Scientific Journal of Agronomy and Plant Breeding, 1(1): 51-58, January 2013
Available online at http://Sjapb.khouzestan.srbiau.ac.ir/en
ISSN: 2322-3227
Evaluate Correlation and Path Coefficient Analysis for Agronomic Traits on
Grain Yield of Canola (Brassica napus L.) Genotypes
Seyed Alireza Seyedmohammadi*
Payammenoor University, Ahvaz Branch, Ahvaz, Iran
* Corresponding Author: Seyedmohammadi.ali@gmail.com
Received Date: 10.7.2012 Accepted Date: 12.23.2012
Abstract
To evaluate most important traits affecting grain yield in canola and the quantity
of direct and indirect effects on grain yield, an experiment was conducted
with rapeseed varieties in a randomized complete block design (RCBD) with
four replications. Correlation coefficients analysis illustrated that the biologic
yield, harvest index, grain weight, the number of grains per pod, number of
pods per plant, plant height, Planting to physiologic maturity and Planting to
flowering trait have a positive significant correlation with grain yield. Stepwise
regression and path analysis indicated that, number of pods per plant had the
highest direct effect on grain yield. In addition, biologic yield, grain weight, and
planting to flowering and Planting to physiologic maturity also had a high direct
effects on grain yield. Thus, direct selection for these traits is suggested.
Key words: Rapeseed, Yield, Correlation, Path analysis
Introduction
Knowing about grain yield issue
and its components plays an important
role for being successful in evaluate
programs. Success in breeding and
having fruitful varieties of agricultural
products with a higher quality depends
on knowledge about genetic control of
grain yield controlling and its relation
with yield components, also to phenologic
traits and forage quality (Jafari,
2001). Rapeseed grain yield is depended
on the capacity of variety
yielding, climatic conditions, the type
of soil and agronomic management.
Also, genetic and agronomic factors
determine growth of the plant and
grain yield (Mirmosavi et al., 2004).
Some statistical methods, such as correlation
analysis, indicate partial role
of each component of yield in the
amount of yield; also, they provide
necessary information for choosing
indirect traits in superior genotypes to
have higher yield breeding value (Farshadfar,
1997). Correlation between
traits is useful for evaluation and planning
on breeding programs. In other
words, when an evaluation is conducted
on a trait, knowing its effects
on the other traits is very important.
Also by knowing if correlation exists
between important traits, interpretation
on previous results would become easier
and could be the basis for effective
future plans. Also correlation between
important and non-important traits
provides plant breeding experts with a
significant assistance in indirect selection
of important traits, through nonimportant
traits which their measurements
are easier (Qulipor et al., 2004).
One of the most important objectives
of rapeseed breeding is to increase
yield in a region.
Seyedmohammadi, Evaluate Correlation and Path Coefficient Analysis for Agronomic … 52
Increase in physiological efficiency
of the plant improves yield in a region.
This is because many other factors indirectly
affect yield increase. Yield is
also affected by all parts of the plant,
and it is considered as the final target
of many characteristics (Farsi and
Baqeri, 1998). In this method, Correlation
coefficient which exists between
two traits is divided in to components
which measure direct and indirect effects.
Using this method requires the
knowledge about cause and effect relations
which exist between traits, and
assuredly must determine the direction
of causes according to previous informations
and experimental evidences
(Garcia Del Moral et al., 1999, Zemmermann
et al., 1994). In fact, path
coefficient analysis depicts a more
complete image of a simple correlation,
and correlation coefficient between
two variables, divides cause and
effect into direct and indirect effects.
Several studies have been conducted
to determine the correlation between
different traits and to divide them by
path coefficient analysis method in
rapeseed (Matlabipour et al., 2000,
Qulipor et al., 2004, Solymanzadeh et
al., 2007). According to the results of
path coefficient analysis in rapeseed
plant, it was illustrated that the duration
of growth had the most direct and
negative effects on oil grain percentage
and the number of pods per plant
had the most positive and direct effect
on grain yield (Mirmosavi et al.,
2004). Some researchers introduced
traits of the number of fertile stems,
and grains per pod, as an index for selecting
yield of rice varieties although
in path coefficient analysis method,
direct effect of 100- grain weight on
yield was relatively high (Abozary
Gasaforodi et al., 2006). By having
rapeseed varieties evaluated, observed
that grain yield has the most correlation
with 1000- grain weight, biological
yield, harvest index and oil yield.
Accumulation of dry matter in plant
causes better assimilate transfer, therefore
the plant makes the best use of
assimilate for grain filling (Rabiee et
al., 2004). A researcher reported that
there is a positive and significant correlation
between Plant height traits,
grains per pod and 1000- grain weight
with grain yield (Shirany rad, 1994).
The purpose of this research is to identify
the correlation between some agronomic
traits, and to recognize traits
with maximum direct and indirect effects
on grain yield by making use of
path coefficient analysis, so that by
using important traits which are related
to yield, we can achieve improvement
on these breeding goals.
Materials and Methods
This research was conducted in the
agronomic year of 2008-2009, in randomized
complete block design
(RCBD) with four replications, in the
experimental field of Payamenoor
University (Salemi Farm). The texture
of the soil was Silt clay, electricity
conductivity of condensed saturation
was 2.2 ds.m-1 and acidity of the soil
was 7.1. The average annual precipitation
was 246 mm, (30-year) daily
temperature was 23.95 0C, the average
precipitation of agronomic year was
136.88 mm and the average temperature
of agronomic year was 20 0C in
this region. Each plot consisted of 8
rows with 30 cm distance from each
other and each plot was 6 meters long.
The average distance between the
plants was considered to be 3 to 4 cm.
The amount of fertilizer used in the
field was according to soil tests before
planting, i.e, 50 Kg N ha -1, 100 Kg
P2O5 ha -1, 100 Kg K2O ha -1 and 100
Kg N ha -1 during stem elongation period
was utilized. In order to deterScientific
Journal of Agronomy and Plant Breeding, 1(1): 51-58, January 2013 53
mine grain yield components, at the
physiological maturity stage, 10 plants
were taken from each plot randomly.
Then traits of pods per plant, grains
per pod and 1000- grain weight were
evaluated. In the final harvest, from
one square meter of each plot, grain
yield and biological yield were estimated,
and harvest index was calculated;
by means of grain yield proportion
to biological yield. For studying
the type of relations between the independent
variables (agronomic traits
and yield components), and the dependent
variable (grain yield), grain
yield path coefficient analysis was
performed in order to achieve the direct
and indirect effects. For doing statistical
calculations, identifying correlation
coefficients and regression
analysis, Minitab 14 was utilized, and
for coefficient analysis, Path74 software
was used.
Results and Discussion
Simple correlation coefficients between
traits
These coefficients were estimated
according to Pearson coefficient The
most positive and significant correlation
was observed in biologic yield
trait (r=0.944), harvest index
(r=0.810**), pods per plant (r=0.963**),
grain weight (r=0.911**), Planting to
flowering (r=0.830**), plant height
(r=0.720**) and Planting to physiologic
maturity trait (r=0.650**). The traits of
grains per pod (r=0.580*) and days to
emergence (r=-0.656*) had correlation
with the grain yield at 5 percent probability
level. Some researchers reported
that traits of -days to flowering
-and -days to maturity- have a significant
and positive correlation with
grain yield of rapeseed varieties
(Solymanzadeh et al., 2007), therefore
varieties with longer Planting to flowering
would have a better chance for
fertilization of flowers and turning
them to pods. In serotinal varieties or
delayed-growing plants the decrease
of length in growing period, poor environmental
conditions (temperature
and humidity) during the Planting to
flowering and fertilization and pod
formation, decreases in the number of
pods per plant, the number and grain
weight finally lead to the decrease of
rapeseed yield (Rahnama and Bakhshandeh,
2005). Maximum correlation
coefficient in grain yield by number of
pods per plant (r=0.963**), is because
Pods are assimilate suppliers for the
grains, therefore, we can consider the
positive and significant correlation of
grains per pod. As a result, the more
this trait is observed, the bigger sink
plant would have for metabolic materials.
These results were conformed by
other researchers (Jorgeh, 2003,
Mendham and scot, 1975). The increase
of biologic yield and its direct
relation with grain yield show the relations
between photosynthesis efficiency
of plant and grain yield, therefore
varieties which have gained more
benefits from production because of
growth conditions and they keep more
photosynthesis materials in their sinks,
have more efficiency. This was in conformity
with the results of some other
researchers (Abozary Gasaforodi,
2002, Matlabipor et al., 2000, Qulipor,
et al., 2004). The significant and positive
correlation between harvest index
and grain yield (r=0.822**) indicated
efficiency and kind of photosynthesis
products distribution in different parts
of plant, especially in grains. Results
obtained by some other researchers the
mentioned issues (Rabiee et al., 2004,
Valadyani et al., 2004).
Scientific Journal of Agronomy and Plant Breeding, 1(1): 51-58, January 2013 53
Table 1- correlation coefficients of grain yield and studied traits of rapeseed genotypes
Traits Day to
emergence
Percent of
emergence
Initiation of
flowering
End of
flowering
Planting to
flowering
Planting to
physiologic
maturity
Plant
heights (cm)
Pods Per
plant
Grains per
pod
Grain
weight (gr)
Harvest
index
(%)
Biologic
yield
(kg.ha-1)
Percent of
emergence -0.676**
Initiation of
flowering -0.197 ns -0.870 ns
End of
flowering
-0.236 ns -0.281 ns 0.862**
Planting to
flowering
-0.851** 0.499* 0.41 ns 0.513*
Planting to
physiologic
maturity
-0.512* 0.190 ns 0.687** 0.790** 0.793**
Plant
heights (cm)
-0.700** 0.390 ns 0.580* 0.662** 0.821** 0.880**
Pods per
plant
-0.780** 0.288 ns 0.490 ns 0.489 ns 0.899** 0.780** 0.808**
Grains
per pod
-0.360 ns 0.299 ns 0.400 ns 0.440 ns 0.731** 0.404 ns 0.626* 0.720*
Grain weight
(gr)
-0.645** 0.188 ns 0.300 ns 0.390 ns 0.932** 0.590* 0.693** 0.893** 0.699**
Harvest
index (%)
-0.622* 0.187 ns 0.439 ns 0.489 ns 0.880** 0.693** 0.688** 0.987** 0.668* 0.890**
Biologic
Yield (kg.ha-1)
-0.777** 0.204 ns 0.430 ns 0.585* 0.799** 0.729** 0.870** 0.973** 0.655* 0.890** 0.959**
Grain yield
(Kg.ha-1)
-0.649** 0.202 ns 0.407 ns 0.430 ns 0.830** 0.650** 0.720** 0.963** 0.580* 0.911** 0.822** 0.944**
ns, * and **: No significant and Significant at 5 and 1% Level of Probability, Respectively
Seyedmohammadi, Evaluate Correlation and Path Coefficient Analysis for Agronomic … 54
Scientific Journal of Agronomy and Plant Breeding, 1(1): 51-58, January 2013 53
Stepwise regression analysis
In stepwise regression analysis,
grain yield was considered as a dependent
variable, while other traits
were considered as independent variables.
All the traits were put into regression
model and finally five traits
i.e. pods per plant, Biologic yield,
Planting to flowering, grain weight
and Planting to physiologic maturity
remained in the regression model. This
model generally justified 94% of
changes, related to the grain yield
(Table 2). Other traits which were
studied did not have significant influences
on this model, therefore variations
on grain yield, are because of
differences in the mentioned traits
above. In another study of project in
stepwise regression analysis of traits
of pods per plant, 64% of coefficient,
the number of grains per pod 67%,
1000- grain weight 72%, oil percentage78%
and the number of nodes in
stem verified 80% coefficient of
changes in regression model, which
were related to rapeseed varieties
comparison (Baradaran., et al., 2006).
In regression model conducted by
some researchers at rapeseed in order
to determine effective traits on seed
yield, traits of grain weight, total
number of pods and the number of
grains per pod were entered into the
model (Rahnamaee tak et al., 2007).
The results of stepwise regression
analysis on bean varieties indicated
that, 5 traits of pod weight, the number
of grains in plant, the total number of
pods, biological yield and harvest index
justified 97% of the changes in
grain yield, yet pod weight devoted
95% of changes by it (Amini, 2009).
Path analysis
In order to have a better conception
and interpretation of the results
achieved by correlation and stepwise
regression analysis, path analysis for
variables was conducted which were
entered into the final regression
model. The highest direct effect was
because of pods per plant (Table 2).
All indirect effects in this analysis
were negligible or negative. Pods per
plant trait had the highest positivedirect
effect on grain yield, also it had
positive indirect effects on Biologic
yield (0.138), grain weight (0.027),
and planting to physiologic maturity
(0.04). Negative indirect effect was
through Planting to flowering. Due to
the fact that, negative- indirect effects
were small, they had no profound impact
on positive-indirect effects. There
was a significant correlation between
pods per plant and grain yield, which
was in conformity with the results of
some other researchers (Baradaran., et
al., 2006, Hakan- Ozer, and Unsal,
1999). It seems that the sink which is
made by the higher number of grains
per pod, is main factor for having
more yield. Changes in the number of
pods, potentially increase yield because
the source has photosynthesis
and increases sink capacity. On the
other hand, earlier characteristics can
have higher direct effects on production
and have indirect effects also, on
yield through other traits. This appears
on the other processes of growth
(Vaeezi et al., 2000). After pods per
plant trait, Biologic yield with direct
effect (0.623) plays an important role.
It seems with an increase in biological
yield (because of having source productions),
preparation i.e. grains for
receiving Biologic yield, through traits
of the number of pods per plant (0.21),
grain weight (0.099) and maturity duration
(0.048), had positive-indirect
effects on grain yield, while its indirect
effect via Planting to flowering
was negative (- 0.036).
55
Scientific Journal of Agronomy and Plant Breeding, 1(1): 51-58, January 2013 53
Table 2 - Stepwise regression analysis for grain yield (dependent variable) with studied
traits in rapeseed genotypes
Table 3 - Path analysis of yield traits between remaining in the stepwise regression
model
Indirect effect
Character Direct
effect
Pod Per
plant
Biologic
yield
Grain
weight
Planting to
flowering
Planting to
physiologic
maturity
Total
effect
Pod per
plant 0.752 ---- 0.142 0.035 -0.034 0.06 0.963
Biologic
yield 0.623 0.21 ---- 0.099 -0.036 0.048 0.944
Grain
weight 0.118 0.640 0.152 ---- -0.036 0.037 0.911
Planting to
flowering 0.540 0.14 0.046 0.054 ---- 0.050 0.830
Planting to
physiologic
maturity
0.063 0.479 0.126 0.069 -0.059 ---- 0.650
Error= 0.54 (Residual effect)
By using path analysis on agronomic
traits of pea, realized that, the
maximum direct effect on grain yield
was related to the following issues;
harvest index trait (95.6), Biologic
yield (48.3) and number of pods per
plant (9.34). Planting to flowering
trait, by having positive-direct effect
on grain yield (0.570), also by having
positive-indirect effects of number of
pods per plant (0.11), Biologic yield
(0.046), Grain weight (0.051) and
planting to physiologic maturity
(0.047). caused positive correlation
with grain yield (Fayaz and Talebi,
2009). These results have been in
agreement with the results of some
other researchers (Solymanzadeh et
al., 2007). Grain weight had a little
positive-direct effect on grain yield
(0.118), while positive-indirect effects
of this trait with the number of pods
per plant (0.640), Biologic yield
(0.152) and planting to physiologic
maturity (0.037) devoted a large
amount of correlation with grain yield.
Grain weight should be considered for
breeding or increasing yield. Due to
the fact that yield components are produced
in physiological order, and lack
of one component can be compensated
by another component, this trait increases
deficiencies of first components
of plant yield.
The Traits added to model stepwise regression
1 2 3 4 5
Constant -1599 -1882 -1759 -1733 -1901
Pods per plant 188 164 155.7 139.1 147.2
Biologic yield 1.03 1.07 0.80 0.71
Planting to flowering -2.4 -2.1 -3.1
Grain weight 108 136
Planting to physiologic
maturity 4.1
Coefficient R2 (%) 47 64 73 86 94
Seyedmohammadi, Evaluate Correlation and Path Coefficient Analysis for Agronomic … 56
Seyedmohammadi, Evaluate Correlation and Path Coefficient Analysis for Agronomic … 52
In such situations, the increase is
considered as an important index for a
variety. Some researchers have also
come to this conclusion in their studies
(Dipenbrock, 2005, Qulipor et al.,
2004, and Solymanzadeh et al., 2007).
The last trait which was gained by
stepwise regression analysis was the
length of planting to physiologic maturity.
The same as Grain weight, positive-
indirect effects of this trait with
traits of pods per plant (0.479), Biologic
yield (0.126) and Grain weight
(0.069), confirmed a large amount of
correlation in maturity duration with
grain yield. However, its direct effect
on grain yield was low (0.063).
Conclusion
According to the results of this research,
pods per plant, Biologic yield
and Planting to flowering, had the
positive-direct effects on rapeseed
grain yield. Therefore, better genotypes
for grain yield improvement can
be made by direct selection of these
traits. Some researchers also obtained
similar results (Baradaran et al., 2006,
Rahnamaee tak et al., 2007).
References
1- Abozary Gasaforodi, A. 2002. Genetic and morphological diversity of correlation
and data storage protein electrophoresis in rice varieties. MS Thesis. Faculty
of Agriculture. Gilan University. 119 pp. (In Farsi).
2- Abozary Gasaforodi, A., R. Honarnejade, M. H. Fotokian and A. Alami.
2006. Correlation and path analysis of agronomic traits in rice. J. Sci. Tech.
Agric. Natural Res. 10 (2): 99-106. (In Farsi).
3- Amini, F. 2009. Correlation between rice quantitative traits with grain yield
through path analysis. MS Thesis. Faculty of Agriculture. Birjand University.
(In Farsi).
4- Baradaran, R., A. Majidi Harvan, F. Darvish and M. Azizi. 2006. Relationship
between correlation and path coefficient analysis of yield in Rapeseed. J. Agric.
Sci. 12(4): 127-139. (In Farsi).
5- Dipenbrock, W. 2005. Yield analysis of winter oil grain rape (Brassica napus L.).
Field Crop Res. 80: 35- 49.
6- Farshadfar, A. A. 1997. Breeding methodology. Razi University Press. 250 pp.
(In Farsi).
7- Farsi, M and A. Bagheri. 1998. Principle of Plant Breeding. Mashhad University
Press. 368 pp. (In Farsi).
8- Fayaz. F and R. Talebi. 2009. Determining relationships among yield and some
yield components using path analysis of field peas. Iranian J. Field Crops Res.
7(1): 137-143. (In Farsi).
9- Garcia Del-Moral, L. f., J. M. Ramos and M. P. Jimenez .1999. Ontogenetic
approach to grain production in spring barley based on path- coefficient analysis.
Crop Sci. 31: 1179-1185.
10- Hakan- Ozer, E. O and D. Unsal. 1999. Relationship between yield and yield
components on currently improved spring rape grain cultivars. J. Agric. Forestry.
23: 603- 607.
Scientific Journal of Agronomy and Plant Breeding, 1(1): 51-58, January 2013 57
Seyedmohammadi, Evaluate Correlation and Path Coefficient Analysis for Agronomic … 52
11- Jafari, A. 2001. Genetic distance of 29 permanent ryegrass genotypes by cluster
analysis based on yield and morphological traits. Genetic modification of
plants for forestry research. Research Institute of Forests and Range Publications.
Tehran. 6: 79-91. (In Farsi).
12- Jorgeh, A. 2003. Determine appropriate planting date on rapeseed cultivars and
promising to study the correlation between yield components. MS Thesis. Faculty
of Agriculture. Islamic Azad University of Dezful. 152 pp. (In Farsi).
13- Matlabipor, Sh., M. Ahmadi and L. Jokar. 2000. Performance of adaptation
and yield traits of rapegrain cultivars in Zarqan (Fars). Iranian J. Crop Sci.
2(3): 59-70. (In Farsi).
14- Mirmosavi, A., H. Zynali and A. Hosinzadeh. 2004. Genetic correlations with
grain yield and oil content of some important agronomic traits in rape grain by
regression and path analysis. Proceedings of the International Crop Science
Congress of Iran. Gilan University. 312 pp. (In Farsi).
15- Mendham, N. J. S and R. K. Scot. 1975. The limiting effect of plant size at inflorescence
initiation on subsequent growth and yield of oil grain rape
(B.napus). J. Agric. Sci. Cambridge. 84: 487–502.
16- Qulipor, A., N. Latifi, K. Qasemi and M. Moqadam. 2004. Comparative
growth and grain yield of rapeseed under dry land conditions in Gorgan. J. Agric.
Sci. Natural Res. 11(1): 5-14. (In Farsi).
17- Rahnama, A. A and A. Bakhshandeh. 2005. Effect of sowing dates and direct
graining and transplanting methods on agronomic characteristics and grain
yield of rapeseed under Ahvaz conditions. Iranian Journal of Crop Sciences. 7
(4): 324-336. (In Farsi).
18- Rabiee, M., M. M. Karimi and F. Safa. 2004. Effect of planting date on grain
yield and agronomic traits of rape grain cultivars as a second crop after rice in
kochesfahan rejoin. Iranian J. Agric. Sci. 35(1): 177-187. (In Farsi).
19- Rahnamaee tak, A., Sh. Vaeezi, J. Masafari and A. A. Boshehri. 2007. Correlation
and causality analysis and yield-related traits in red beans. Iranian J. Res.
Development (Agron. J.). 80(2): 95-102. (In Farsi).
20- Solymanzadeh, H., N. Latifi and A. Soltani. 2007. Phenological and physiological
traits associated with grain yield of rapeseed cultivars under rain fed
conditions. J. Agric. Sci. Natural Res. 14(5): 67-76. (In Farsi).
21- Shirany rad, A. H. 1994. Effect of planting date and plant density on growth
and agronomic traits of rapegrain cultivars. MS Thesis. Faculty of Agriculture,
Tarbiat Modarres University. (In Farsi).
22- Vaeezi, Sh., S. Abdmyshani, B. Yazdisamadi and M. R. Qanadha. 2000. Correlation
and path analysis and yield-related traits in maize. J. Agric. Sci. 31(2):
82-71. (In Farsi).
23- Valadyani, A., M. Tajbakhsh and M. R. Zardoshti. 2004. Evaluation of agronomic
characteristics and yield of rapeseed cultivars in Urmia region. J. Agric.
Sci. 14(2): 31-43. (In Farsi).
24- Zemmermann, M. S., J. G. Warnesand and K. W. Foster. 1994. A heritability
and correlation study of grain yield, yield components, and harvest index of
common bean in sole crop and intercrop. Field crop Res. 9(2): 109-118.
58
