Yield dynamics of soybean relative to plant population

MISSING IMAGE

Material Information

Title:
Yield dynamics of soybean relative to plant population
Physical Description:
xv, 155 leaves : ill. ; 29 cm.
Language:
English
Creator:
Thompson, Peter G., 1959-
Publication Date:

Subjects

Genre:
bibliography   ( marcgt )
theses   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis (Ph. D.)--University of Florida, 1990.
Bibliography:
Includes bibliographical references (leaves 145-154).
Statement of Responsibility:
by Peter G. Thompson.
General Note:
Typescript.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 001613901
notis - AHN8319
oclc - 23536042
System ID:
AA00003761:00001

Full Text










YIELD DYNAMICS OF SOYBEAN RELATIVE TO PLANT POPULATION


By
PETER G. THOMPSON












A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA


1990













TABLE OF CONTENTS

page
ACKNOWLEDEGMENTS ................................................................................... v
LIST OF TABLES................................................................................................... vii
LIST OF FIGURES................................................................................................. xi
ABSTRACT...................................................................................................... xvi
CHAPTERS
1 INTRODUCTION................................................................................ 1

2 LITERATURE REVIEW......................................................................... 4

The Effect of Plant Population on Soybean................................... 5
Stems and Branches.............. .......... .............................. 5
Leaves and Leaf Area Index................................... ......... 6
Flower Production and Development ................................... 7
Nodes, Pods and Seeds.................................................. 7
Biomass....................................................................................... 10
Lodging................................................................................. 10
Nitrogen and Phosphorus Accumulation............................. 10
The Structural and Yield Implications of
Density-Mediated Plastic Responses............................ 11

Explaining Soybean Yield in Plant Population Studies................. 12
Regression Analyses and Correlations of
PYCs and Yield.............................................................. 12
Plant Population Density Effect on Productivity on a
per Plant and per unit Area Basis. ................................ 13
The Yield-Plant Population Density Response.................... 19

3 MATERIALS AND METHODS..................................................... 27

Field Experiments ............................................................................. 27
Data Analysis.......................................................................................... 32
The Evaluation of Variables and Relationships
Operating in Yield-Plant Density Studies...................... 35






ii










4 RESULTS AND DISCUSSION-1987 EXPERIMENT ..................... 41
ANOVA Results..................................................... .......................... 41
Plant Population Effect ............................................. ......... 41
Cultivar Effect......................................................... ................... 46
An Explanation for Cultivar Differences
Based on ANOVA.................................... .... ........... 52
Correlation Analysis........................................................................ 56
An Explanation for Cultivar Yield Differences
Based on Correlation Analyses........................... ........ 59
The Yield Plant Population Density Response.................... 61
Conclusion........................................................................................... 65

5 RESULTS AND DISCUSSION-1988 EXPERIMENT...................... 67
Plant Population Density Effects, Yield Correlations and their
Implications for Yield.............................................................. 67
Vegetative Structures...................... .................................. 69
Nutrient Status and Linear Regression ................................ 84
Reproductive Structures.......................................................... 94
Results and Discussions of Cross Tables ............................... 102

6 AN INVESTIGATION INTO THE YIELD-PLANT DENSITY
RESPONSE WITH SOYGRO..................................................... 114
A Crop Simulation Perspective on Duncan's Assumptions ............ 116
Simulation Exercise .......................................................................... 118
Results of Model Calibration .................................................................. 123
An Investigation of Duncan's Assumptions....................................... 128
Total Light Use Efficiency--Glucose Equivalent...................... 135
Seed Yield Efficiency and Seed Yield Efficiency
-Glucose Equivalent............................................................ 135
Conclusions.............................................................................................. 140
7 SUMMARY AND CONCLUSIONS....................................................... 142








REFERENCES.............................................................................................. 145
BIO G RAPH ICAL SKETC H................................................................................. 155














ACKNOWLEDGMENTS


I am grateful to the Institute of Food and Agricultural Sciences for
providing funding; without which my studies at the University of Florida would not
have been be possible. I would also like to acknowledge the Graduate School;
specifically the Minority Program, for their generous financial support.
I would like to thank Dr. J. L. Fry who introduced me to the University of
Florida, for his constant encouragement and unwavering commitment to the
success of my program His concern, advice and thoughts also helped greatly
in resolving my frustrations as a student.
I am grateful to Dr. R. McDavis who always provided invaluable and
timely advice. His philosophy on education was always encouraging, especially
in difficult times.
I am deeply appreciative of Dr. Darell E. McCloud, whom as the chairman
of my committee was particularly understanding and wholeheartedly supported
my quest for a well-rounded education.
I am indebted to Dr. Ken J. Boote, the cochair of my committee, who
contributed significantly to my intellectual maturity by persistently placing before
me the challenge of considering academic and scientific issues that I would not
have otherwise addressed.











I would like to express my gratitude to Dr. Kuell Hinson. It has always
been a pleasure to interact with him. He has always been concerned about my
welfare as a student and his open-mindedness is reassuring.
I am indebted to Dr. Chris. O. Andrew--a mentor. Our discussions and his
sound advice always provided valuable insight. But most importantly, he has
inspired me to take a broader view of life.
I am grateful to Dr. William G. Blue, with whom I have an extremely
comfortable relationship, and whose openess and warmth never diminished his
rigorous and spirited examination of my work.
I am thankful to Dr. Goran Hyden who patiently instilled in me the
confidence to pursue issues pertinent to agricultural development. He
contributed greatly to my intellectual development by always challenging me to
take my ideas one step further.
I would like to thank Rick Hill who introduced me to the operations of the
Agronomy Farm which facilitated my field experiments. I am grateful to Frank
McGraw for his willing and invaluable practical advice on my field experiment.












LIST OF TABLES


Table aga
3-1 Summary of the plant variables measured during
1987 and 1988.................................................................. 33
3-2 A table for the interpretation of plant density-yield cross-
referenced data.................................................................... 39
4-1 The effect of plantpopulation and cultivar on total dry
weight (g m ) at R5, 1987.................................. .......... 42
4-2 Cultivar effect on total dry weight at R5,1987. ............................. 42
4-3 The effect of plant population and cultivar on soybean
plant traits at R6 stage, 1987................................................... 43
4-4 Cultivar effect on soybean plant traits at R6 stage, 1987.............. 44
4-5 The effect of plant population and cultivar on plant
traits at R8 stage, 1987.................................... .......... 45
4-6 The effect of plant population (X) on pod number (m-2)()
for each cultivar at R8 stage, 1987......................................... 53
4-7 The effect of plant population on seed number (m-2)
for each cultivar at R8 stage, 1987.......................................... 53
4-8 The effect of cultivar on pod number (m-2) for each
plant population at R8 stage, 1987........................................ 54
4-9 The effect of cultivar on seed number (m-2) for each
plant population at R8 stage,1987....................................... 54
4-10 Cultivar effect on soybean plant traits at R8 stage, 1987.............. 55
4-11 Correlation coefficients for yield and its fundamental
components (X1) at R6 with variables (X2)at
R5 and R6, 1987................................................................ 57








4-12 Correlation coefficients for the fundamental yield components
(X1) at stage R8 with variables (X2) at R5 and
R8, 1987........................................................................................ 58
4-13 Mean values and correlation coefficients for R5 total dry
weight (g m- ), seed dry weight (g m-2) and its
fundamental components at R8, 1987................................ 60
4-14 Statistical models for per plant responses (Y) of soybean
with plant density (X).............. .............................................. 60
5-1 Statistical models for soybean plant variables (Y)
with plant density (X) at R3 stage, 1988................................ 70
5-2 Statistical models for soybean plant variables (Y)
with plant density (X) at R5 stage, 1988.................................... 70
5-3 Correlation coefficients of soybean plant variables at
R3 stage with seed dry weight (g m-4) at R8, 1988.............. 74
5-4 Correlation coefficients of soybean plant variables at
R5 stage with seed dry weight (g m- ) at R8, 1988............. 74
5-5 Statistical models for soybean plant variables (Y)
with plant density (X) at R3 stage,1988....................................... 77
5-6 Statistical models for soybean plant variables (Y)
with plant density (X) at R5 stage, 1988.................................. 77
5-7 Correlation coefficients of soybean plant varjibles
at R3 stage with seed dry weight (g m") at R8,1988................. 82
5-8 Correlation coefficients of soybean plant variables
at R8 stage with seed dry weight (g m-2) at R8,1988......... 82
5-9 Statistical models for soybean plant variables (Y) at R5
stage with plant density (X), 1988..................... 86
5-10 Correlation coefficients of soybean plant variables
at R5 stage with seed weight (g m" ), 1988........................ 86
5-11 Statistical models for soybean plant variables (Y)
at R5 stage with plant density (X), 1988.............................. 93
5-12 Statistical models of plant density (X) with soybean
yield components (Y) at R8, 1988............................................. 93
5-13 Correlation coefficients of soybean plant vriables at
R5 stage with seed dry weight (g m-4) at R8,1988............ 101








5-14 Correlation coefficients of soybean plant variables at
R8 with seed dry weight (g m) at R8,1988......................... 101
5-15 Yield-Plant Density Cross-Table for soybean plant
variables highly associated with yield and
exhibiting high plant density effect at R8, 1988.................... 103
5-16 Yield-Plant Density Cross-Table for soybean plant
variables highly associated with yield and
exhibiting no plant density effect at R8, 1988..................... 105
5-17 Yield-Plant Density Cross-Table for soybean plant
variables not associated with yield and
exhibiting no plant density effect at R8, 1988............................ 106
5-18 Yield-Plant Density Cross-Table for R5 soybean plant
variables exhibiting plant density effect and high
yield association, 1988.................................................. 107
5-19 Yield-Plant Density Cross-Table for soybean plant
variables not associated with yield and not
affected by plant density at R5 and R8,1988................................ 108
5-20 Yield-Plant Density Cross-Table for soybean plant
variables not associated with yield and
exhibiting no plant density effect at R5,
and which show high association to
each other, 1988...................... ................................................. 110
5-21 Yield-Plant Density Cross-Table for soybean plant
variables highly associated with yield and
exhibiting high plant density effect at R3, 1988......... 111
5-22 Plastic and non-plastic responses of reproductive,
remobilization and photosynthetic parameters
and their impact on seed dry weight (g m-2), 1988............... 112
6-1 Yield data for Cumberland cultivar,1986. .................................... 120
6-2 Irrigation schedule for Cumberland cultivar,1986........................ 120
6-3 Definitions of soil parameters and characteristics
of M aury silt loam ........................................................................ 121
6-4 Summary of changes in genetic traits of Williams cultivar
and crop parameters in calibrating SOYGRO using
Cumberland cultivar,1 986................................. 124
6-5 Simulated yield data for Cumberland cultivar, 1986..................... 125








6-6 Total light use efficiencies as determined by SOYGRO
for Cumberland cultivar, 1986, at seventeen
plant densities............................................................................... 136
6-7 Seed yield efficiencies as determined by SOYGRO for
Cumberland cultivar,1986, at seventeen plant
densities..................................................................................... 138









LIST OF FIGURES


Figure page
2-1 Typical response for yield per unit area (a) and yield
per plant (b) to plant density................ ..................................... 21
3-1 Precipitation profile over the 1987 and 1988
growing seasons, Gainesville, FI................................................. 28
4-1 The effect of plant population density on total dry weight
at the R5 stage in soybean, 1987............................................. 47
4-2 The effect of plant population density on seed dry weight
of soybean at maturity, 1987............................................. ...... 48
4-3 The effect of plant population density on seeds pod-1
of soybean at maturity, 1987......................................................... 49
4-4 The effect of plant population density on weight seed-1
of soybean at maturity, 1987.............................................. ...... 50
4-5 The effect of plant population density on pod number m-2
of soybean at maturity, 1987...................... .................... ... 51
4-6 The effect of plant population density on dry weight per
plant of soybean at the R5 stage, 1987..................................... 62
4-7 The effect of plant population density on seed and pod
weight per plant of soybean at maturity, 1987......................... 63
4-8 The effect of plant population density on seed and pod
number per plant of soybean at maturity, 1987.................... 64
5-1 The effect of plant population density on seed and pod
weight of Kirby at maturity, 1988........... ..................... 68
5-2 The effect of plant population density on stem, leaf and total
dry weights of Kirby at the R3 stage, 1988.............................. 71
5-3 The effect of plant population density on stem, leaf and total
dry weights of Kirby at the R5 stage, 1988.................................. 72
5-4 The effect of plant population density on vegetative
dry weight of Kirby at the R3 and R5 stages,1988............... 73









5-5 The effect of plant population density on stem and leaf
fractions of Kirby at the R3 and R5 stages,1 988....................... 76
5-6 The effect of plant population density on specific leaf
area of Kirby at the R3 and R5 stages,1 988.............................. 79
5-7 The effect of plant population density on the leaf area
index of Kirby at the R3 and R5 stages, 1988............................ 80
5-8 The effect of plant population density on percent light
interception of Kirby at the R3 and R5 stages,1988.................. 81
5-9 The effect of plant population density on leaf and stem N
concentrations of Kirby at the R5 stage,1988............................ 85
5-10 The effect of plant population density on leaf and stem
N content of Kirby at the R5 stage,1 988.................................. 87
5-11 The effect of plant population density on leaf and stem P
concentrations of Kirby at the R5 stage,1988............................... 88
5-12 The effect of plant population density on leaf and stem
P content of Kirby at the R5 stage,1988.................................. 89
5-13 The effect of plant population density on N content per
unit leaf area of Kirby at the R5 stage, 1988................................. 90
5-14 The effect of plant population density on P content per
unit leaf area of Kirby at the R5 stage, 1988.............................. 91
5-15 The effect of plant population density on seed and pod
number of Kirby at maturity, 1988.................................. 95
5-16 The effect of plant population density on weight per seed
of Kirby at maturity, 1988.......................................................... 96
5-17 The effect of plant population density on seed number
per pod of Kirby at maturity, 1988..................................... 97
5-18 The effect of plant population density on seed and pod
weight per plant of Kirby at maturity, 1988................................ 98
5-19 The effect of plant population density on seed and pod
number per plant of Kirby at maturity, 1988.............................. 100
6-1 Observed and simulated effects of plant population
density on seed yield of Cumberland,
cultivar, 1986..................................... 126










6-2 Observed and simulated effects of plant population
density on weight per seed of Cumberland
cultivar,1986....................................................................................... 127
6-3 Observed and simulated effects of plant population
density on seed number per unit area of
Cumberland cultivar,1986.............................................................. 129
6-4 Observed and simulated effects of plant population
density on seed yield per plant of
Cumberland cultivar,1986.............................................................. 130
6-5 The relationship between light intercepted between R5 and
R7, and percent light absorbed at noon (R5), daily (R5)
and between R5 and R7 for Cumberland,1986....................... 131
6-6 The simulated effect of plant population density on N
mobilized per unit area for Cumberland cultivar,1986............. 133
6-7 The simulated effect of plant population density on percent
mobilized-N in seeds of Cumberland cultivar, 1986................... 134
6-8 The simulated effect of plant population density on total
light use efficiency of Cumberland cultivar,1986..................... 137
6-9 The simulated effects of plant population density on seed
yield efficiency of Cumberland cultivar,1986 with and
without the benefit of protein remobilization................................ 139













Abstract of Dissertation Presented to the Graduate School of the
University of Florida in Partial Fulfillment of the Requirements for the
Degree of Doctor of Philosophy

YIELD DYNAMICS OF SOYBEAN RELATIVE TO PLANT POPULATION

By

PETER G. THOMPSON

August 1990

Chairman: Dr. Darell E. McCloud
Cochairman: Dr. Kenneth J. Boote
Major Department: Agronomy


Soybean [Glycine max (I.) Merr.] yield has been shown to respond to
changes in plant population. Results from numerous plant population studies
on soybean conclusively show that various crop and plant parameters which
are known to influence yield, are themselves subject to plant population effects.
However, the mechanism by which plant population exerts its influence on yield
is not well understood.
Field experiments were conducted in Gainesville, Florida (29038N)
during 1987 and 1988 to identify crop parameters and physiological and
ecological relationships which mediate the effect of plant population on yield.
SOYGRO--a process-oriented crop growth simulator, was also used to
investigate the effect of plant population on seed yield and to test published
hypotheses regarding the efficiency with which the soybean crop uses
photosynthetically active radiation (PAR) to produce seed yield.







The soybean cultivars Kirby [MG VIII], Centennial [MG VI] and Forrest
[MG V], were planted in 1987 at populations of 11.1, 25.0 and 44.4 plants m-2 in
a square arrangement. The parameters recorded at growth stages R5, R6 and
R8 included total dry weight, pod dry weight and number, and seed dry weight
and number. Plant population did not have any effect on the variables
recorded; however, all variables were significantly affected by cultivar. Seed
size and total dry weight m-2 at R5 accounted for the differences observed in
yield between the cultivars.
In 1987, Kirby [MG VIII] was planted at populations of 2.0, 4.0, 6.3, 11.1,
16.0, 25.0, 44.4 and 69.4 plants m-2 in a square arrangement. The variables
most sensitive to plant population, and which were also closely associated with
yield, were seed number and pod number at R8 and total dry weight and leaf
dry weight at R5.
The SOYGRO model which was calibrated to give responses
approximating those of observed mean yield and canopy traits to plant density,
confirmed the hypothesis that seed yield efficiency (seed yield / unit PAR
intercepted) tends to increase with plant density. This increase can be
attributed in part to an increasing ratio of mobilized-N to total-N in seeds at
higher plant densities.











CHAPTER 1
INTRODUCTION


Soybean [Glycine max (L.) Merr.], the prodigious legume which was first
domesticated approximately 3,000 years ago by the Chinese has come to be
known as the "yellow jewel" and the "wonder bean." Soybean has a tremendous
range of uses. It is used in the production of such varied products as cardboard,
paint, fire extinguisher foam and glue. However, because of its high protein and
oil content, soybean finds its most important use either directly or indirectly as
food for both humans and animals. Soybean is viewed by some as a major
weapon against world hunger. But if the full potential of soybean as a weapon
against hunger is to be realized, the processes which determine its grain yield
need to be better understood.
Soybean is grown under diverse cultural conditions and usually at plant
populations which, through trial and error, have been found to be desirable.
The study of the effect of plant population or plant population density on soybean
grain yield is of interest for two main reasons. First, for a given cropping system,
the quantitative relationship between yield and plant population density is
important in estimating i) at what density maximum or optimum yield can be
achieved and ii) in economic terms, which plant population will provide the most
profitable yield.







Secondly, plant population studies allow examining the process of yield
achievement and its determinants. The determinants of yield or potential yield
correlates (PYCs) can be defined as any morphological or physiological
characteristic of a crop, which may influence final seed yield at one time or
another during the development of the crop. An understanding of the relationship
between PYCs and yield is basic to any appreciation of the process of yield
achievement. The elucidation of the nature and the extent of the functional
relationship between PYCs and yield is facilitated by plant population studies. A
sufficiently wide range of plant populations may provide a range in magnitude of
both PYCs and yield, and this allows comprehensive regression analyses to be
conducted on the relationship of yield to PYCs.
Therefore plant population studies are important in understanding both
the physiological basis of yield and crop improvement. By establishing the
relationships i) of both PYCs and yield with plant population density, ii) between
PYCs themselves and iii) between PYCs and yield, a basis and direction for
physiological inquiry into the process of yield achievement can be attained.
Once the major determinants or PYCs of yield have been established, these
factors may also serve as effective selection criteria in crop improvement
programs which seek to achieve higher yields, and foster socioeconomic
development.
It has been pointed out by Duncan (1986) that there has yet to emerge a
theory capable of explaining soybean yield within the context of plant population
studies. In an attempt to appreciate the dynamics of yield within the context of
plant population studies, the objectives of the investigation conducted here are;
first, to develop an analytical framework for the analysis of yield-plant population




3


density data; secondly, to ascertain which biological and ecological variables
and relationships of the soybean crop mediate the influence of plant population
density on final grain yield; and thirdly, to elucidate the basis for the response
surface which characterizes yield-plant population density relationships. This
final objective which focuses on a relatively unexplored area (the response
surface) of plant population studies, will be carried out using SOYGRO a
process oriented crop growth simulation model for soybean.











CHAPTER 2
LITERATURE REVIEW


The research literature on yield-plant population density studies will be
broken down into three sections. First, the effect of plant population on the
soybean crop will be discussed. The next two sections come under the umbrella
of explaining yield, and consist of discussions on the observation that yield per
unit area increases and yield per plant decreases with increasing plant
population density, and discussions on the asymptotic or parabolic shape of
yield-plant population density responses. It is within the context of the latter
section that SOYGRO--a process oriented plant growth simulator (Jones et al.,
1989) will be discussed. Of specific interest with SOYGRO are estimations of the
impact of plant population density on the efficiency with which plants use light
inputs for seed production. Lastly, the identification of the reproductive growth
stages of soybean and the square planting pattern which are relevant to
experimental procedures will be discussed.


The Effect of Plant Population on Soybean


In plant population or yield-plant population density studies, the main
ecological force which influences crop performance is interplant competition.
With increasing density, plants tend to compete more intensely for available







resources. Plant competition is subject to a number of differing interpretations;
however, here it will be defined according to Grime (1973)--the tendency for
neighboring plants to utilize the same quantum of light, ion of a mineral, molecule
of water or volume of space.
Competitive interactions between organisms are considered to be among
the most important ecological phenomena (Snell and Burch, 1975). Competition
between organisms occupies a place of prominence in evolutionary theory,
where "survival of the fittest" determines the procreation possibilities of species.
At a more immediate level, competition holds various consequences for a
developing plant. In reacting or adapting to changes in the pressures of
competition, plants may become modified both physiologically and
morphologically. When this occurs, the plant is said to exhibit plastic responses
(Hutchings and Budd, 1981). Plastic responses may affect the productive
performance of plants and crops. It has been amply demonstrated by
researchers, that in soybean, both yield and PYCs are subject to
density-mediated plastic responses. The impact of plant population density on i)
vegetative and reproductive structures, ii) lodging and iii) nitrogen (N) and
phosphorus (P) accumulation will be discussed.


Stems and Branches


The height and thickness of soybean stems have been shown to be
affected by plant population density. It has been reported that plant height
increases with plant population density (Ramseur et al., 1985; Doss and Thurlow,
1974; Wilcox, 1974; Johnson and Harris, 1967). This increase in plant height has






been attributed to an increase in the length of the internodes, and not to
increases in the number of nodes (Dominguez and Hume, 1978; Basnet et al.,
1974; Hinson and Hansen, 1962). In some cases, plant height has been
unaffected by density (Hoggard et al., 1978; Lueschen and Hicks, 1977; Probst,
1945). In one atypical case, Dominguez and Hume (1978) reported that with the
cultivar 'Vansoy', increased density resulted in shorter plants. The diameter of
the main stem in soybean decreases with increased density (Miura and Gemma,
1986; Ramseur et al., 1985; Wright et al., 1984; Fontes and Ohlrogge, 1972).
Oba et al. (1961) also found that stem weight per plant decreases with density.
The number of branches per plant decreases with density (Basnet et al.,
1974; Fontes and Ohlrogge, 1972; Hinson and Hansen, 1962; Lehman and
Lambert, 1960). However, in their work Chaudhry and Cheema (1985) found
that the effect of density on the number of branches per plant was not significant.


Leaves and Leaf Area Index


Leaf weight per plant decreases (Oba et al., 1961) while the leaf area
index (LAI) increases with plant population density (Parvez et al., 1989; Weber et
al., 1966). Of relevance to population studies is the fact that the degree of
shading or exposure to light affects leaf anatomy. At lower light intensities leaves
develop fewer layers of palisade cells (Xu and Miao, 1988), while higher light
intensities have the opposite effect, resulting in thicker leaves (Chabot et al.,
1979; Fails et al., 1982). At higher population densities where the shading of
lower leaves occurs, it can be expected that there will be greater differences







between the specific leaf weight of leaves on the periphery of the canopy and
those deeper in the canopy itself.


Flower Production and Development




Increased population results in a reduction in the proportion of flowers
which result in mature pods (Buttery, 1969). Dominguez and Hume (1978) also
found that increased density increased the percentage of flowers aborted. They
further observed that the decrease in the number of flowers per plant observed at
high densities, was due to fewer flowers produced per node. This observation
has also been made by Lehman and Lambert (1960). Fewer flowers at higher
densities has also been attributed to fewer nodes per plant (Dominguez and
Hume, 1978).


Nodes. Pods and Seeds


The number of nodes per plant decreases with plant population density
(Dominguez and Hume, 1978; Basnet et al., 1974). However, on a per unit area
basis, the number of nodes increases with density (Parvez et al., 1989; Hruska
and Labounek, 1985). Consistent with these observations are those that the
number of pods per plant decreased (Olsen, 1986; Nakagawa et al., 1986;
Hoggard et al., 1978; Lueschen and Hicks, 1977; Basnet et al., 1974; Lehman
and Lambert, 1960), while the number of pods per unit area increased (Udoguchi
and McCloud, 1986) with plant population density. In their work, Chaudhry and







Cheema (1985) reported that the number of pods per plant was not affected by
plant population density. Pod set is also greater at lower populations (Oba et al.,
1961; Weber et al., 1966).
Seed weight per unit area increases with plant population density (Parvez
et al., 1989; Udoguchi and McCloud, 1986; Parker et al., 1981; Spilde et al.,
1980; Doss and Thurlow 1974). However, in their work, Hoggard et al. (1978)
found that seed weight per unit area decreased with increasing plant population
density. Basnet et al. (1974) and Caviness (1966) reported that seed weight per
unit area was unaffected by plant population density.
Johnson and Harris (1967) in their work, found that of the several cultivars
investigated, weight per seed decreased with plant population density only with
'Lee'. This decrease has also been reported by Parks and Manning (1980),
Wilcox (1974), and Wright et al. (1987). In thinning experiments, Weil and
Ohlrogge (1976) also reported that weight per seed increased by 11.5% with
decreased plant population density. Nedic et al. (1987) also found that the
number of filled seeds per plant decreased with density. The reduction in seed
size with increasing plant population density may be the result of increased pod
production per unit area. Other researchers reported no relationship between
weight per seed and plant population density (Dominguez and Hume, 1978;
Fontes and Ohlrogge, 1972).
The number of seeds per plant decreases with density (Lueschen and
Hicks, 1977; Fontes and Ohlrogge, 1972). Blumenthal et al. (1988) observed
that the number of seeded pods per branch decreased with population. The
number of seeds per pod decreases with plant population density (Blumenthal et
al., 1988; Herbert and Litchfield, 1982; Fontes and Ohlrogge, 1972; Lehman and







Lambert, 1960). However, in some cases it has been shown that seeds per pod
is not affected by plant population density (Chaudhry and Cheema, 1985;
Dominguez and Hume, 1978).
The distribution of pods and hence seeds within the crop and on plants is
also affected by plant population density. The height of the first pod increases
with density (Nedic et al., 1987; Olsen, 1986; Tsuchiya et al., 1986; Lueschen and
Hicks, 1977; Basnet et al., 1974). This fact is of importance in mechanical
harvesting, where it is advantageous to have the pods located away from the
ground. Parks and Manning (1980) found that the ratio of the number of seeds on
the main stem to those on the branches increased with plant population density.
Ramseur et al. (1984) also found that with decreased plant population density,
the seed number and yield on branches increased while pod number, seed
number and yield on stems decreased. Dominguez and Hume (1978) in their
work, observed that for the determinate cultivar 'Fiskeby V', an increase in plant
population density resulted in a shift in the majority of yield from the upper third to
the middle of the plant. Ramseur et al. (1984) also found that increased plant
population density reduced the contribution of the lower nodes to yield.
According to Weil and Ohlrogge (1976) thinning resulted in weight per seed
being 23.9% greater at the top of the canopy than in unthinned stands. Increased
density resulted in a reduction of the bottom portion of plants in their contribution
to total seed number (Dominguez and Hume, 1978).










It has been reported that biomass per unit area increases with plant
population density (Parvez et al., 1989; Blumenthal et al., 1988; Udoguchi and
McCloud, 1986). Biomass per plant decreases with plant population density
(Holler and Abrahamson, 1977; Wilcox, 1974). It should be noted that biomass
per unit area tends to increase with plant population density at all stages of plant
development. However, biomass per plant will decrease with plant population
density only when there is appreciable interplant competition.


Lodging


Lodging increases with plant population density (Tsuchiya et al., 1986;
Fontes and Ohlrogge, 1972; Probst, 1945). Lueschen and Hicks (1977) found a
significant linear relationship between plant population and lodging. Johnson
and Harris (1967) in studying a number of soybean cultivars reported that
cultivars with larger stem diameters, such as 'Hardee', tended not to lodge in
response to increased plant population density. Increased lodging was also
associated with taller plants (Cooper, 1971). Increased density increased plant
height and decreased stem diameter (Wright et al., 1984).


Nitrogen and Phosphorus Accumulation


According to Abrahamson and Caswell (1982), biomass allocation in
plants has been correlated to various degrees to the allocation of minerals and







energy. Hanway and Weber(1971) from their work, also showed that N, P and K
accumulation mirrored biomass allocation. Nelson and Weaver (1980) reported
that the lowest planting density exhibited the highest rate of acetylene reducing
activity per plant. This finding has been attributed to the greater nodule mass on
plants at the lowest plant population density, since the specific activity of nodules
was only slightly affected by plant population density. Buttery (1969) reported that
N concentration decreased with density, and he attributed this to increased
competition for N between plants at higher plant population densities.
The nutrient status of plants is directly related to crop productivity in two
ways. First, the photosynthetic capacity of plant tissue depends upon the
concentration of nutrients, especially in the leaves. Secondly, the status of N
and P also determines the amount of nutrient which may be available for seed
production via the process of nutrient remobilization.


The Structural and Yield Implications of Density-Mediated Plastic Responses


Plastic responses, which are embodied in morphological and
physiological changes, hold both structural and yield implications for soybean.
Structurally, plants are exposed to various physical stresses. Plants which
undergo density-mediated plastic responses must compensate for changing
stresses to avoid the mechanical failure of their axes (that is, breakage and
lodging). Accordingly, plants preferentially form and lay down specialized tissues
of varying material properties, in different regions of the plant (Niklas and
O'Rouke, 1982). Pianka (1972) suggested that in organisms with "plastic







abilities," it can be expected that the manner and pattern of resource allocation
will be influenced by different competitive environments.
The process of adjusting to the forces of interplant competition, therefore,
influences canopy arrangement, light interception, stem thickness, pod number
and distribution, mineral concentrations and contents, the amount of nutrients
which can be remobilized, the number of seeds and so forth. The entire process
whereby light energy is converted into chemical energy and is allocated and
stored, is affected by plant population density.


Explaining Soybean Yield in Plant Population Studies


In attempting to explain soybean yield within the context of plant
population studies, three areas of overlapping research can be identified: i) the
correlation or regression of PYCs with yield, ii) the exploration of the
physiological basis for the observation that yield per unit area increases, while
yield per plant decreases with increasing plant population density, and iii) the
examination of the physiological basis of asymptotic and parabolic yield-plant
population density responses.


Regression Analyses and Correlation of PYCs and Yield


The correlation or association of various plant and crop characteristics
with seed yield is standard practice in yield improvement programs and in yield
analysis. Outside the context of plant population density studies (using different
soybean genotypes and productive environments), the following characteristics







are correlated with yield per plant in soybean: stem weight, stem thickness and
branch number (Taguchi et al., 1958), pod number and the number of fertile
nodes (Board, 1985). Yield per unit area has been shown to be highly
associated with LAI and number of seeds per unit area, but only moderately
correlated with dry matter production (Basuchaudhuri, 1987). In their work, Kaw
and Menon (1972) found that pod number and seed number per unit area were
highly correlated with yield per unit area. Characteristics showing the lowest
correlation with yield per plant include seeds per pod and weight per seed
(Board, 1985; Tandojam,1986; Johnson et al., 1955). Hartwig and Edwards
(1970) in their studies on morphological characters which affect yield per unit
area, reported that seed number per pod did not affect yield. However, Ma
(1946) (cited by Anand and Torrie (1963)) found that seed number per pod was
correlated with yield, and Weber and Moorthy (1952) reported that there was a
high positive correlation between weight per seed and yield per unit area.
Within population studies the following characteristics are highly
correlated with yield per plant: stem diameter and pod number per plant (Fontes
and Ohlrogge, 1972). The following were correlated with yield per unit area: pod
number and total dry matter (Parvez et al., 1989; Udoguchi and McCloud, 1986).
Leaf area index was not correlated with yield per unit area (Ramseur et al., 1985;
Costa et al., 1980).


Plant Population Density Effect on Productivity on a per Plant and per Unit Area
Basis


Light interception and canopy structure form the main basis of discussions
which seek to explain the impact of plant population density on yield per plant







and yield per unit area. To a lesser extent, the significance of N and P status has
also been addressed. Writing in 1966, Weber et al. argued that in order to
understand how plants respond to changes in density, it is necessary to
investigate the 'spatial relationships' in canopies, solar energy interception and
leaf area accumulation. The amount of light intercepted by a crop is determined
by the arrangement of the canopy (branching, leaf area, etc.)
Lueschen and Hicks (1977) reported that soybean plants are capable of
compensating for low plant population densities by producing more branches
and pods per plant. Acock and Acock (1987) found that the growth and
development of branches on a plant depended on the amount of light received.
Shading reduced the photosynthetic rate of plants at high densities, resulting in
carbon shortages (per plant) and a tendency to produce fewer branches. These
findings are consistent with those of Charles-Edwards (1984), Charles-Edwards
and Beech (1984), Bunce (1988) and Johnston et al. (1969).
Yield per unit area has been positively correlated with vegetative dry
matter production per unit area (Egli et al., 1987). It has also been established
that dry matter production is a function of solar radiation intercepted (Shibles and
Weber, 1966; Brougham, 1956; Davidson, 1958; Watson, 1958). Brougham
(1956) found that both the rate of dry matter production per unit area and the
percent solar radiation intercepted increased with LAI. Johnston and Pendleton
(1968) argued that due to its large leaves and phyllotaxy, soybeans tend to have
a closed canopy. Soybeans intercepted as much as 90% of the incoming solar
radiation near the top and periphery of the canopy (Sakamoto and Shaw, 1967).
Thus there is a tendency for lower leaves to be shaded. If increased plant
population density is accompanied by increased shading, it can be expected that






the light intercepted per plant will decrease which will subsequently result in a
reduction in yield per plant.
Leaf area index increases with plant population density (Parvez et al.,
1989; Weber et al., 1966; Costa et al., 1980). The extent of photosynthetic activity
not only depends on LAI and the amount of light intercepted, but also upon the
photosynthetic capacity of the leaves intercepting the light. Bunce (1988) found
that photosynthetic capacity was lower at higher densities, even though LAIs are
higher. Reduced photosynthetic capacity resulted from the shading of lower
leaves, and the longer period of time that young developing leaves spent in the
shade at high stand densities. These findings corroborate the work of
Nichiporovich et al. (1969) and the suggestion by Johnson et al. (1969) that due
to both inter- and intra-plant competition for light, middle and bottom soybean
leaves do not reach their full photosynthetic potential.
Differences in light intensity also results in photomorphogenic changes in
leaves. Phytochrome provides the sensory pigment which enables plant cells to
react to light quality and intensity. The light-regulated ratio of
phytochrome-far-red to total phytochrome serves to influence protein synthesis
and enzyme activity (Song 1984; Schopfer 1984). Photomorphogenic effects
have implications for photosynthetic activity. Increased irradiance results in
thicker leaves, and this has been attributed to increased thickness of the palisade
mesophyll layer (Reginer et al., 1988; Chabot et al., 1979). Therefore,
photosynthetic tissue per unit leaf area is increased. Ford et al. (1983) found
that carbon dioxide exchange rate was moderately and positively correlated with
specific leaf weight. Wells et al. (1986a) also found that differences in canopy
photosynthesis between soybean cultivars was associated with differences in







specific leaf weight. However, by contrast, Secor et al. (1982) found that specific
leaf weight was able to explain only 4% of the variation in total photosynthesis
among 110 lines of soybean. Nelson and Schweitzer (1988) found that specific
leaf weight in soybean is an effective selection trait, as it tends not to be greatly
influenced by the environment.
Wells et al. (1986b) reported that the quantity of RUBPcase per unit leaf
area was positively correlated with specific leaf weight. Reginer et al. (1988)
reported that when soybean plants were grown at low versus high irradiance, the
total soluble protein per unit leaf area decreased by as much as 40%, RuBPcase
per unit area decreased by 31%, and chlorophyll content per unit leaf area
decreased by 15%. These factors contribute to reduced photosynthetic rates at
lower irradiance. However, in soybean, shaded leaves are not 'parasitic,'
because their respiration rates decrease along with reductions in photosynthetic
capacity (Shibles and Weber 1965; Duncan et al., 1967).
The status of N in both leaf tissue, and the entire plant, is important to the
processes of photosynthesis and remobilization, both of which are related to
seed yield. The rate of carbon dioxide assimilation, chlorophyll content and
RuBPcase activity are proportional to leaf N (Evans, 1983). It has been shown by
Boote et al. (1978) that a decline in photosynthesis was correlated with
decreases in leaf N. Lugg and Sinclair (1981) reported that in lower leaflets
and below a critical protein content per unit leaf area (13 to 16 g m-2) in
uppermost leaflets, there existed a linear relationship between estimated protein
content per unit area and photosynthesis. Buttery (1969) found that N
concentration decreased with increasing plant population density and attributed
this to an inadequate supply of N. Soluble protein and N content in the leaf reach







a maxima during the early reproductive phase (Boon-Long et al., 1983,
Thibodeau and Jaworski, 1975). Pod number and seed yield are correlated with
N accumulation in leaves in the early reproductive phase (Zheng et al., 1987).
Phosphorus status also affects the productive capacity of soybean
canopies. Fredeen et al. (1989) found that low P decreased soybean growth
primarily through effects on the rate of expansion of leaf surfaces. Photosynthetic
rates were also reduced (Fredeen et al., 1989). Sawada et al. (1983) in their
P-stress experiments reported that decreases in photosynthetic activity can be
partially attributed to P deficiency (lack of phosphate compounds in the
photosynthetic machinery). Phosphorus deficiency reduced photosynthetic
capacity (Foyer and Cooper, 1986). Photosynthetic CO2 fixation diminished with
reduced P concentration in leaves (Terry and Ulrich, 1973). Keogh et al. (1972)
found that P levels in 10 soybean cultivars from three maturity groups ranged
from 3.4 to 3.7 g kg-1, with low standard deviations. Nutrient concentration did
not vary much except Ca which was higher in higher-yielding cultivars.
Miller et al. (1961) reported that in the canopy of soybean, upper leaves
tend to have a higher P concentration than lower leaves. They also found a
positive relationship between P concentration (in upper leaves) and yield when
percent K was greater than 1.2 g kg-1. Saturating irradiance for photosynthesis
in P-deficient plants was found to be 30% less than in the control, however, P
deficiency did not affect photosynthetic activity at low irradiances (Sawada et al.,
1983). Thus at higher plant population densities (if there are no deficiencies due
to interplant competition) low P would affect leaves atop the canopy.
Photosynthetic capacity and seed production are also affected by the
process of nutrient remobilization. Nitrogen and P are accumulated in the leaves,







petioles and stems until the pod production stage, after which they are rapidly lost
from these parts and accumulated in the developing seeds (Hanway and Weber,
1971). In upper leaves, N and P concentrations were 50.5 and 3.7 g kg"1
respectively, while in the bottom leaves (older and undergoing translocation) N
and P concentrations were 23.0 and 3.0 g kg"1 respectively (Terman, 1977).
The extent to which developing fruits draw upon nutrient reserves of
vegetative tissue has led to the suggestion that nutrient remobilization leads to
senescence. Depodding prevented senescence and the gradual reinstatement
of senescence as an increasing number of pods is allowed to develop on each
plant (Lindoo and Nooden, 1977). However, Nooden et al. (1978) indicate that
the development of seeds is separable from the senescence response in
soybeans and that seeds may function as more than sinks. They found that sink
size does not parallel foliar senescence; the senescence response is saturated at
a level far below the maximum level of dry weight and nitrogen accumulation in
the seeds. Although drain and diversion may be involved in the monocarpic
senescence of soybeans, it seems unlikely that developing seeds exert such a
remarkable correlative influence simply by functioning as a sink (Nooden et al.,
1978).
Carbon assimilated early in the life cycle has virtually no direct relevance
to fruit nutrition. Therefore, carbon fixed during reproductive development must
be primarily relied upon to furnish the seed's requirement for this element. The
extremely low rate of transfer of early-fixed carbon to seeds is resolved as being
partly due to the fact that much of this carbon has been dissipated in respiration
before flowering has commenced, and partly to the fact that carbon which







survives until fruiting is bound into materials from which it cannot be readily
retrieved during the senescence of vegetative parts (Pate and Flinn, 1973).
On the other hand, Pate and Flinn found that N assimilated early in the
life cycle is released for seed development with great efficiency. Therefore, it
should be theoretically possible to increase rates of transfer to seeds by
arranging for larger reserves of N to have accumulated by the time when
flowering commences. Mobilization is a gradual process which gathers
momentum as fruiting proceeds. Hocking and Pate (1977) in studying legumes
reported that N, P and K are highly mobile and 60 to 90% are usually retrieved.
Approximately half of the N and P in mature seeds appears to have been in other
plant parts prior to pod development (Hanway and Weber,1971). Therefore,
since nutrient accumulation mirrors biomass accumulation, at higher plant
population densities more substrate will be available per unit area for
remobilization making higher seed yields possible.


The Yield-Plant Population Density Response


Explanations of the dynamics of yield in population studies, go beyond
justifying that yield per area tends to increase while yield per plant tends to
decrease with plant population density. The asymptotic or parabolic shape of the
yield-plant population density response needs to be investigated. The main
areas of interest in explaining the shape of yield-plant population density
reponses are light interception, the efficiency with which the crop and plant use
production inputs, biomass accumulation, lodging and barreness. In this
subsection, first, the fundamental and obvious characteristics of the yield-plant







population density response will be discussed; secondly, Duncan's (1986)
theory on yield-plant population density response in soybean will be critically
reviewed; thirdly, the relevance of SOYGRO to investigations of yield-plant
population density responses will be outlined; and fourthly, the possible basis for
changes in the efficiency with which plants use inputs for seed production and its
relevance to yield-plant population density responses will be discussed.


The fundamentals of yield-plant population density responses


In Fig. 2-1 (a) and Fig. 2-1 (b) the typical yield plant population density
responses are shown. Using plant population density as a basis for categorizing
sections of these responses, the following phases are identified: between plant
population densities DO and D1, yield per plant is at a maximum and is constant,
and yield per area assumes a strict linear relationship with plant population
density. At density D1, interplant competition commences and yield per plant
begins to decrease. After plant population density D1, yield per area begins to
deviate from linearity and the slope of the response begins to decrease.
Between plant population densities D1 and D4, although yield per plant
decreases, yield per area increases because plant population density increases
at a rate greater in magnitude than the rate of reduction in yield per plant. At
plant population density D2 the yield per plant response assumes its steepest
negative slope, and between plant population densities D2 and D3 (which occurs
at an early stage of interplant competition) interplant competition has its greatest
impact on the reduction of yield per plant. This impact is given by the magnitude





















- ~(a


DO D1 D2


D4 DS


Plants m-2


Fig. 2-1. Typical responses for yield per unit area (a) and yield per
plant (b) to plant density.







of the negative slope of the relationship between yield per plant and plant
population density between D2 and D3.
At plant population density D3, there is a reduction in the rate of decrease
in yield per plant with plant population density. This reduction may be indicative
of plant adaptation to interplant competition, and possibly implies an increased
efficiency with which the plant is using inputs for seed production.
Between plant population densities D4 and D5 yield per area assumes a
maximum and constant value. A number of crop and plant requirements for
achieving maximum yield per area have been suggested. It has been suggested
that insolation interception must approach 100% early in reproductive growth to
maximize yield per area (Shibles and Weber, 1965; Hawkins, 1982), specifically,
prior to the period of grain production (Shibles and Weber, 1966). Duncan
(1986) argued that complete interception needed to occur only before fruit or
seed number reached a maximum, in order to maximize fruits per unit area.
Duncan's view may be consistent with the findings of Egli et al. (1985) that fruit
number increased even after growth stage R5. However, Johnson et al. (1982)
have suggested that complete light interception is required at the earlier growth
stage of R1, if yield is to be maximized.
Vegetative mass has also been correlated to seed number. This
observation reveals the possible importance of vegetative dry weight in
contributing to differences in yield where light interception is constant, as pointed
out by Duncan (1986). Egli et al. (1987) reported that increasing the vegetative
dry weight to a limit of 500 to 600 g m-2 increased seed number per unit area.
Fruit and seed numbers may be associated with the availability of assimilate
during flowering and fruit set (Stephenson 1981; Heitholt et al., 1985). Christy







and Porter (1982) further point out that seed number per unit area is closely
associated with canopy photosynthesis during flowering and pod set.
At plant population density D5 yield per plant continues to decrease.
However, yield per unit area may also begin to decrease. This decrease in yield
per area is attributed to the traumatic events of lodging and barrenness. Lodging
and barrenness at high densities, are important in explaining parabolic
yield-plant population density responses. At high densities, plants tend to lodge
because of the failure of their mechanical axes as they undergo plastic
responses caused by increased interplant competition. Apart from the reduced
yield that occurs when mechanical harvesting is used on lodged crops, lodged
plants in comparison to upright plants have lower yields. This yield reduction has
been attributed to reduced light interception in lodged plants (Johnston and
Pendleton, 1988). At very high populations plants become barren. Barrenness
can probably be attributed to the low levels of light interception per plant. Fontes
and Ohlrogge (1972) stated that these barren plants utilize light and nutrients, but
do not produce any yield. At high densities, both yield per unit area and yield per
plant decrease resulting in a parabolic response. Leffel (1961), however,
reported that natural lodging of soybeans did not significantly affect yield when
compared to unlodged plants.


Duncan's theory on yield-plant population density responses


Duncan (1986) developed a theory to explain yield-plant population
density responses based on two postulates. First, within soybean planting
patterns there is a range of densities within which seed yield per unit area







increases without an increase in light interception by the fully developed leaf
canopy (this occurs at growth stage R5). Secondly, within limits, the greater the
vegatative weight of a soybean plant during the seed initiation period (which
begins at R5) the more seed it will yield, all other conditions remaining the same.
Duncan further argued that seed yield per unit area was affected by two dominant
considerations: i) the fraction of total photosynthetically active radiation (PAR)
intercepted by the crop and ii), the efficiency with which the intercepted PAR is
used for seed production. Based on the statements above, Duncan defined three
Phases (which correspond to plant population density ranges), which are used to
explain yield-plant population density responses.
In Phase I, there is no appreciable competition between the plants for
light ; light intercepted per plant and seed yield per plant are constant and seed
yield per unit area is influenced only by the fraction of total light intercepted.
Between Phases I and II, there is an increasing degree of mutual shading as
plant population density increases. Light interception per plant decreases;
however, the total light intercepted (per area) increases. Increases in yield per
unit area in this transitional region are due to both increased light intercepted per
unit area, and increased efficiency with which intercepted light is used.
In Phase II, all of the incoming radiation is intercepted at full canopy
development (this occurs at growth stage R5). Therefore, in Phase II light
intercepted at seed initiation is constant, and increases in seed yield per unit
area in this phase are due to the increased efficiency with which intercepted light
is used to produce seed. Phase II ends where seed yield per unit area reaches a
maximum.







In Phase III seed yield per unit area remains constant at a maximum for
the cultivar and the environment. Duncan (1986) was able to identify Phases II
and III using data from Wiggans (1939) and Parks et al. (1983). Egli (1988a) in
his work with 'Cumberland' found that at R5 and a density of 5.1 plants m -2 there
was an apparent increased efficiency with which intercepted light was used for
seed production. This finding supports the existence of Duncan's (1986) Phase
II. Photosynthetically active radiation intercepted attained a maximum of 95%,
after which yield continued to increase with density. The R5 growth stage (seed
initiation) is important in light interception studies because it is at this stage that
LAI attains a maximum (Koller et al., 1970). Egli's (1988a) data supports
Duncan's (1986) argument that plant population densities providing complete
insolation interception by the fully developed leaf canopy, may not be high
enough to maximize yield.

A basis for changes in the efficiency with which plants use inputs for seed
production


In appreciating the possible effect of density-mediated plasticity on the
efficiency with which a plant produces seed, an obvious point of departure is
Huxley's law of allometry, which states that the relative distribution of biomass
among the parts of a plant depends on plant size (Willey and Heath, 1969). Kira
et al. (1956) also observed that the weight of a plant part could be related to the
weight of the whole plant via

wp = kwh w = whole plant weight

w = weight of a plant part k = a constant
P








Plant population density has an obvious effect on plant size, in addition to
other photomorphogenic and plastic effects. A plant's biomass is allocated to
vegetative, photosynthetic, and structural tissues and reproductive organs and.
Obviously, interplant competition has much potential for affecting the efficiency
with which inputs are used by plants in seed production.
Several studies illustrate that soybeans respond to reduced competition
by increasing the size of the metabolic sink sites (Buttery, 1969; Johnston et al.,
1969; Lehman and Lambert, 1960 ) where seed number and seed growth rate
are increased. At higher densities, however, the observed tendency for plants to
utilize intercepted light more efficiently may be of evolutionary significance.
Soybean plants may be exhibiting vestiges of an evolutionary strategy. Holler
and Abrahamson (1977), Ogden, (1974) and Thomas (1974) reported that at high
densities plants tend to maximize seed production. Abrahamson (1975)
suggested that this is a strategy to promote seed dispersal away from an
unfavorable site. Egli's (1988b) work, however, suggested that the partitioning of
assimilate during this growth stage (flowering and fruit set), may not be altered by
changes in plant size and plant population density. Shibles and Weber (1966)
suggested that at higher populations, the period of vegetative production is
longer than at lower densities. As a consequence, the vegetative production
period encroaches upon the seed production period, resulting in competition
within the plant for available carbohydrates. Hence there is less carbohydrate for
seed production. Whatever the possible basis for changes in seed yield
efficiency, it is probably one of the least understood areas in yield-plant
population density relations.












CHAPTER 3
MATERIALS AND METHODS


The experimental procedures used in the field are described prior to
discussions of data analysis. The discussion on data analysis involves the
development of an analytical framework for the analysis of yield-plant population
density data. The materials and method used with SOYGRO will be discussed
along with crop simulation results in chapter 6.


Field Experiments


This study was conducted at Gainesville, Florida (290 38') in 1987 and
1988. The soil was classified as a Kendrick fine sand (a loamy siliceous
hyperthermic family of Arenic Paleudults). The previous crop consisted of small
grains. Rainfall data were collected from a nearby Agronomy Meteorological
Station (Fig. 3-1).
In all the field experiments a square planting arrangement was used.
Wiggans (1939) found that the nearer the arrangement of plants on a given area
approaches a square pattern, the greater the yield per unit area. Safo-Kantanka
and Lawson (1980) observed that the number of pods per plant, seed size and





















600
550 1987 growing season
500
? 450 U 1987 1988 growing season
E 400 [ 1988

g 350
S300
g 250







MAY JUN JUL AUG SEP OCT
MONTH



Fig. 3-1. Precipitation profile over the 1987 and 1988 growing
seasons, Gainesville, Fl.
I-- / V











seasons, Gainesville, Fl.









yield per unit area increased as a square planting pattern was approached. This
finding was corroborated by Parvez et al. (1989) in their work. Buttery (1969), in
citing Donald (1963) and Holliday (1963) stated that it was difficult to
demonstrate any depressing effects of small deviations from squareness on yield.
Square planting is of both analytical and production importance. A
constant square planting arrangement ensures that the differences observed in
plant population studies, resulted from plant population and not plant
arrangement. The square planting arrangement minimizes inter-plant competition
and maximizes insolation interception per unit land area (Miura and Gemma,
1986; Harper, 1983).
In 1987 the soybean cultivars Kirby [MG VIII], Centennial [MG VI] and
Forrest [MG V] were planted on 27-29 May, using planting boards which gave a
square planting arrangement. The inter-row distances of 30, 20 and 15 cm gave
the plant population densities of 11.1, 25.0 and 44.4 plants m-2, respectively.
Vermiculite was used to fill seeded-holes to ensure even germination. The
experimental design was a split-plot with plant population density as the main
plot and cultivar as the subplot. Subplots measured 3.7 by 3.0 m. The resulting
three main plots with their combined total of nine subplots were replicated four
times. Fertilizer, 10-10-10 (N-P205-K20) was applied at a rate of 785 kg ha"1 at
seed-bed preparation. An overhead sprinkler irrigation system was used during
periods of water stress. The plots were hand weeded. In cases where gaps
appeared in stands, soybean plants were transplanted in from border rows.
Transplantation often resulted in weak plants.









The corn ear worm, Heliothis zea (Boddie) was controlled with lannate,
(methomyl S-methyl N [(methylcarbamyl) oxy] thioacetamidate which was applied
in a solution containing 2 g L"1 of active ingredient at a rate of 140 L ha"1. The
cyst nematode, Heterodera glycines affected crop productivity, especially for the
less nematode resistant cultivars, Forrest and Centennial.
The following reproductive growth stages of soybean described by Fehr et
al. (1971) were used in sampling procedures:


R1-One flower at any node
R2-Flower at node immediately below the uppermost nodes with a
completely unrolled leaf.
R3-Pod 0.5 cm long at one of the four uppermost nodes with a
completely unrolled leaf
R4-Pod 2 cm long at one of the four uppermost nodes with a
completely unrolled leaf.
R5-Beans beginning to develop (can be felt when the pod is squeezed
at one of the four uppermost nodes with a completely unrolled leaf.
R6-Pod containing full size green beans at one of the four uppermost
nodes with a completely unrolled leaf.
R7-Pods yellowing; 50% of leaves yellow: physiological maturity.
R8-95% of pods brown: harvest maturity.


Whole plants were hand sampled at the R5, R6 and R8 growth stages (89,
132 and 144 d after planting, respectively). The samples were dried to constant
weight in a forced-draft oven at 60 C. Total dry weights and where appropriate,
pod and seed weights and numbers were determined.
Because of the severe effects of nematode infestation on soybean
production in 1987, the nematode resistant cultivar Kirby was used in 1988.








Plantings took place between 28 -29 June, with planting boards, which gave a
square planting pattern. The plant population densities of 2.0, 4.0, 6.3, 11.1,
16.0, 25.0, 44.4, and 69.4 plants m-2 were achieved. Seeded-holes were filled
with builder's sand to ensure even germination.
The experimental design was a randomized complete block with plots
measuring 4.3 by 3.7 m. Each block consisted of eight plots and was replicated
four times. Fertilizer, 0-10-20 (N-P205-K20) was applied at a rate of 504 kg
ha-1. Aldicarb, [2-methyl-2 (methylthio) propionaldehyde o-(methylcarbamyl)
oxime] was applied at a rate of 27 kg ha"1, at seed bed preparation for nematode
control. The herbicide alachlor, 2-chloro 2' 6'-diethyl-N-(methoxymethyl)
acetanilide was applied two days after planting at a rate of 4.67 L ha1, in a
solution containing 480 g L -1 of active ingredient. An overhead sprinkler
irrigation system was used. Lannate was effective in controlling the corn ear
worm, Heliothis zea (Boddie), however, it was ineffective in controlling the severe
infestation by white flies, Aleyrodidae aleurocanphus woglumi, Ashby. The 'sooty
mold', Capnodium spp. also blackened leaves as they thrived on carbohydrate
exudates of the white flies. It has been shown that the 'sooty mold' interferes with
normal photosynthesis (Vaishampayan and Kogan, 1980). Arioglu et al. (1989)
found that there was a significant negative correlation between white fly
infestation and seed yield.
At R3 and R5, (56 and 68 d after planting, respectively) light interception
above and below the canopy, as well as reflectance, were measured, at or near
solar noon, with a line quantum sensor (LI-COR model 188). Whole plant
samples were taken at R3, R5 and R8. It was not possible to take samples









between R5 and R8 due to the deterioration of vegetative structures brought on
by white flies and extremely wet conditions (Fig. 3-1).
Pods and seeds were retrieved at R8. Subsamples were separated into
their component parts (stems and petioles were kept together). Leaf area was
determined using a leaf area meter (LI-COR 3100). Samples were dried in a
forced-draft oven at 60 OC to constant weight. Vegetative dry weights, pod and
seed weights and numbers were measured.
Subsamples of stems and leaves from R5 were analyzed for N and P.
The samples were chopped in a hammer mill, and then ground in a Wiley mill,
using a 1 mm screen. The samples were digested using a modification of the
aluminum block digestion procedure of Gallaher et al. (1975). To 0.3g of sample,
3.2 g of 9:1 K2S04:CuSO4, were added and digestion carried out for 4 h at
400 OC, using 10 ml H2SO4 and 2 ml H202. Ammonia in the digestate and P
were determined by semiautomated colorometery (Hambleton, 1977) using a
Technicon Autoanalyzer II. The plant variables measured during 1987 and 1988
are provided in Table 3-1.


Data Analysis


The analytical framework developed here for the analysis of yield-plant
density data is based on:


i) the identification of the biological and ecological variables
and relationships existing in yield-plant density studies,










Table 3-1. Summary of the plant variables measured during 1987 and
1988.

-1987
Variables Growth stage

Total dry weight (g m"2) R5,R6,R8
Seed dry weight (g m2) R6,R8
Pod dry weight (g m-2) R6,R8
Seed number ( m2) R6,R8
Pod number ( m2) R6,R8
.--------------------1988------------
Total dry weight (g m-2) R3,R5
Leaf dry weight (g m-2) R3,R5
Stem dry weight (g m-2) R3,R5
Leaf area ( m2) R3,R5
Percent light intercepted R3,R5
Leaf N concentration (g kg -1) R5
Stem N concentration (g kg -1) R5
Leaf P concentration (g kg -1) R5
Stem P concentration (g kg 1) R5
Seed dry weight (g m-2) R8
Pod dry weight (g m-2) R5,R8
Seed number (m-2) R8
Pod number (m-2) R5,R8








the use of analyses of variance (ANOVA), linear regression
and correlations to determine the importance and strength
of these relationships,


the cross referencing of plant population density effects
and yield associations to assess the effect of plant
population density on a) ecological and b) biological
relationships which determine or are associated with yield.


The basic data set
calculated variables:

Stem fraction

Leaf fraction

Leaf area index(LAI)

Specific leaf area (m2 g-1)

Specific leaf weight (g m-2)

Seed number Pod-1

Weight seed"1 (g)

Leaf N (g m-2)

Leaf N (g m'2(leaf area))

Stem N (g m-2)


(Table 3-1.) was expanded to include the following


= Stem weight / Total dry weight

= Leaf dry weight / Total dry weight

= Leaf area / Harvested area

= Leaf area / Leaf dry weight

= Leaf dry weight / Leaf area

= Seed number/ Pod number

= Seed dry weight / Seed number

= Leaf dry weight X Leaf N concentration X 0.01

= Leaf N (g m-2) / LAI

= Stem dry weight X Stem N concentration X 0.01








Leaf P (g m-2) = Leaf dry weight X Leaf P concentration X 0.01

Leaf P (g m-2(leaf area)) = Leaf P (g m-2) / LAI

Stem P (g m-2) = Stem dry weight X Stem P concentration X 0.01




The Evaluation of Variables and Relationships Operating in Yield-Plant
Population Density Studies


In studying the yield determining relationships (both ecological and
biological) existing in crops which are subject to varying plant population
densities, these relationships first need to be classified; secondly, they need to
be measured; and thirdly, their influence on yield needs to be evaluated.


The classification of variables and relationships


In investigating the process of yield achievement, variables and the
relationships between variables, may contribute to or affect three spheres of crop
productivity. These spheres are, photosynthetic activity, nutrient remobilization
and reproductive parameters (i.e., the size and number of 'sinks'). Variables and
their inter-relationships in contributing to these various productive spheres, in
yield-plant density studies, may be either plastic or non-plastic. For example,
weight seed-1 is known to be non-plastic, being largely under genetic and not
under environmental control. In this investigation, variables and relationships
described as not being related nor associated with yield, refer specifically to the
tendency of these variables not to undergo changes similar in magnitude to








changes in the magnitude of yield. The extent of the contribution of 'plastic' and
'non-plastic' variables to yield is not being investigated.
The number of relationships between measured responses (including
those which have been calculated) is given by,



nCr = nPr ( r!)-1 = n (r!(n-r)!)-1

nCr = the total number of relationships, taking two responses at a
time
nPr = the total number of permutations of n responses taken r at a
time
n = number of measured or calculated responses
r = the number of responses assessed at a time (i.e., two)


For the 1988 experiment, there is a combined total of 28 variables from
the R5 and R8 stages. The total number of possible relationships between the R5
and R8 variables is 378 ( 28! (2! x 26!) -1). There are 406 (378 + 28 (plant
population density relationships)) relationships which may be analyzed.


Techniques for determining relationships in yield-plant density studies


The three basic techniques which will be used in measuring the extent of
the relationships operating in yield-plant density studies are the analysis of
variance (ANOVA), linear regression and correlations. The ANOVA will be used
to determine density and cultivar effects on measured variables. Differences
between cultivars will be determined using Duncan's Multiple Range Test








(DMRT). Ordinarily, it is the aim of linear regression analysis to describe a
functional relationship between Y and X (i.e., plant population density) ( Sachs,
1982). However, results of the linear regression analyses obtained in this
investigation will not be seen as describing the functional relationship between
plant population density (X) and a given variable (Y). This is because linear
regressions are not always appropriate for providing exact relationships
between variables in complex biological systems. Linear regression analyses of
plant population density effects, are used here strictly as statistical tests, which
establish the extent of the impact of plant population density on a given variable.
Therefore, of specific interest are the R2 values obtained in linear regressions.
In using linear regression as a statistical test, logarithmic transformations serve
the dual function of linearizing the data and stabilizing the variance (Montgomery,
1984). The Pearson product moment correlation coefficient (r) which will be
used here as a statistic for the strength of the relationship between two variables.
Correlations will be used to assess the strength and nature (whether it is positive
or negative) of primary, secondary and tertiary relationships.


The evaluation of the relationships operating in yield-plant density studies


The evaluation of the ecological and biological relationships which may
be yield-determining will be conducted by cross referencing i), plant population
density effects on variables, and ii), the association of these variables with yield.
The first step is the determination of the strength of the effect of plant population
density on measured variables. Secondly, the strength of the association








between measured variables and yield per unit area is determined. These
relationships are then categorized based on R2 (from linear regression) and r
(from correlations) values into high (H) (0.75 1.00), moderate (M) (0.50 0.74)
and low (L) (0.00 0.49) categories. The next step which will be illustrated using
only the 'high' (H) and 'low' (L) categories, involves the correlation of 'responses
classified with respect to their relationship with yield' against 'responses
classified with respect to their relationship with plant population density'. The
resulting r, is then used to determine the possible contribution of primary,
secondary and tertiary relationships to yield per unit area, in plant population
density studies.
Based on R2 and r values, variables can be classified into those; affected
by plant population density (HDV); not affected by plant population density (LDV);
associated with yield (HYV); and not associated with yield (LYV). In some
instances, HDVs and HYVs may be the same. The types of relationships which
may exist between these variables are,



HDVs x HYVs
HDVs x LYVs
LDVs x HYVs
LDVs x LYVs


The interpretation of the strength of these relationships as determined by r
values are outlined in Table 3-2. The relationships which are of greatest interest
are those involving HDVs X HYVs, and LDVs X LYVs with high r values. The
former identifies variables and relationships which are particularly plastic, and









Table 3-2. A table for the interpretation of plant
cross-referenced data.


population density yield,


Type of relation A high r A low r


A strong relationship which
is most likely to be yield
determining and which is
under plant population
density effect.





A strong relationship,
however, this relationship,
cannot explain yield as a
plant density-mediated
process.




A strong relationship,
however, this relationship,
cannot explain yield as a
plant density-mediated
process.






A strong relationship between
variables which are not
associated with yield and
are not affected by plant
density.


A weak relationship, however
the HYVs are greatly associated
with yield, and the HDVs are
under plant population density
effect. This relationship cannot
be yield determining in plant
population density studies.



A weak relationship not
capable of explaining yield,
however, the HDVs are
under plant population density
effect.




Although the HYVs are
closely associated with yield,
this weak relationship is not
capable of explaining yield
as a plant density-mediated
process.





A weak relationship between
variables which are not
associated with yield
and are not affected
by plant population density.





40


which are most likely to be responsible for the changes in the magnitude of yield
per unit area with changes with plant population density. The latter identifies
variables and relationships, which are not particularly plastic (which are probably
under much genetic control) and cannot explain changes in the magnitude of
yield per unit area with changes in plant population density.












CHAPTER 4
RESULTS AND DISCUSSION -1987 EXPERIMENT


Plant population had no significant effect on any of the variables
measured. In the absence of significant plant population effect on crop variables,
the cross referencing of the effect of plant population density on variables with the
'yield association' of variables is not possible. However, the cultivar effect on crop
variables and the effect of plant population density at levels above the 5%
significance level, allows for the use of ANOVA and correlation analyses for yield
investigation.


ANOVA Results


Plant Population Effect
Plant population had no effect on any of the variables investigated at R5
(Table 4-1), R6 (Table 4-3), and R8 (Table 4-5). In instances where this result
was unexpected (such as with seed and total dry weight per unit area), it may be
attributed to a combination of i), the plant populations used were not sufficiently
discriminating, and ii), nematode infestation. The nematode resistant soybean
varieties recommended for Florida (IFAS, 1983), does not include Kirby and
Forrest. Centennial, however, is recommended. Analyses conducted on root
and soil samples showed that the cyst nematode (Heterodera) (race unidentified,
and mostly juveniles), and root knot nematode (Meloidogyne incognita,









Table 4-1. The effect of plant population and cultivar on total dry
weight (g m'") at R5, 1987.

Source of variation F

Population 5.56
Cultivar 31.8*
Population X Cultivar 1.96

*, *, significance at the 5 and 1% levels, respectively.








Table 4-2 Cultivar effect on total dry weight at R5,
1987.

Cuttivar Total dry weight (g m-2)

Kirby 957a*
Centennial 723b
Forrest 551 c

* observations followed by the same letter are not
significantly different (DMRT, 5%).













Table 4-3 The effect of plant population and cultivar on soybean plant traits at R6
stage, 1987.

Variable Source of variation F


Total dry weight (g m-2)


Pod weight (g m-2)


Seed dry weight (g m-2)


Pod number (m-2)


Seed number (m-2)


Seed number Pod -1



Weight seed 1 (g)


Population
Cultivar
Population X cultivar
Population
Cultivar
Population X cultivar
Population
Cultivar
Population X cultivar
Population
Cultivar
Population X cultivar
Population
Cultivar
Population X cultivar
Population
Cultivar
Population X cultivar

Population
Cultivar
Population X cultivar


*, *, significance at the 5 and 1% levels, respectively.


0.21
10.84**
1.91
0.15
12.06**
2.28
0.27
10.26**
1.92
0.34
1.90
2.18
0.28
2.69
2.24
2.32
0.56
1.27

1.07
33.37**
0.73


















soybean plant traits at R6 stage, 1987.


Cultivar


Kirby


Centennial


Forrest


Total dry weight (g m"2) 644 a* 598 a 443 b
Pod weight (g m'2) 298 a 318 a 223 b
Seed dry weight (g m"2) 159 b 198 a 141 b
Pod number (m-2) 1236 a 1213 a 1056 a
Seed number (m-2) 2400 a 2294 a 2015 a
Seed number Pod -1 1.95 a 1.90 a 1.94 a
Weight seed -1(g) 0.07 b 0.09 a 0.07 b

* observations followed by the same letter are not significantly
different (DMRT, 5%).


Variable


Table 4-4. Cultivar effect on









Table 4-5. The effect of plant population and cultivar on soybean plant traits at R8
stage, 1987.

Variable Source of variation F


Pod weight (g m-2)


Seed dry weight (g m-2)


Pod number (m-2)


Seed number (m-2)


Seed number Pod -1



Weight seed -1 (g)


Population
Cultivar
Population X cultivar
Population
Cultivar
Population X cultivar
Population
Cultivar
Population X cultivar
Population
Cultivar
Population X cultivar
Population
Cultivar
Population X cultivar

Population
Cultivar
Population X cultivar


*, ", significance at the 5 and 1% levels, respectively.


2.14
11.82**
1.92
1.08
12.76**
1.85
0.73
1.95
4.03*
1.53
2.29
4.36*
0.57
8.67**
0.42

0.40
14.69**
1.73








'Southern') populations were sufficient to hinder grain production in all three
cultivars.
At R5, total dry weight (g m -2) tended to increase with population (Fig.
4-1). Although such an observation is consistent with the results of Udoguchi and
McCloud (1986), Blumenthal et al. (1988) and Parvez et al. (1989), the
population effect was not significant (p = 0.07) (Table 4-1). There was no
significant difference in seed dry weight between plant populations at R8 (Fig.
4-2) (Table 4-5), and this can be attributed to the effect of plant population density
on yield components. The number of seeds per pod (R8) was not affected by
plant population (Fig. 4-3) (Table 4-5). This result corroborates those of
Chaudhry and Cheema (1985) and Dominguez and Hume (1978). Weight per
seed (R8) (g) remained unchanged with plant population (Fig. 4-4) (Table 4-5),
and this is similar to the results obtained by Dominguez and Hume (1978), and
Fontes and Ohlrogge (1972). Pod number m'2 was unaffected by plant
population density (Fig 4-5) (Table 4-5). With none of the yield components
being affected by plant population, the absence of plant population effect on seed
dry weight (g m-2) at R8 is expected.


Cultivar Effect
Total dry weight (g m-2) at R5 was affected by cultivar (Table 4-1) with
Kirby, Centennial and Forrest providing 970, 723 and 551 g m-2, respectively
(Table 4-2 ). At R6, pod number m-2, seed number m"2 and seed number pod"1
were not affected by cultivar (Table 4-3). However, total dry weight (g m-2), pod
dry weight (g m-2), seed dry weight (g m-2) and weight per seed (g), were all
























900


850-


800-


750 -


700


650-


n -


uIJIJ


I I I I *I I I I
10 15 20 25 30 35 40 45


Plants m-2


Fig. 4-1. The effect of plant population density on total dry
weight at the R5 stage in soybean, 1987.







































10 15 20 25


30 35 40 45 50


Plants m"2


Fig. 4-2. The effect of plant population density on seed dry
weight of soybean at maturity, 1987.


260


250


240


230


220


210




49

















1.96


1.94-


7 1.92

CL
0
0 1.90


1.88


1.86- .
10 15 20 25 30 35 40 45 50

Plants m-2



Fig. 4-3. The effect of plant population density on seeds pod" of
soybean at maturity, 1987.




50
















0.091


0.090


7 0.089


Ca 0.088


0.087


0.086 -
10 15 20 25 30 35 40 45 50

Plants m-2


Fig. 4-4. The effect of plant population density on weight seed of
soybean at maturity, 1987.





















1600


1550


1500


E 1450
E
^ 1400-


1350


1300 ,. ,i i
10 15 20 25 30 35 40 45 50

Plants m-2


Fig. 4-5. The effect of plant population density on pod number m-2
of soybean at maturity, 1987.








affected by cultivar. These results are summarized in Tables 4-3 and 4-4. At R8
(Table 4-5), all variables were affected by cultivar. The interaction of cultivar and
population on seed number m-2, and pod number m'2 warranted further
investigation; upon analysing the cultivars individually, for both pod number m'2
(Table 4-6) and seed number m"2 (Table 4-7) there was no population effect in
either case. At each plant population, there was no cultivar effect on pod number
m"2 (Table 4-8) and seed number m'2 (Table 4-9). A summary of the cultivar
effects at R8 is provided in Table 4-10.


An Explanation for Cultivar Yield Differences Based on ANOVA


In Table 4-10. a comparison of seed dry weight (R8) (g m-2) between
Kirby, Centennial and Forrest shows that yields were 268, 270 and 178 g m -2,
respectively, with yield of Forrest being significantly lower. A review of yield
components and total dry weight (R5) (g m-2) shows that pod number m-2, and
seed number pod -1, are unable to account for differences in seed dry weight (g
m-2). Weight seed -(g) and total dry weight (R5) (g m"2) appear to be
responsible for the seed dry weight (g m-2) differences observed for these
genotypes.
It has been shown that pod number m-2, and seed number pod-1, are
unable to account for differences in seed dry weight (g m-2). Kirby and
Centennial are from higher maturity groups than Forrest and, therefore, have
longer grain filling periods, and this may have attributed to Kirby and Centennial
having larger seeds. Also, of the three cultivars, Forrest was the least nematode
resistant and this may have taken away further from the ability of Forrest to








Table 4-6. The effect of plant population (X) on pod number (m-2) (Y) for each
cultivar at R8 stage, 1987.


Cultivar Statistical model R2


Kirby InY = 0.01 InX + 7.29 0.00
Centennial InY = -0.32 InX + 8.27 0.35
Forrest InY = 0.26 InX + 6.29 0.37


*,* significance at the 5 and 1% levels, respectively.





Table 4-7. The effect of plant population on seed number (m-2) for each cultivar
at R8 stage, 1987.


Cultivar Statistical model R2


Kirby In Y = -0.05 InX + 8.03 0.02
Centennial In Y = -0.32 InX + 8.94 0.37
Forrest In Y = 0.23 InX + 7.04 0.35


*,** significance at the 5 and 1% levels, respectively.










Table 4-8. The effect of cultivar on pod number (m-2) for each plant population at
R8 stage,1987.


Plant population Source of variation F


44.4 plants m"2 Cultivar 3.53

25.0 plants m"2 Cultivar 2.20

11.1 plants m"2 Cultivar 4.22

*,* significance at the 5 and 1% levels, respectively.


Table 4-9. The effect of cultivar on seed number (m-2) for each plant population
at R8 stage, 1987.


Plant population Source of variation F


44.4 plants m"2 Cultivar 3.01

25.0 plants m-2 Cultivar 3.11

11.1 plants m'2 Cultivar 4.50


*,significance at the 5 and 1% levels, respectively.













Table 4-10. Cultivar effect on soybean plant traits at R8 stage, 1987.


Cultivar
Variable Kirby Centennial Forrest
Pod weight (g m-2) 405 a 385 a 266 b
Seed dry weight (g m-2) 268 a 270 a 178 b
Pod number (m-2) 1493 a 1499 a 1265 a
Seed number (m-2) 2701 a 2964 a 2457 a
Seed number Pod -1 1.81 b 1.99 a 1.95 a
Weight seed -1(g) 0.10 a 0.09 a 0.07 b

* observations followed by the same letter are not significantly different
(DMRT, 5%)







perform. Forrest also had the lowest vegetative dry weight at R5, and this means
that in comparison to Kirby and Centennial, less nutrients were available for
remobilization.



Correlation Analysis


The correlation of variables at R5 and R6, and R5 and R8, with seed dry
weight (g m-2) and yield components, are provided in Tables 4-11 and 4-12,
respectively. Total dry weight (g m-2) at R5 was poorly correlated with seed dry
weight (g m-2) at R6 (r = 0.34, p < .05). Total dry weight (g m-2) at R5 was
moderately correlated with seed dry weight (g m-2) at R8 (r = 0.56, p < .01). Total
dry weight (g m-2) at R5 was also moderately correlated with all yield
components at R8.
At both R6 and R8, pod dry weight (g m-2), pod number (m-2) and seed
number (m-2) were highly correlated with seed dry weight (g m-2).
Basuchaudhari (1987), Kaw and Menon (1972), Fontes and Ohlrogge (1972),
and Parvez et al. (1989) had similar findings. Weight seed -1 (g) was
moderately correlated with seed dry weight (g m-2) at both R6 and R8. These
results are contrary to those obtained by Board (1985), Tandojam (1986),
Johnson et al. (1955). However, Moorthy (1952) found that there was a high
correlation between weight seed '1(g) and seed dry weight (g m-2). At R6 there
was moderate correlation between seed dry weight (g m-2) and seed number
pod'1, this result was also obtained by Ma (1946). At R8, seed number pod"1
was not correlated with seed dry weight (g m-2). This result corroborates those
obtained by Hartwig and Edwards (1970), Board (1985), Tandojam (1986) and







Table 4-11. Correlation coefficients for seed yield and its fundamental
components (X1) at R6 with variables (X2) at R5 and R6, 1987.

Xl X2 r


Seed dry weight (R6)(g m-2)







Pod number (R6) (m-2)







Seed number Pod-1 (R6)







Weight seed -1 (R6) (g)


Total dry weight (R5)(g m-2)
Total dry weight (R6)(g m-2)
Pod dry weight (R6)(g m-2)
Pod number (R6)(m-')
Seed number (R6) m-2)
Seed number Pod- (R6)
Weight seed -1 (R6) (g)

Total dry weight (R5)(g m-2)
Total dry weight (R6)(g m-2)
Pod dry weight (R6) (g m-2
Seed dry weight (R6) (g ma )
Seed number (R6) (m-)
Seed number Pod- (R6)
Weight seed -1 (R6) (g)

Total dry weight (R5)(g m-2)
Total dry weight (R6)(g m-2)
Pod dry weight (R6) (g m-2
Seed dry weight (R6) g m)
Pod number (R6) (m-
Seed number R6) (m-)
Weight seed (R6) (g)

Total dry weight (R5)(g m-2)
Total dry weight (R6)(g m-2)
Pod dry weight (R6) (g m-2
Seed dry weight (R6) g m")
Pod number (R6) (m-
Seed number (R6) (m-')
Seed number (R6) Pod-1


'," significance at the 5 and 1% levels, respectively.


0.34*
0.82**
0.93**
0.83**
0.81**
-0.48*
-0.56**

0.55**
0.87**
0.88**
0.83**
0.97**
-0.56**
0.06

-0.21
-0.35
-0.42*
-0.48*
-0.56**
-0.34
-0.35

-0.12
0.14
0.33
0.56*
0.06
-0.03
-0.35








Table 4-12. Correlation coefficients for the fundamental yield components (X1) at
stage R8 with variables (X2) at R5 and R8, 1987.


Xl X2 r


Seed dry weight (R8)(g m-2)


Pod number (R8) (m-2)





Seed number Pod-1 (R8)





Weight seed -1 (R8) (g)


Total dry weight (R5) (g m-2)
Pod dry weight (R8) (g m'2)
Pod number (R8) (m'"I
Seed number (R8) m' )
Seed number Pod' (R8)
Weight seed -1 (R8) (g)

Total dry weight (R5) (g m-2)
Pod dry weight (R8) (g m"2
Seed dry weight (R8) (g m")
Seed number (R8) (m' )
Seed number Pod" (R8)
Weight seed -1 (R8) (g)

Total dry weight (R5) (g m-2)
Pod dry weight (R8) (g m"2,
Seed dry weight (R8)Pg m' )
Pod number (R8) (m'-
Seed number R8) (m'")
Weight seed (R8) (g)

Total dry weight (R5) (g m-2)
Pod dry weight (R8) (g m-2
Seed dry weight (R8) g m')
Pod number (R8) (m'
Seed number (R8) (m" )
Seed number Pod' (R8)


*,** significance at the 5 and 1% levels, respectively.


0.56"
0.98"
0.83"
0.83"
-0.13
0.45*

0.43*
0.79"
0.83"
0.97"
-0.28
-0.06

-0.56*
-0.15
-0.13
-0.28
-0.04
-0.16

0.55"
0.47*
0.45*
-0.06
-0.11
-0.16








Johnson et al. (1955). Among yield components at both R6 and R8, seed number
pod-1 and weight seed1 (g) are the least correlated with the variables measured,
and this may indicate that these yield components are largely under genetic
control.

An Explanation for Cultivar Yield Differences Based on Correlation
Analyses


Based on DMRT, it has already been shown that weight seed -1(g) and
total dry weight (R5) (g m-2) appear to be associated with cultivar seed dry weight
(g m"2) differences (Table 4-11). However, a number of important points are
brought out in attempting to explain differences observed in seed dry weight
(g m-2) between Kirby, Centennial and Forrest using correlations. First, weight
seed -1 and total dry weight (R5) (g m-2), which, based on DMRT, can explain
cultivar differences in seed dry weight (g m-2) are not highly correlated with seed
dry weight (g m-2) (Table 4-13). This raises the point that the investigation of the
effect of individual variables on yield, as i), with DMRT analyses (Table 4-10) or
ii), the correlation of a single variable with seed dry weight (g m-2) has limitations.
Cultivar differences observed in seed dry weight (g m-2) are evidently attributable
to a number of variables acting together, and the investigation of the effect of one
variable on seed dry weight (g m-2) yields an incomplete result.
Secondly, in Table 4-13, the combined data from all cultivars, when
regressed against seed dry weight (g m-2), provides r values which do not
represent r values obtained for each cultivar. This discrepancy is particularly the
case for seed pod -1 and total dry weight (g m-2). Therefore, the individual yield
dynamics of cultivars and the differential response of cultivars to plant population









Table 4-13. Mean values and correlation coefficients for R5 total dry weight
(g m-2), seed dry weight (g m ') and its fundamental components at R8,1987.


CuLIi


Kihtu


nrntanrnil


Fnrrect


All
^i intiu!re


Seed dry weight~g m-2)
Pod number (m")
Seed Pod -1
Weight seed -1(g)
Total dry weight(R5)(g m-2)


268a+
1493a (.77*)@
1.81b (.33)
0.10a (.48)
957a (.30)


270a
1499a (.96"*)
1.99a (-.52)
0.09a (-.47)
723b (.35)


178b
1265a (.93**)
1.95a (-.33)
0.07b (-.21)
551c (.56)


+ observations followed by the same letter are not significantly different
(DMRT, 5%).

@Correlation coefficients (r) of variables with seed dry weight (g m-2) are in
parentheses.

*,** significance at the 5 and 1% levels, respectively.



Table 4-14. Statistical models for per plant responses (Y) of soybean with plant
density (X)


Variable (Y) Statistical model R2


Vegetative dry weight Plant -1 (R5) InY = -0.811nX + 5.99 0.74**

Pod dry weight Plant -1 (R8) InY = -1.021nX + 5.90 0.84**

Seed dry weight Plant -1 (R8) InY = -1.041nX + 5.55 0.80**

Pod number Plant -1 (R8) InY = -1.021nX + 7.28 0.82**

Seed number Plant -1 (R8) InY = -1.041nX + 8.01 0.84*


*, significance at the 5 and 1% levels, respectively.


(.83")
(-.13)
(.45*)
(.56")


ra~crr~ I\PA








density, indicates that in depth yield-plant population density studies should
focus on a single cultivar.


The Yield-Plant Population Density Response.


Seed weight (g m-2) was not affected by plant population density, and no
lodging was observed at the highest plant population densities. However, the
vegetative dry weight (R5) (Fig 4-6), seed weight and pod weight (Fig. 4-7 ), and
seed and pod number (Fig. 4-8), on a per plant basis were all significantly
affected by plant population density (Table 4-14). The reduction in vegetative dry
weight per plant with density (Fig. 4-6), indicates a reduction in the amount of
nutrients which may be remobilized for seed production. Individual plants clearly
adjusted plastically to increased plant competition; at higher plant population
densities they reduced the number of sink sites which is consistent with reduced
inputs at higher plant population densities. It appeared that interplant
competition had its greatest effect between the plant population densities of 11.1
and 25.0 plants m-2, where vegetative dry weight and yield reduction occurred at
the greatest rate. As population density approached 44.4 plants m-2, plastic
adjustments within the plant reduced the rate of yield decrease with plant
population density--that is, the plants were adapting and responding positively to
the negative forces of interplant competition.





62















60

55-
50-

45-
40

r 35

30
0 25

20
15

10 -
10 15 20 25 30 35 40 45 50

Plants m-2

Fig. 4-6. The effect of plant population density on dry weight per
plant of soybean at the R5 stage, 1987.






















40

35-

30-

25-

20-

15-


----Pod
-- Seed


10 15 20 25 30 35 40 45
10 15 20 25 30 35 40 45


Plants m -2

Fig. 4-7. The effect of plant population density on seed and pod
weight per plant of soybean at maturity, 1987.




64















300

SP-od
250 Seed


200

E
S 150

z 100


50


0-
10 15 20 25 30 35 40 45 50

Plants m-2


Fig. 4-8. The effect of plant population density on seed and pod
number per plant of soybean at maturity, 1987.








Conclusion


Although plant population had no significant effect on the variables
measured at R5, R6 and R8, the effect of cultivar on variables allowed for yield
investigations. Total dry weight (g m -2) was moderately correlated with all yield
components, however, it was negatively correlated with seed number pod -1
Correlation analyses on the data sets for each cultivar, and for the combined
data set of all cultivars, at both R6 and R8, showed that pod dry weight
(g m -2), seed number (m-2), and pod number (m -2) were all highly correlated
with seed dry weight (g m-2). Seed number pod-1, was unaffected by plant
population density, and Kirby had the lowest value. However the greater weight
seed -1 (g) of Kirby, contributed to Kirby having a greater seed dry weight
(g m -2) at R8 than Forrest.
Weight seed -1 (g) was neither plastic nor was it highly correlated with
seed dry weight (g m-2). However, among yield components, weight seed -1
accounts for the yield difference observed in cultivar. This apparent contradiction
in the influence of weight seed -1 (g) highlights the following point. Weight
seed-1 (g) determines final seed dry weight (g m-2); however, being under
genetic control, it can account for differences in genotypes, but weight seed -1
(g) is unable to account for the plastic response of seed weight (g m-2).
The combination and subsequent analysis of data obtained from different
cultivars in yield-plant population density studies, poses a major challenge.
Cultivars may differ greatly in their yield dynamics, and as such they may respond
differently to changes in plant population. If this is the case, then the analysis of




66


combined data sets from different cultivars may yield results which misrepresent
biological and ecological realities.













CHAPTER 5
RESULTS AND DISCUSSION-1988 EXPERIMENT


The first section of this chapter focuses on i) the effect of plant population
density on crop variables ii), the extent of the correlation of these variables with
seed dry weight per unit area and iii), the yield implications of i) and ii). In the
second section the cross referencing of plant population density effects with the
strength of the yield association of variable, is used to establish yield determining
relationships (between variables) which are of importance in yield-plant
population density studies.


Plant Population Density Effects. Yield Correlations and their Implications
for Yield


Seed dry weight (g m-2) increased with plant population density
(Fig. 5-1). Attempts to explain this result were done by categorizing the variables
which have been measured as being i) vegetative ii), related to nutrient status
and iii), reproductive. The effect of plant population density on R3, R5 and R8
variables was determined by conducting linear regressions on data which were
linearized through logarithmic transformations. These models only show if plant
population density had a significant effect on the variables in question, and with
the aid of the model's slope, whether this effect was positive or negative.




68
















500

450 -- Seed
400 -- Pod
350
E
S 300 -
S 250
S 200
150-

100-
50-

0 10 20 30 40 50 60 70 80

PI3M-2
Plants m-2


Fig. 5-1. The effect of plant population density on seed and pod
weight of Kirby at maturity, 1988.








Correlations of variables with seed dry weight (g m -2) were determined for
untransformed data.


Vegetative Structures

Dry weight response to plant population density


At R3, total dry weight (g m-2) leaf dry weight (g m-2) and stem dry matter
(g m-2) (Fig 5-2) (Table 5-1), were all positively affected by plant population
density. At R5, total dry weight (g m-2), leaf dry weight (g m-2) and stem dry
weight (g m-2) (Fig. 5-3) (Table 5-2) were all positively affected by plant
population density. These results are consistent with those of other researchers
(Parvez et al., 1989; Blumenthal et al., 1988; Udoguchi and McCloud, 1986).
At R3 and R5 total dry weight per plant decreased (Fig. 5-4) and was
significantly affected by plant population density (Tables 5-1 and 5-2). Total dry
weight decreased sharply up to a density of 25 plants m-2, and tended to level off
after 44.4 plants m-2, for both R3 and R5 (Fig. 5-4). Obviously at higher densities,
increases in plant population density were not as effective in reducing plant dry
weight.


Dry weights (a m-2) and correlations to seed yield.


Total dry weight (g m-2), leaf dry weight (g m-2) and stem dry weight
(g m-2) at R3 were positively and highly correlated with seed dry weight (g m-2)
at R8 (Table 5-3). Similar results were obtained at R5 where total dry weight










Table 5-1. Statistical models for soybean plant variables
density (X) at R3 stage, 1988.


Variable (Y)


(Y) with plant


Statistical model


Total dry weight(g m'2)
Leaf dry weight (g m'-2
Stem dry weight (g m'
Total dry weight plant *(g)


InY = 0.411nX + 5.14
InY = 0.341nX + 4.43
InY = 0.471nX + 4.47
InY = 0.581nX + 5.14


* ,** significance at the 5 and 1% levels, respectively.









Table 5-2. Statistical models for soybean plant variables (Y) at R5 stage with
plant density (X), 1988.


Variable (Y)


Total dry weight (g m-2)
Leaf dry weight (g m-2)
Stem dry weight (g m'
Total dry weight plant (g)


Statistical model


InY = 0.431nX + 5.34
InY = 0.411nX + 4.49
InY = 0.411nX + 4.78
InY = 0.521nX + 5.28


*,significance at the 5 and 1% levels, respectively.


0.86*
0.78*
0.88**
0.92*


0.79"
0.75"
0.56"
0.87"






















1400

1200 ---- Stem
Leaf
1000 --- Total

E 800

S 600

C 400

200

0-
0 10 20 30 40 50 60 70 80

Plants m"2


Fig. 5-2. The effect of plant population density on stem, leaf and
total dry weights of Kirby at the R3 stage,1988.





















2000

1750 --- Total
Leaf
. 1500 -- Stem

>E 1250

. 1000

" 750

500

250-

0
0 10 20 30 40 50 60 70 80

Plants m-2


Fig. 5-3. The effect of plant population density on stem, leaf and
total dry weights of Kirby at the R5 stage, 1 988.






















140

120

100-

80

60

40

20

0 -
0


10 20 30 40


Plants m-2


5 6 7I 8
50 60 70 80


Fig. 5-4. The effect of plant population density on vegetative dry
weight of Kirby at the R3 and R5 stages, 1 988.


---a- R5
R3


I*1*1*I














Table 5-3. Correlation oefficients of soybean plant variables at R3 stage with
seed dry weight (g m'") at R8, 1988.
Variable r
Total dry weight (g m-2) 0.80"
Leaf dry weight (g m~ 0.77**
Stem dry weight (g m') 0.81

*, significance at the 5 and 1% levels, respectively.







Table 5-4. Correladon coefficients of soybean plant variables at R5 stage with
seed weight (g m'") at R8, 1988.
Variable (Y) r

Total dry weight (g m"2) 0.63*
Leaf dry weight (g m-22 0.65"
Stem dry weight (g m'2) 0.54*


*,* significance at the 5 and 1% levels, respectively.









(g m-2), leaf dry weight (g m-2) and stem dry weight (g m-2) at R5, were
moderately correlated with seed dry weight (g m-2) at R8 (Table 5-4). Udoguchi
and McCloud (1986) and Parvez et al. (1989) also had similar findings.
However, Basuchaudhari (1987) found only moderate correlations.




Yield implications.


Total vegetative dry weight per plant decreased with plant population
density and therefore, there was less potential for higher yields on a per plant
basis. However, the increase in plant population offset the reduction in total dry
weight per plant, and as a result total dry weights (g m-2) increased with plant
population density. The high correlation of dry weight (g m-2) with seed dry
weight (g m -2), indicates that accompanying such dry weight increases per unit
area are i) increased photosynthetic capacity per unit land area and ii), increased
amounts of nutrients per unit land area for remobilization. A better appreciation
of the role of dry weight in productive processes comes with examining measures
of its distribution--specifically leaf fractions, the leaf area index, etc.


Leaf fraction, leaf area and light intercepted and linear regressions


At R3, the stem fraction was positively affected while leaf fraction was
negatively affected by plant population density (Fig. 5-5) (Table 5-5). However,
at R5 plant population density had no significant effect on the stem and leaf




















1.0-

0.9-

0.8-

0.7-

0.6-

0.5-

0.4-

03


0 10 20 30 40 50 60 70 80


Plants m'2



Fig. 5-5. The effect of plant population density on stem and leaf
fractions of Kirby at the R3 and R5 stages, 1988.


---- R5 leaf
R5 stem
-- R3 leaf
--- R3 stem








_- ----


."













Table 5-5. Statistical models for soybean plant variables
density (X) at R3 stage, 1988.


(Y) with plant


Variable (Y) Statistical model R2


Stem fraction
Leaf fraction
Percent light intercepted
Leaf area index
Specific leaf area (m2 g-1
Specific leaf weight (gm m )


InY = 0.051nX .66
InY = -0.071nX .71
InY = 0.091nX + 4.29
InY = 0.401nX + .63
InY = 0.071nX 3.81
InY = -0.071nX + 3.81


*,* significance at the 5 and 1% levels, respectively.


Table 5-6 Statistical models for soybean plant variables
(X) at R5 stage, 1988.


(Y) with plant density


Variable (Y) Statistical model R2


Percent light intercepted
Stem fraction
Leaf fraction
Leaf area index
Specific leaf area (m2 g-1
Specific leaf weight(g m' )


InY = 0.071nX + 4.35
InY = -0.021nX .56
InY = -0.021nX .85
InY = 0.381nX + .80
InY = -0.031nX 3.69
InY = 0.031nX + 3.69


*,** significance at the 5 and 1% levels, respectively.


0.61"*
0.63"
0.59"
0.83"
0.25*
0.25*


0.56"
0.01
0.02
0.67**
0.01
0.01








fractions (Fig. 5-5) (Table 5-6), although the plant density trends were similar to
those at R3. At R3, specific leaf area (m2 g'l) (Fig. 5-6) and specific leaf weight
(g m-2) were affected by plant population density (Table 5-6). At R5, specific leaf
area (m2 g -1) (Fig. 5-6) and specific leaf weight (g m-2) were not affected by
plant population density (Table 5-6). By comparison, Nelson and Schweitzer
(1988), observed that specific leaf weight (g m-2) (and hence, specific leaf area
(m2 g-1) ) was not greatly affected by environments.
At higher plant population densities, due to soybean's phyllotaxy, it is
expected that a greater proportion of leaves will be shaded. However, the lack of
differences in specific leaf weight at R5, may have been due to the fact that there
was probably greater differences within canopies, than between canopies. The
specific leaf weights and specific leaf areas taken were average measures over
the whole canopy and they do not indicate differences with height in canopies.
At R3, the leaf area index (Fig. 5-7) and percent light intercepted (Fig. 5-8), were
affected by plant population density (Table 5-5). At R5, both the leaf area index
( Fig. 5-7 ,Table 5-6 ) and the percent light intercepted ( Fig. 5-8, Table 5-6 ) were
affected by plant population density. Light interception reached a maximum of
99% and leveled off at a population density of 25 plants m2.


Leaf fraction, leaf area and light Intercepted and correlations with seed yield.


At R3, stem and leaf fractions were highly correlated with seed dry weight
(g m-2) at R8 (Table 5-7). However, at R5 stem and leaf fractions were not
correlated with seed dry weight (g m-2) (Table 5-8). At R3, specific leaf weight























---- R5
- R3


4. -


0.032-

0.030-

0.028-

0.026-

0.024-

0.022-

0.020-

0.018-

0.016


50 60 70 80


Fig. 5-6. The effect of plant population density on specific leaf area
of Kirby at the R3 and R5 stages, 1988.


10 20 30 40

Plants m2


























10-


. R 5s
8R3


2 I I I I I I I
0 10 20 30 40 50 60 70 80

Plants m-2

Fig. 5-7. The effect of plant population density on the leaf
area index of Kirby at the R3 and R5 stages, 1988.




81















100

95

.2 90

S 85

-c 80-
= --- R5
R3
a- 70 -

65-

60 i -, ,, -, ,
0 10 20 30 40 50 60 70 80

Plants m -2



Fig. 5-8. The effect or plant population density on percent
light interception of Kirby at the R3 and R5 stages, 1988.
















Table 5-7. Correlation coefficients of soybean plant variables at R3 stage
with seed weight (g m'") at R8, 1988.

Variable r

Stem fraction 0.69*
Leaf fraction -0.69"
Percent light intercepted 0.67*
Leaf area index 0.79*
Specific leaf area (m2 g-1 0.41'
Specific leaf weight (g m ") -0.46*

*,* significance at the 5 and 1% levels, respectively.







Table 5-8. Correla~jon coefficients of soybean plant variables at R5 stage with
seed weight (g m'") at R8, 1988.

Variable r

Percent light intercepted 0.65"
Stem fraction -0.11
Leaf fraction 0.07
Leaf area index 0.65
Specific leaf area (m2 g'l~ -0.10
Specific leaf weight (g m'-) -0.10


*," significance at the 5 and 1% levels, respectively.








(g m-2) and specific leaf area (m2 g -1) were moderately correlated with seed
weight (g m-2) at R8. At R5 specific leaf weight(g m-2) and specific leaf area
(m2 g -1) were poorly correlated with seed weight (g m-2) at R8.
Leaf area index and percent light intercepted at R3, were highly
correlated with seed weight (g m-2) at R8 (Table 5-7). Leaf area index and
percent light intercepted at R5 were highly correlated with seed weight (g m-2) at
R8 (Table 5-8). This is consistent with the findings of Basuchaudhari (1987).
However, Ramseur et al. (1985) and Costa et al. (1980) found that LAI was not
correlated with yield.


Yield implications


At R3 with increasing plant population density the stem fraction increased.
However, at R5 the plant population effect on the stem fraction was not
significant. Specific leaf weights tended to be poorly correlated with seed dry
weight.
Bunce (1988) indicated that even though LAI increased with plant
population density, the photosynthetic capacity per unit leaf area decreased,
because of the reduced photosynthetic capacity of shaded leaves. The lower
respiration rate of shaded leaves results in them not being parasitic (Shibles and
Weber, 1965; Duncan et al. 1967). However, it is likely that the increase in leaf
photosynthetic area per unit land area, more than makes up for reduced
photosynthetic capacity per unit leaf area, thereby, increasing the potential for
increased yields with increased plant population density.








Nutrient Status and Linear Regression.


At R5, plant population density had no effect on N concentrations in the
stems and leaves (Fig. 5-9 and Table 5-9 ). However, N content (g m-2) in the
stems and leaves were affected significantly (Fig. 5-10, Table 5-9 ). At R5, plant
population density had no effect on P concentrations in either stems or leaves (
Fig. 5-11 and Table 5-9 ). Phosphorus content (g m-2) in the stems and leaves
were affected significantly ( Fig. 5-12 and Table 5-9). These results corroborate
those of Abrahamson and Caswell (1982) and Hanway and Weber (1971).
Unlike the findings of Buttery (1969), N concentration did not decrease because
of inter-plant competition for limited N resources. Nitrogen (g) per unit leaf
area ( Fig. 5-13, Table 5-9 ), and P (g) per unit leaf area ( Fig. 5-14, Table 5-9)
were not affected by plant population density. Leaf N (g m-2) appeared to be
greater in the leaves than in the stems, (the intercept and slopes of the statistical
models for N (g m-2) and P (g m-2) imply this) and this observation is consistent
with the findings of Wang and Liu (1987).


Nutrient status and correlations.


The correlation results are provided in Table 5-10. The concentrations of
N in both stems and leaves at R5 were not correlated with seed dry weight
(g m-2) at R8. The concentration of P in both stems and leaves at R5 were not
correlated with seed dry weight (g m-2) at R8. N and P content per unit leaf area
was not correlated with seed dry weight, and therefore were not a basis for






















60


50 -


40


30


20


lu 1 I I I I I I i
0 10 20 30 40 50 60 70 80
Plants m-2



Fig. 5-9. The effect of plant population density on leaf and stem
N concentrations of Kirby at the R5 stage, 1988.


----- Leaf
--- Stem