Genetic Path Analysis of Early Yield in Cucumber

admin
Cucurbit Genetics Cooperative Report 22:1-2 (article 2) 1999

Meng Zhang, Xiaofen Wang and Hongwen Cui
Department of Horticulture, Northwestern Agricultural university, Yangling Shaanxi, 712100, P.R. China

Early high yield in cucumber would be an important way for solving the problem of the absence of fresh vegetables in spring cultivation in china. Thus, breeders have paid more and more attention to the development of high yielding early lines and hybrids. It is essential to characterize the physiological mechanisms of early yield in order to increase the effectiveness of direct and indirect selection. Here is one of a series of studies on this aspect (1).

Materials and Methods

An experiment was conducted at the Horticulture Station of the Northwestern Agricultural University. Twenty-four cultigens were planted in a randomized block design with three replications. Ten plants of each cultigen were randomly chosen to evaluate for early yield (Y) and 14 relevant traits in the early stage of plant development. Traits were selected as follows: (1) node position of the first pistillate flower (x1);(2) days from sowing to the first pistillate flowering plant in the population (x2); (3) days from sowing to pistillate flowering of 50% of the plants (x3);(4) days from sowing to first male flowering plant in the population (x4); (5) days from sowing to staminate flowering in 50% of the plants (x5);(6) leaf area per plant (x6); (7) fruit length (x7); (8) leaf number (x8); (9) pistillate flower density (main vine) (x10); (11) number of staminate flowers (main vine( (x11); (12) number of harvested fruits per plant (x12); (13) average fruit weight (x13); and (14) downy mildew index (x14).

The early yield was considered as the primary trait while the other traits were considered secondary traits. On the basis of genetic correlation analysis, the path coefficients of each trait to early yield were obtained by partitioning the genetic correlation coefficient according to the formulas below:

P1+r12P2 …………………………………………………………………. +r1mPm = r1y
R21P1+ P2+ ………………………………………………………………….. +r2mPm = r2y
Rm1P1+rm2P2+ ………………………………………………………………….. +Pm = Pmy

Path coefficient descriptors were designated as:

  • Pi is the path coefficient of trait x1 to early yield;
  • rij is the genetic correlation coefficient between x1 and xj , and
  • riy is the genetic correlation coefficient between xi and yield.

The path coefficients obtained were tested, and if one or more were not significant, the trait with the smallest t or F value was eliminated and then coefficients were calculated again. This process continued until all path coefficients were significant (2).

Results

The results (Table 1) show the path coefficient obtained from progressive elimination of genetic correlation coefficients. Some coefficients were small. The first coefficient to be eliminated was pistillate flower density (x9) and then traits x10, x2, x1, x5, x6, x11 were eliminated sequestially thereafter. Only seven traits were retained after examination. The results of path analysis are shown in Table 2. The three traits with the largest direct action to early yield were average fruit weight (x13), number of harvested fruit per plant (x12_ and average fruit length (x7). Their actions were positive. Positive selection can be carried out for early maturity. Downy mildew index (x 14) had the least direct action in the path analysis, and primarily exerted an indirect influence on early yield via fruit length (x7). The days from sowing to pistillate flowering of 50% of the plants (x3) had a negative direct effect on early yield, and had a slightly larger negative indirect effect via the way of number of harvested fruit (x12). Days from sowing to first male flowering plant (x4) had a positive effect via fruit length (x7) and fruit weight (x13). Thus, it is suggested that selection be applied to earlier pistillate flowering and later male flowering plant when breeding for maturity. Although leaf number (x8) had a positive correlation with early yield, it had a certain negative direct action on early yield. On the other hand, it positively influenced early yield indirectly via fruit length (x7) and fruit weight (x13). These plant growth variables reflected the antagonism and unity between vegetative and reproductive plant development.

Discussion

Path analysis can be used to identify direct and indirect action of traits, and thus can be helpful during breeding. Because path analysis results are often not tested we suggest that the utility of path coefficients be thoroughly evaluated in order to document their utility before adapt ion to breeding programs. In addition, it should he mentioned that the traits eliminated during the path coefficient analysis have no direct action on the objective trait. Thus, we cannot rule out the possibility that traits are “pre-eliminated” during such analyses have no indirect action through the expression of other traits. If we take a multi-stage path analysis approach (i.e., analysis based on the action mechanism and the physiological principal), we may learn more about the status and correlation among the traits examined. The total determination (R) was R = 1.011. This made it impossible to estimate the standard error of path coefficient and the environment (Pe). This phenomenon has been elucidated by classical variance and covariance analysis. A better method should be sought to solve this problem.

Table 1. Genetic correlation coefficient of 14 traits to early yield.

Traits X1 X2 X3 X4 X5
(Genetic correlation coefficient) -0.0329 -0.0265 -0.2187 0.336 -0.0414
Traits X6 X7 X8 X9 X10
(Geneticcorrelation coefficient) 0.1215 0.4577 -0.1404 0.0126 0.0085
Traits X1 X12 X13 X14
(Genetic correlation coefficient) 0.1464 0.5778 0.4368 0.0205

Table 2. Direct and indirect action of seven traits to early yield.

Traits
Corella
Direct

Indirect action of related traits
Corr. coef.
Dir. act.
(Riy)z
(Pi)y
X3 -> Y
X4 -> Y
X7 -> Y
X8 -> Y
X12 -> Y
X13 -> Y
X1 -> Y
X3 -0.218 -0.253 0.075 0.183 -0.066 -0.310 0.170 -0.018
X4 0.367 0.225 -0.084 0.175 -0.066 -0.075 0.241 -0.048
X7 0.741 0.393 -0.118 0.100 -0.095 0.067 0.434 -0.041
X8 0.460 -0.128 -0.130 0.117 0.292 -0.010 0.350 -0.031
X12 0.658 0.453 0.173 -0.042 0.058 0.003 0.027 0.008
X13 0.758 0.463 -0.093 0.007 0.369 -0.097 0.027 -0.028
X14 -0.237 0.086 0.052 -0.127 -0.187 0.046 0.042 -0.150

z Correlation coefficient.
y Direct action of trait.

Literature Cited

  1. Meng, Z. 1993. Application of factor analysis to cucumber breeding. Cucurbit Genet. Coop. Report 16:27-29.
  2. Huidong, M. 1992. Agriculture experimental statistics. Shanghai Science and Technology Publishing House.
  3. Jingfu, L. 1985. Correlation and path analysis of main agronomic traits to yield in tomato. Acta Northeastern Agr. Col. 2:59-64.
  4. Tails, G.M. Sampling errors of genetic correlation coefficients calculated from the analysis of variance and covariance. Aust. J. Stat. 1:53-58.