23 0.07 0.01

Pistillate nodes ­0.38 0.19 0.00 ­0.03

Fruit set ­0.24 ­0.07 ­0.13

Total fruit 0.71** 0.45

Marketable fruit 0.68**

zStrong correlations were considered to be 0.7 to 1.0 and ­0.7 to ­1.0.

*,**,***Significant at P = 0.05, 0.01, 0.001, respectively.

Cucurbitaceae '98


selection (000, 0+0) (Table 3). Another 5% of the correlations were strong (7 of 8 were positive) at the early cycle and remained strong with selection

(+++, ­0­). All of the positive correlations involved total or marketable yield per plant. In addition, the correlation between total and mar

Table 3. Changesz in correlation coefficients between branches per plant, nodes per branch, percentage of pistillate nodes, percentage fruit set and total, marketable and early fruit yield per plant from early to intermediate to late selection cycles in each population.

Yield Nodes/ Pistillate Fruit Fruit/plant (no.)

Population component branch nodes (%) set (%) Total Marketable Early

NCWBP Branches/plant 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

Nodes/branch 0 0 0 0 0 0 + + 0 0 0 0 0 0 0

Pistillate nodes ­ 0 ­ 0 + 0 0 + 0 0 0 0

Fruit set 0 0 0 0 0 0 0 0 0

Total fruit + + + + 0 0

Marketable fruit + 0 0

NCMBP Branches/plant 0 0 0 0 0 0 0 0 0 0 0 + 0 0 0 0 0 0

Nodes/branch 0 0 0 0 0 0 0 + 0 0 + 0 0 0 0

Pistillate nodes 0 0 0 + 0 0 + 0 0 0 0 0

Fruit set 0 0 0 0 0 0 + 0 0

Total fruit 0 0 0

Marketable fruit 0 0 0

NCEP1 Branches/plant 0 0 0 0 0 0 0 0 0 0 + 0 0 + 0 0 + 0

Nodes/branch 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

Pistillate nodes 0 0 0 0 0 0 0 0 0 0 0 0

Fruit set 0 0 0 0 0 0 0 0 0

Total fruit + + + + + 0

Marketable fruit + + 0

NCH1 Branches/plant 0 0 0 + 0 0 0 0 0 + + + + + + 0 0 0

Nodes/branch 0 0 ­ 0 0 0 0 0 0 0 0 0 0 0 0

Pistillate nodes 0 0 0 0 0 0 0 0 0 0 0 0

Fruit set 0 0 0 0 0 0 0 0 0

Total fruit + + + 0 + 0

Marketable fruit 0 + +

NCWBS Branches/plant 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

Nodes/branch 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

Pistillate nodes 0 0 0 0 0 0 0 0 0 0 0 0

Fruit set 0 0 0 0 0 0 0 0 0

Total fruit 0 + 0 0 0 0

Marketable fruit + 0 0

NCMBS Branches/plant 0 0 0 0 0 0 0 0 0 + + + + 0 0 0 0 0

Nodes/branch 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

Pistillate nodes 0 0 0 0 0 0 0 0 0 0 0 0

Fruit set 0 0 0 0 0 0 0 0 0

Total fruit 0 + + 0 0 0

Marketable fruit 0 + +

NCES1 Branches/plant 0 0 0 0 0 0 0 0 0 0 + 0 0 0 0 0 0 0

Nodes/branch 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

Pistillate nodes 0 ­ 0 0 0 0 0 0 0 0 0 0

Fruit set 0 0 0 0 0 0 0 0 0

Total fruit + + + 0 0 0

Marketable fruit 0 0 0

NCBA1 Branches/plant ­ 0 0 0 0 0 0 0 0 0 + + 0 0 0 0 0 0

Nodes/branch 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

Pistillate nodes 0 0 0 0 0 0 0 0 0 0 0 0

Fruit set + 0 0 0 0 0 + 0 0

Total fruit 0 0 + + 0 0

Marketable fruit 0 0 0

zWeak correlation (0) (­0.69 to 0.69), positive correlation (+) (0.70 to 1.00), negative correlation (­) (­0.69 to ­1.00). Correlations in bold face indicate a change in sign occurred from early to late cycle of selection.

Cucurbitaceae '98


ketable yield per plant remained positive with selection for four populations (Table 3).

The remaining 13% of correlations were either strong initially and weakened with selection (+00. ++0), or were weak initially and strengthened with selection (00­, 00+, 0++) (Table 3). Of the changes in correlations with selection, 68% of the correlations weakened with selection. A majority of correlations were positive and weakened with selection. In addition, a majority of the correlations that strengthened with selection, became positive by the latest cycle. With regards to the changes in correlations between yield components with selection, the percentage of pistillate nodes of the NCH1 populations became negatively correlated with nodes per branch by cycle 10. However, the positive correlation with branches per plant weakened with selection. A negative correlation between the number of branches per plant and the number of nodes per branch of the NCBA1 population also weakened with selection.

For the NCMBP and NCBA1 populations, weak correlations between branches and total yield per plant became positive with selection (Table 3). For the NCMBP population, positive correlations of the percentage of pistillate nodes with total and marketable yield per plant weakened from cycle 2 to cycle 10 (Table 3). For the NCBA1 population, fruit set was positively correlated with total and early yield per plant at cycle 0 and both correlations had weakened by cycle 8 (Table 3). The correlation between total and marketable yield per plant became positive by the latest cycle for both the NCMBS and NCBA1 populations. For the NCWBP, NCEP1, and NCBA1 populations, the positive correlation between total and early yield per plant at the early cycle weakened with selection in each population. The correlation between marketable and early yield per plant was positive initially and weakened with selection in the NCWBP, NCEP1, and NCWBS populations. However, the correlation was weak initially and became positive with selection in the NCH1 and NCMBS populations (Table 3).

Discussion

Based on the results of this study, selection of yield components to improve fruit yield would

not increase yield for the populations studied and for particular yield components measured. Of the yield components studied, the number of branches per plant exhibited the strongest, positive correlation with yield in five of the eight populations. Selection for an increased number of branches per plant would increase the number of total fruit per plant for several populations. The strong, positive correlation between branch number and fruit yield might be related to plant density. Each of the populations tested in this study produced more branches and fruit per plant than when the populations were tested at a higher planting density (Cramer and Wehner, 1998a, 1998b, 1998c). In addition, strong, positive correlations between branches per plant and fruit yield per plant were observed for only two populations when tested at a higher planting density. With more branches and fruit produced at the lower density, the correlation between the two traits became stronger. The strong correlation might also be related to the phenomenon of first-fruit inhibition. In cucumber, the development of the first fertilized pistillate flower inhibits the subsequent development of other fertilized flowers (McCollum, 1934). This inhibition was hypothesized to be related to a limited amount of photosynthates that can support the growth of one fruit at a time (Pharr et al., 1985). However, if photosynthate production were to increase, through increased photosynthetic area (more branches per plant), the number of developing fruit per plant would also increase.

The strong, positive correlation between the number of branches and fruit per plant of the NCH1 population might be related to germplasm used in the formation of this population. LJ 90430, a multibranched, multifruiting accession of C. sativus var. hardwickii, was used in the formation of the NCH1 population (Wehner, 1998b). This accession can support the development of multiple fruit at one time. However, the fruit are small and a large photosynthetic area is developed before fruiting begins. C. sativus var. hardwickii plants were able to produce multiple fruit because of high leaf area and multiple branching (Ramirez et al., 1988). Even though the NCH1 population did not produce small fruit and fruit development occurred soon after flowering, the strong positive

Cucurbitaceae '98


correlation between branch number and fruit production of LJ 90430 had been maintained in the population. Branch and fruit number were also positively correlated when the population was tested at a higher planting density (Cramer and Wehner, 1998a, 1998d).

Cucumis sativus var. hardwickii germplasm might also be responsible for the negative correlation between the number of nodes per branch and the percentage of pistillate nodes for the NCH1 population (Table 2). Plants with this germplasm produced many branches with few nodes per branch, and produced a pistillate flower at each node. With an increase in the number of nodes per branch, the plant might not be able to support a pistillate flower at each node. This response would explain the negative correlation between number of nodes per branch and percentage of pistillate nodes. However, if this hypothesis were true, one would expect to observe a negative correlation between the traits at cycle 0, and not at cycle 10, since the population probably had more C. sativus var. hardwickii germplasm at cycle 0 than at cycle 10.

Selection for increased fruit yield and quality was particularly effective for improving total, marketable and early yield when averaged among populations (Table 1). Selection did not effect yield component means. However, selection did influence the correlations between yield and its components. When strong correlations were present at early selection cycles, selection reduced those correlations to weak correlations. In addition, a majority of the correlations between yield components and yield were weak at the latest cycle of selection. Weak correlations and weakening of strong correlations with selection was also observed for the same populations tested at a higher planting density (Cramer and Wehner, 1998a, 1998b, 1998d). Those observations would suggest that for most populations, selection for improved yield favors weak correlations between yield and its components, or that no particular yield component has more influence than other components in its contribution to fruit yield.

Regarding correlations between yield components, strong correlations were observed at the latest cycle among populations in only four in

stances. Three of those correlations involved the percentage of pistillate nodes. In each case, the particular yield component was negatively correlated with the percentage of pistillate nodes. In addition, negative correlations between pistillate nodes and other yield components were also observed for the same populations tested at a higher planting density (Cramer and Wehner, 1998a, 1998b, 1998d).

The low planting density used in this study would allow for easy data collection and selection, but was much lower than the density used by growers. In addition, Cramer and Wehner (1998d) did not, however, observe many similarities in the correlations between yield components and yield when populations were tested at two different planting densities. Thus, low density selection for yield components may not be effective in breeding programs involved in yield improvement.

In order for yield component selection to result in more yield gains than selection for yield itself, heritability of yield components must be higher than heritability of fruit yield. Narrow-sense heritability of fruit yield in cucumber has been determined to be 0.17 to 0.25 (Smith et al., 1978; Strefeler and Wehner, 1986). Heritability of yield components in cucumbers has not been determined. However, the general heritability can be estimated based upon the number of effective factors involved in each trait. For example, sex expression in cucumbers is controlled by two major genes with several modifiers (Pierce and Wehner, 1990). Even though the number of effective factors involved in cucumber yield has not been determined, the number of effective factors for yield in other crops has been numerous and the same would be expected in cucumbers. With fewer genes involved in sex expression than fruit yield, heritability of sex expression would be expected to be higher than heritability of yield. Even with high heritability of yield components, the product of heritability by correlation for yield components must still be higher than heritability of yield in order for indirect selection of yield components to be effective.

In conclusion, selection for yield components would not increase fruit yield in most cucumber populations, since the correlations were weak. At

Cucurbitaceae '98


fruiting and vegetative plants of Cucumis sativus L. Plant Physiol. 77:104­108.

Pierce, L.K. and T.C. Wehner. 1990. Review of genes and linkage groups in cucumber. HortScience 25:605­615.

Prasad, V.S.R.K. and D.P. Singh. 1994a. Diallel analysis of yield components in slicing cucumber (Cucumis sativus L.). J. Res. (BAU) 6:151­154.

Prasad, V.S.R.K. and D.P. Singh. 1994b. Genetic association and interrelationship between yield components in cucumber. J. Maharashtra Agr. Univ. 19:147­148.

Ramirez, D.R., T.C. Wehner, and C.H. Miller. 1988. Source limitation by defoliation and its effect on dry matter production and yield of cucumber. HortScience 23:704­706.

Schultheis, J.R. 1990. Pickling cucumbers. N.C. State Agr. Ext. Hort. Info. Lflt. No. 14­A.

Siefker, J.H. and J.F. Hancock. 1986. Yield component interactions in cultivars of the highbush blueberry. J. Amer. Soc. Hort. Sci. 111:606­608.

Solanki, S.S. and A. Shah. 1989. Path analysis of fruit yield components in cucumber. Prog. Hort. 21:322­324.

Smith, O.S., R.L. Lower, and R.H. Moll. 1978. Estimates of heritabilities and variance components in pickling cucumbers. J. Amer. Soc. Hort. Sci. 103:222­225.

Strefeler, M.S. and T.C. Wehner. 1986. Estimates of heritabilities and genetic variances of three yield and five quality traits in three fresh-market cucumber populations. J. Amer. Soc. Hort. Sci. 111:599­605.

Swallow, W.H. and T.C. Wehner. 1986. Optimum plot size determination and its application to cucumber yield trials. Euphytica 35:421­432.

Swallow, W.H. and T.C. Wehner. 1989. Optimum allocation of plots to years, seasons, locations, and replications, and its application to once-over-harvest cucumber trials. Euphytica 43:59­68.

U.S. Department of Agriculture. 1998. Agricultural statistics. U.S. Government Printing Office, Washington D.C.

Wehner, T.C. 1989. Breeding for improved yield in cucumber, p. 323­359. In: J. Janick (ed.). Plant breeding reviews. vol. 6. AVI Press, Stamford, Conn.

Wehner, T.C. 1997. Three pickling cucumber populations: NCWBP, NCMBP, and NCEP1. HortScience 32:941­944.

Wehner, T.C. 1998a. Three slicing cucumber populations: NCWBS, NCMBS, and NCES1. HortScience 33:168­170.

Wehner, T.C. 1998b. Two special cucumber populations: NCH1 and NCBA1. HortScience 33 (in press).

Wehner, T.C. and C.S. Cramer. 1996a. Gain for pickling cucumber yield and fruit shape using recurrent selection. Crop Sci. 36:1538­1544.

Wehner, T.C. and C.S. Cramer. 1996b. Ten cycles of recurrent selection for fruit yield, earliness, and quality in three slicing cucumber populations. J. Amer. Soc. Hort. Sci. 134:322­326.

Yin, M. and H. Cui. 1994. Analysis of component traits for early yield in cucumber. Cucurbit Genet. Coop. Rpt. 17:27­29.

Zhang, M. and H. Cui. 1994. Some morphological parameters involving the mechanism of early yield in cucumber. Cucurbit Genet. Coop. Rpt. 17:24­26.

this planting density, selection for an increased number of branches per plant would increase the number of fruit produced per plant for several populations. Selection for improved fruit yield and quality resulted in increased total, marketable and early yield and no change in yield components means or correlations with yield. Further work on yield component selection methods and heritability is needed for the benefits of indirect selection for cucumber yield to be realized.

Literature cited

AbuSaleha and O.P. Dutta. 1988. Interrelationship of yield components in cucumber. Veg. Sci. 15:79­85.

Cramer, C.S. and T.C. Wehner. 1998a. Fruit yield and yield component means and correlations of four pickling cucumber populations improved through recurrent selection. J. Amer. Soc. Hort. Sci. (in press).

Cramer, C.S. and T.C. Wehner. 1998b. Fruit yield and yield component means and correlations of four slicing cucumber populations improved through six to ten cycles of recurrent selection. J. Amer. Soc. Hort. Sci. 123:388­395.

Cramer, C.S. and T.C. Wehner. 1998c. Planting density effects on cucumber fruit yield and yield components. Plant Breeding (in press).

Cramer, C.S. and T.C. Wehner. 1998d. Planting density effects on the correlation between cucumber fruit yield and yield components. Euphytica (in press).

Cramer, C.S., T.C. Wehner, and S.B. Donaghy. 1998. PathSAS: A SAS computer program for path coefficient analysis of quantitative data. J. Hered. (in press).

Hancock, J.F., M.P. Pritts, and J.H. Siefker. 1984. Yield components of strawberries maintained in ribbons and matted rows. Crop Res. 24:37­43.

Lertrat, L. and R.L. Lower. 1983. Pickling cucumber population improvement for increased fruit yield. Cucurbit Genet. Coop. Rpt. 6:18­19.

Lertrat, L. and R.L. Lower. 1984. Pickling cucumber population improvement for increased fruit yield II. Cucurbit Genet. Coop. Rpt. 7:9.

McCollum, J.P. 1934. Vegetative and reproductive responses associated with fruit development in the cucumber. Memoir 163, Cornell Univ. Agr. Expt. Sta. Ithaca, N.Y.

McGiffen, Jr., M.E., D.J. Pantone, and J.B. Masiunas. 1994. Path analysis of tomato yield components in relation to competition with black and eastern black nightshade. J. Amer. Soc. Hort. Sci. 119:6­11.

Miller, C.H. and G.R. Hughes. 1969. Harvest indices for pickling cucumbers for once-over harvested systems. J. Amer. Soc. Hort. Sci. 94:485­487.

Nienhuis, J. and R.L. Lower. 1988. Comparison of two recurrent selection procedures for yield in two pickling cucumber populations. J. Amer. Soc. Hort. Sci. 113:272­276.

Pharr, D.M., S.C. Huber, and H.N. Sox. 1985. Leaf carbohydrate status and enzymes of translocate synthesis in

Cucurbitaceae '98