Diss. ETH No. 9791
AGRONOMIC AND PHYSIOLOGICAL ASPECTS OF
POSTFLOWERING DROUGHT TOLERANCE OF PEARL MILLET
(PENNISETUM GLAUCUM (L.) R.Br.)
IN THE SAHEL
A dissertation submitted to the
SWISS FEDERAL INSTITUTE OF TECHNOLOGY ZURICH
for the
degree of
Doctor of Natural Sciences
presented by
PETER BIELER
Dipl. Ing. Agr, ETH Zurich (Switzerland)
born March 7, 1963
citizen of Solothurn
accepted
on
(SO)
the recommendation of
Prof.Dr.P.Stamp, examiner
Prof.DrJ.J.Oeilli, co-examiner
Zurich, 1992
L'etranger
ne
voit que ce
qu'il sait.
Proverbe afncain
Contents
1.
Summary
1
Resume
3
2. ICRISAT
-
Sahelian Center
6
3. General Introduction
3.1.
3.2.
4.
8
Drought Stress
Drought Stress
Postflowering Drought
in General
9
in Pearl Millet
Stress
11
Screening in
West Africa
13
4.1. Introduction
13
4.2. Materials and Methods
14
4.2.1. Location and Soil Characteristics
14
4.2.2. Plant Material
14
4.2.3.
Experimental Design and Irrigation Treatments
15
4.2.4.
Crop
16
Treatments
4.2.5. Observations and Measurements
4.2.6. Statistical
16
Analysis
17
4.3. Results
20
4.4. Discussion
29
5. Grain Growth Under
Postflowering Drought Stress
32
5.1. Introduction
32
5.2. Materials and Methods
33
5.2.1.
Experimental Design
and
Crop Management
33
5.2.2. Observations
5.2.3. Data
33
Analysis
34
5.3. Results
36
5.3.1. Grain Growth Under Well-Watered Conditions and
Drought
Stress
36
5.3.2. Grain Growth and
Drought Response
44
5.4. Discussion
6.
Carbohydrates
Reserves
45
During Postflowering Drought Stress
49
6.1. Introduction
49
6.2. Materials and Methods
50
6.2.1.
6.2.2.
Experimental Design
Sampling and Processing
6.2.3. Chemical
Analysis
50
of Plant Parts
of Nonstructural
51
51
Carbohydrates
6.3. Results
52
6.3.1. Yield and Yield Related Parameters in the Yield Plot
6.3.2. Total Nonstructural
6.3.3.
Carbohydrates
6.4. Discussion
52
Carbohydrates
Under Unlimited Versus Limited
53
Tillering
.
.
59
62
7. Further
Aspects
of Plant
Development
After
Flowering
66
7.1. Introduction
66
7.2. Materials and Methods
67
7.2.1.
7.2.2.
Experimental Design
Vegetative Growth Observations
68
68
7.2.3. Root Growth Observations
7.2.4. Neutron Probe Measurements
69
69
7.3. Results
7.3.1.
67
Vegetative
and Generative
mance
Development, and Yield Perfor¬
69
7.3.2. Root Growth
74
7.3.3. Neutron Probe Results
80
7.4. Discussion
81
8. General Discussion and Conclusion
85
9. References
89
Annex I
97
Annex II
99
Annex III
102
Annex IV
103
Acknowledgement
Curriculum Vitae
Table of Abbreviations
DAS
days after sowing
DRI
drought response
EF
early flowering
FL
time to
index
flowering
GFP
grain filling period
GGR
grain growth
ICRISAT
International
rate
Crop Research
Institute for the
Tropics
ISC
ICRISAT Sahelian Centre
LAI
leaf
area
LF
late
flowering
LGGP
linear
LP
lag phase
index
grain growth phase
M
maturity
SAT
semi-arid
TNC
total nonstructural
tropics
carbohydrates
Semi-Arid
1
1.
Summary
Pearl millet
Sahelian
(Pennisetum glaucum (L.) R.Br.)
farming
subsistence
the most
as
faces various
biotic
and
constraints. One of the most cited and also most feared is
stress, due to the
Grain
has been
filling
drought
irregular
stress in
programs. Therefore
previously described
step is
for their
reliable
screening technique
effect of
drought
being developed
drought stress,
approach
to
identify parameters responsible
(drought tolerance) by multiple regression
for their individual
drought response
Three hot off-season trials on the
International
Sadore
Crop
(Niger,
is
a
and to characterize
millet
genotypic
are
West
Africa)
were
origin.
differences in
probably important for
in the literature. An
positive drought response
to enable a classification of
experimental
site of the Sahelian Center of the
carried out
including
Grain
under
yield
a
Tropics (ICRISAT)
large
drought
stress was reduced
50%
individual
panicle grain yield under postflowering drought conditions
to the
in
number of millet
by almost
compared
genotypes
suggested.
Research Institute for the Semi-Arid
genotypes of West African
postflo¬
large scale breeding
A first
reported
for
region.
developed.
response, and to look for
tolerance. Various attempts have been
in
has to be
agronomic and physiological parameters which
drought
postflowering drought
the most sensitive stage to
being
as
crop in
production
For this reason it is necessary to include
in new varieties
to describe the
genotypes
various
a
abiotic
inter- and intra-seasonal rainfall pattern in the
pearl millet.
wering drought tolerance
important
irrigated control. Correlation analysis suggested
as a
viable
parameter for selection of drought tolerant genotypes. Inconsistency of the
drought response
over
the three years data could be reduced in
simple stability
analysis.
Individual
panicle yield
investigated.
In a
by grain number and grain
a
large
grain filling characteristics
their reaction under
dry
determined
grain growth study
examined, for their
of
as
postflowering drought
matter accumulation in
grains
size was further
number of millet genotypes
under
were
optimum conditions; and for
stress. It could be shown that the rate
under
drought
stress remained
unchanged.
2
grain filling period
However the
grain
size.
panicle
Additionally grain
studied for
significantly reduced, resulting
abortion after
stress. The
drought
under
was
physiology
possible genotypic differences
grain filling. Therefore,
water soluble
of
grain
reduced
flowering
in
smaller
a
number per
contrasting genotypes
in the assimilate
carbohydrates
was
supply responsible for
were
in three
analyzed
genotypes under well-irrigated and under postflowering drought stress conditions.
75% to 100% of the total
anthesis
dry
matter accumulation of the whole
explained by dry
was
matter increase in the
genotypes with different drought response showed
supply
conditions. No change in
well-irrigated
under
occurred from before anthesis until the end of
assimilates for
grain filling
came
stress the abilities to maintain
drates
important
were
to
partitioning carbohydrates
from current
photosynthesis
maintain
were
panicle.
differences in
no
plant
carbohydrate
carbohydrate
reserves
grain filling. Therefore
Under
photosynthesis.
after
The three
and to mobilize stored
most
drought
carbohy¬
grain filling. Genotypic differences
identified but could not be
in
adequately explained
with the available information.
Further aspects of
plant growth under drought
performance
well-irrigated conditions. Tillers
% to total
under
grain yield
panicle
was
Rooting pattern
give
wilting point
the
at
sandy
of
of
in tiller
to the
found to contribute 33
control and 21 %
flowering
only
under
panicles and
drought
in the main
for
flowering,
and the utilization of soil water after anthesis did not
pearl
of
grain yield under postflowering drought
genotypic behavior due
millet
to environmental
stress.
variability.
The
between 6% and 7% volumetric water content in
was
soils of the research station.
Although postflowering drought tolerance
in this
were
compared
important
precise idea
a
well-irrigated
Synchronization
stress conditions.
stem
under
stress were
study
it
was found that
an
eventual
to be done in a
drought nursery.
criteria
with
together
a
of
pearl millet could
screening
Individual
of
not be
genotypes
panicle grain yield
stability analysis provides
a
has
quantified
necessarily
as a
basis for future
selection
breeding.
However, the physiological mechanisms of grain filling, particularly influenced by
the
photoperiod sensitivity
investigated
for
pearl
millet.
of
dry
matter
partitioning,
needs to
be
further
3
Resume
Le
mil
(Pennisetum glaucum (L.) R.Br.),
I'agriculture
de subsistance
production de
biotique
nature
craintes est la secheresse
saisons et
decrit
la
le
comme
abiotique. Une
et
due
post-florale
plus sensible
plus importante culture
diverses contraintes
rencontre
des
plus citees
I'irregularite
a
sein des saisons. Le stade du
au
du mil. On
Sahel,
au
point.
Une
premiere
demarche
de la secheresse et
a
secheresse,
decrire des parametres
de
et
responsable des
a
regressions multiple; elle
des
reactions
specifiques
Trois essais
en
a
un
a
conduit
saison-seche chaude
grand nombre
50%
de
de
grains,
en
a
Sadore
a
secheresse
partir
aux
sur
doit
la
a
physiologique
et
de
la secheresse. Les
pourraient
des donnees
a
identifier les parametres
(resistance)
a
ete faite a
partir
classer des genotypes selon leurs
le site
sur
experimental
du centre sahelien
les Cultures des Zones
de
I'Ouest)
mil provenant de
temoins. Une
grains par panicule
Des recherches
criblage
Tropicales
ont ete menes avec
I'Afrique
de I'Ouest. Le
dans des conditions de secheresse, a diminue de presque
post-floraison peut servir
resistants
a
(Niger, Afrique
genotypes de
comparativement
rendement
de
decrire les effets
la secheresse.
a
(ICRISAT)
rendement en
a
la secheresse
a
sure
consiste
agronomique
la secheresse
de I'lnstitut International de Recherche
Semi-Arides
ete
reactions des genotypes differents. Differents essais ont ete
reaction positive
une
deja
classer des genotypes de mil, selon leur reaction
decrits dans la litterature. Une demarche visant
permettant
a
cycle de developpement
post-florale dans programmes de selection. Une technique
done etre mise au
plus
entre les
pluies,
done realise la necessite d'inclure la resistance
a
la
a
et des
remplissage des grains
la secheresse dans le
a
des
dans
etre
sur
a
montre que le
problemes
a
de variation de la reaction
travers
une
analyse
de stabilite
la
a
simple
a
ans.
plus poussees
fonction du nombre des
correlation
parametre pour selectionner des genotypes
expliques
trois
analyse.de
individuelle dans des conditions de secheresse
grains
ont ete faites sur le rendement par
panicule
et de leur masse. En suivant 1'evolution
en
ponderale
4
des
grains,
on a
modalites de
pu caracteriser
un
grand
nombre de
dans leurs reactions
a
la secheresse
notoirement reduite
avec
au
duree de
physiologique
remplissage des grains
est
moins gros. De
en
grains
grains apres la floraison
Ceci
a amene a
analyser
solubles pour trois genotypes dans des conditions
les
totale de matiere seche de la
I'anthese, etait due pour 75
100%
du
developpement
des
reactions differentes
a
pas varie
jouant
un
courante.
La
quantite des
pendant
le
role dans le
Les
assurer
le
remplissage
capacites
des
en
fonction de leurs
hydrates
des
remplissage
grains
la
des assimilats
de la
photosynthese
sous
conditions de
photosynthese
de carbone accumules sont
et pour que le
genotypes dans leurs capacites
grain puisse
de
expliquees de
importantes
concurrencer
ont ete decelees au
repartition des hydrates de
maniere
appropriee,
plante
sous
les
des conditions de secheres¬
se, ont ete examines et compares a la situation dans des conditions
a
avec
disponibles.
D'autres aspects de la croissance de la
II
qui
plupart
respiration. Des differences
etre
ce
de carbone avant I'anthese n'a
grains proviendraient
maintenir la
a
carbone, mais n'ont pas pu
informations
d'hydrates
remplissage des grains. Done,
d'autres facteurs comme la
niveau des
plante entiere, apres
de carbone, dans de tres bonnes conditions
reserves
secheresse et de mobiliser les
pour
et de secheresse
de la matiere seche du fait
panicules. Trois genotypes choisis
hydrates
lors
de carbone
hydrates
la secheresse n'ont pas montre de differences pour
est de la devolution des
d'irrigation.
('augmentation
a
reduit le
photosynthates
d'irrigation
post-floraison. L'accumulation
a
en
plus,
etudie pour deceler d'eventuelles
a ete
niveau des genotypes pour la devolution des
remplissage des grains.
du
grains demeure inchange dans des
pour consequence, des
de secheresse, I'avortement des
nombre. Le contexte
II a ete decouvert que le taux
post-florale.
Cependant la
conditions de secheresse.
differences
mil dans leurs
remplissage des grains sous des conditions sub-optimales ainsi que
d'accumulation de matiere seche dans les
cas
genotypes de
d'irrigation.
ete constate que les talles contribuaient pour 33% au rendement total en
grains dans des conditions sub-optimales
et pour 21% seulement dans des
conditions de secheresse. La simultaneity de floraison entre les
talles et celles de la
tige pnncipale
est
importante pour le
panicules des
rendement
dans des conditions de secheresse. L'examen du systeme racinaire
a
en
grains
la floraison
5
et ('observation de I'utilisation de I'eau apres I'anthese n'ont pas donne une idee
precise du comportement des genotypes
point
de fletrissement pour le mil
Bien que la resistance du mil
dans
necessite de
mener un
de selection ainsi
dans
remplissage
la
repartition
des
avancees.
a
ces travaux de
avec
a
I'environnement. Le
7% de contenu
le
la secheresse
post-florale n'ait pas
recherche, ceux-ci ont
en
eau
en
Cependant
future.
grains
sous
sous
grains par panicule individuelle
qu'une analyse de stabilite constitueront
les
mecanismes
influence notamment de la
de la matiere seche, doivent faire
I'objet
pu etre
de degager la
permis
criblage de genotypes dans une pepiniere
de secheresse. Le rendement
selection
relation
dans les sols sablonneux de la station de recherche.
volumetrique
quantifiee
se
en
situe entre 6
conditions
comme
une
critere
base pour la
physiologiques
photosensibilite
de recherches
du
et de
plus
6
2. ICRISAT
•
Sahelian Center
The present studies
carried out at the
were
Research Institute for Semi-Arid
Africa. The
experimental
southwest of Niger
near
6 km west of the river
which is
currently
ISC is situated
slope.
(see
the town of
Niger.
plateau
on a
The
rainy
The rest of the year
mm
(Niamey)
of 32°C in
primary
food crop is
grown
grazing
of
as a
cowpea
and
slight
north-south
the soils in this
area are
low
fertility and
organic
is
dry,
is characterized with an
irregular convective
storms.
winds
bearing
often with
dry
a
(350
are
-
600
mm
vegetation
a
characterized
by
lower
second of 34°C in December and
population
pearl
maxima of up to 44°C occur
periods
and
are warm
millet
are
all
during
daytime
January.
annual rainfall) climatic zone,
of grasses,
a
thorny bushes and few
involved in subsistence
agriculture.
(Pennisetum glaucum (L.) R.Br.) often
(Vigna unguiculata L). Groundnuts (Arachis hypogaea)
cash crop
mainly
in the southern
part of the country. Pastoral
sheep, goats and cattle is the other major agricultural branch. Almost
all field work is done
practically
Daytime
two
characterized by
intercropped with
are
capital Niamey
(called 'Harmattan'). Temperatures
August
trees. More than 90 % of the
The
in the form of
(October through May)
average of 29°C.
area
September
from June to
ISC lies in the sahelo-sudanian
semi-arid
tropics (SAT)
often shallow, with low
season
March, April and May, while
one
45 km south of the
included by the Center is 500 ha, half of
at about 240 m elevation with a
dust from the north and east
maxima,
Say,
area
found in the semi-arid
average rainfall of 561
an
Niger, West
4.2.1).
also
The four months
year with
in
site is situated at Sadore at 13° 15'N and 2° 18'E in the
yellowish-colored sandy soils,
matter
(International Crop
Tropics) Sahelian Center (ISC)
cultivated.
frequently
As
ICRISAT
not
by hand labor. Animal traction
existing.
is rare, mechanical traction
7
The institute's
objectives
of
and
grain yield
are
quality
(1)
to serve as a world center for the
including: pearl millet, sorghum, groundnut, chickpea
develop improved farming systems; (3)
to
production
in the semi-arid
and
transfer of
technology
research
as
improvement
of ICRISATs mandate crops grown in West Africa
tropics (SAT);
well as
training.
identify
(4)
and
pigeonpea; (2)
constraints to
to assist the
to
agricultural
development and
All of them include intensive basic
including the collaboration of the National Research Programs.
8
3. General Introduction
Pearl millet
secondary
It is
(Pennisetum glaucum (L.) R.Br.) originates
center of varietal diversification in Asia
tillering C4-plant,
a
80 cm
better
1.5
-
5
Millet is grown in
long.
m
(Rachie
tall. The inflorescence is
areas
with
a
large
range of
genetic diversity
ranges between 200 and 1200
is of excellent
in
on
a
Majmudar 1980).
dense
spike up
under
to
a
dry conditions.
in West Africa. Grain
yield
rainfall conditions. The
grain
grain yield
kg/ha depending
in terms of nutrition with
quality
a
rainfall of 125-900 mm, and has
growth than sorghum {Sorghum bicolor{L.) Moench)
It has a
from Africa with
and
protein
a
content of 11-19%
(Kumar 1989).
Pearl millet is the
Africa.
predominant cereal of the drought-prone Sahelian
Forty-nine percent
in West Africa
(12.2
of the world's cultivated
Mio ha in 1986, FAO
zone
of West
pearl millet
is located
production yearbook), mainly
in Burkina
area
with
Faso, Chad, Mali, Niger, Nigeria and Senegal. It is the primary food crop and the
basis of subsistence
The Sahel has
a
farming
very low
tion and
increasing
improve
this
soil
situation.
constraints to millet
not
growing
published).
season
high inter-
agricultural potential (Sanders 1989).
fertility
ISC
More
as
Centre)
drought (30%),
(15%),
soil
insect pests
Pearl millet in the SAT
principal ways
scientists
to
estimate
fertility (20%), Striga
(10%) and diseases (5%)
(Semi
Arid
Tropics) faces high
a
intra-seasonal variation of amount and duration of rainfall
Annual rainfall in this
below average since the late 1960s
rainfall
grain filling period
declined by
much
as
zone
(Forest
during August when the
water for the
as
to be the two
Sahelian
Water conserva¬
temperatures and very frequently erratic rainfall resulting in
and
importantly,
ly, pearl
perceived
(ICRISAT
bird attack
(Sivakumar 1986, 1989a).
persistently
are
production
(parasite weed, 20%),
(data
in these countries.
soil
(300-600mm)
has been
1982; Sivakumar
1989b).
profile normally fills, storing
of the crop
during September/October, has
40% in some locations
(Sivakumar 1989b). Consequent¬
millet grown in the Sahelian zone is now
frequently subjected
to
drought
9
during grain filling. Therefore cropping systems
capable
of
3.1.
with this variation,
coping
of the crop
are
given
with
a
yield
in General
in a crop is the
environment. Effects of
yield reduction, the
expression of
drought
its
is taken
single aspect.
to be
is
generally
expressions. Therefore
as
plant generally responds
following major accepted
era/.
drought
1987; Sagar
et at.
to
drought,
drought
mechanisms
and
as a
be
resistance
reported
a
and
as
response
agronomic
agronomic parameter
an
selection tool.
as a
by invoking
in
one
of the
1981):
drought
escape
escape mechanisms. The
one
Earliness however is
Time to
flowering
by
or
one
negatively correlated to yield
determines the
two genes
controls the
length
length
of the
only (Appa Rao
et al.
of the
developmental cycle
flowering. Drought escape
add to
drought
is often and
yield consistency
'resistance'.
of two
important drought
plant reaches maturity before drought stress affects its
generative development and therefore escapes from
can
in terms of
since
1984)
helpful
constraint
a
(Ugherughe 1987; Annerose 1988; Jones
Earliness, controlled by the time of flowering, is
it
drought
physiological
might
in
expression of drought
to treat
keen interest still exists to find
that covers crucial responses to
A
et al.
an
resistance has been
combination of various
a
a
for
breeding
extremely difficult (O'Neill
drought
to
However
genetic yield potential
mostly expressed
stress are
extent of which
susceptibility. Screening programs have often tried
one
high development plasticity
required.
Drought Stress
The observed
together
and crop cultivars that are
the effects of late
drought.
in years with
adequate rainfall.
growing period
and is controlled
1988). Photoperiodic sensitivity
of
a
genotype and
easily exploited
in
so
the time to
breeding programs
but is often used under the
misleading
name
as
of
10
A second
drought escape mechanism
phenology during
plant makes up
main culm with an increased tiller number
drought
stress
Such
period.
or
drought escape
breeding
programs in
sensitive
developmental stages.
developmental plasticity
concerns
mid-season stress. The
tiller
for the
yield, produced after the
mechanisms
'avoiding' possible harmful
yield
in
loss of the
be included in
can
effects of
dry periods during
drought avoidance
Drought avoidance is
high tissue
(e.g.,
water
avoiding tissue dehydration by maintenance of
potentials. This
can
be achieved due to restricted water loss
stomatal control; reduction in the amount of radiation absorbed
movements;
a
based on
change
better water supply
in
evaporative
by
an
surface
area
through
leaf
through reduced leaf area)
extensive root system (e.g., increase root:shoot
or
ratio).
drought tolerance
The third mechanism is
drought
tolerance and is
of turgor pressure at low tissue water
osmotic
adjustment,
however is
dependent
apparently possible
on
which is
species
in all
drought
expressions
is often used in
general
In the literature
genotypes
following
yield potential (yield
a
a
physiological
'drought tolerance'
'resistance' for
an
abiotic
appropriate.
description of drought response
measurements included the ratio of
under
optimum conditions)
threshold value with the average
to describe a stress index
adjustment using
protoplasmatic
humidity.
studies the term
for the
(Bruckner
correction factor per
day
not
only dependent
includes
cells to air
and does not seem
physiological
imposed and is
to describe a range of
come
practical approaches
the calculation of
plant
sense, since the term
quantified
labor intensive
under stress to
of
as
plant cells. This
It is therefore not
tolerance
drought
dehydration
tolerance has
a
accumulation of solutes in
Further
to water deficit. In the
constraint cannot be
avoiding
a
an
maintenance
This includes mechanisms
the rate of which stress is
physiological stages.
varieties.
or
resistance that allows
However
on
potentials.
expressed in the
and
yield
of
a
yield
genotype,
or
of a number of
Frohberg 1987a). Yield
advance in
flowering
was
used to
11
calculate
drought susceptibility index
a
experiments (Fischer
Maurer
and millet
1978)
highly significant
Drought Stress
3.2.
Generally, drought
intensity
(Bidinger
stress affects
as
demonstrated for
millet reduced
pearl
developing
1987).
of tillers.
tiller
Seemingly
is the
crop
in tiller number and
stress this
described
yield
due to tillers
did not
occur
yield components of millet
by
Mahalakshmi and
flowering. Delay
resulted in
affected
panicle
but
development
an
achieved
increased
a
grain
panicle.
For terminal
and
after mid-season
as
was
An
days
by both
area
yield
number
et
therefore
to
increase
an
(postflowering)
drought
were
important factor
panicle initiation
and biomass. Grain
both main stem
the most sensitive
found that millet
of
Bidinger 1986).
tillers) thus the harvest index
flowering decreases
postflowering drought
(1988)
panicle,
resulted in no difference in final
was
Bidinger (1985b).
leaves, tillers, leaf
Mahalakshmi ef al.
during panicle
develop
can
time and
yield
was
further
was
and
initiation of main stems due to mid-season
(since compensated by
Stress after
makes
more
in
the
for
reason
(Mahalakshmi
stage of the crop's development expressed
to
drought
by Oertli (1988).
due to mid-season stress
number per tiller
grain
compensation
Growth and
had
(Mahalakshmi and Bidinger 1986; Mahalakshmi
grain yield
yield. Compensation
on
main stem
during
stress
compensated by changing tiller performance and
of
wheat
water deficit
panicles; this way tillers
to the main shoot
loss of
spring
grain yield of the
panicle (mid-season stress)
The
(Fischer and
wide range of individual
(1985a) found that
the main stem
al.
a
plant development dependent
grain yield
comparable
across
potential.
could increase
of the later
wheat
in Pearl Millet
Mahalakshmi and Bidinger
of
consistent
was
1987a) grain yield under drought
showing
wide selection
a
of water deficit
development
et al.
cultivar differences
response and therefore
for wheat, which
1978). Barley (Blum 1989),
and Maurer
the
days
drought
was
not
reduced.
yield and tiller yield which
stage
grain yields
of millet
were
development.
linearly reduced by
12
terminal
(postflowering)
dependent
intensity),
was
stress with
increasing intensity of
lower for
grain yield. This shows
phenology
Bidinger
ering
the
(1987a) reported
yield
of the
a
timing
0.9% reduction in
of stress relative to
maturity groups planted in
a
reduction in
pearl
millet. A
under stress and the time to
grain yield
a same
after a
potential and
time to
flowering
grain yield under
postflow¬
significant correlation between
flowering
was
observed. This fact
account for more than 50% of the total variation
stress.
Little research has been carried out to understand the mechanisms of
tolerance in
pearl
millet. The purpose of the present
breeding strategy including drought
measured parameter with
a
allows a breeder to select
agronomic
for the
field
and
is rather
probably
investigations
high genotypic variability
a
large
drought
was to test a
criterion. Therefore
as a
an
easily
had to be identified that
number of genotype entries. Furthermore,
be
investigated that
the chosen parameter. The
experiments including
genetic material
increase the
tolerance
physiological aspects should
expression of
ments are
trial.
possibility of drought escape. Bidinger et al. (1987a) found that yield
indicates the
in millet
day caused
one
importance
that
stress can reach up to 50% in
relative
by
for genotypes of different
et al.
(yield
flowering
grain number and grain size the later the stress occurred.
Advance in time of onset of stress
relative
stress. Yield reduction
time of onset of stress in relation to time to
on
West African
pearl
are
presented results
responsible
come
from
millet genotypes. Since most of the
heterogeneous, sophisticated physiological
measure¬
very difficult to interpret. Therefore the main purpose
was to
knowledge about morphological growth characters under controlled
well-watered conditions
compared to postflowering drought stress conditions
to make use of the information for a
description
of
drought
tolerance.
and
13
4.
in West Africa
Postflowering Drought Stress Screening
4.1. Introduction
It is
highly desirable
being developed by
to include
a
postflowering drought tolerance
breeding program
of
pearl
drought
availability
of
have to
be
technique. Specific agronomic parameters
genotypes in
expression
time of 50%
lines of Indian
potential
flowering
differences in
for
of
ef al.
required
pearl
(DRI)
a
consistency
of
of the defined
materials
drought
able to show varietal
an
analytical technique
DRI to
found
exclusively
for
agronomic parameters suggested
and
to test for West Africa the
analysis developed by Bidinger
of West African
response of individual genotypes
are
to calculate a
multiple regression
postflowering drought-screens. The
chapter were undertaken
screening parameter
were
specific drought response
from the residual of this
postflowering drought screening method
(1987b), using
were
yield potential (yield under well-watered
relationships of the
in this
they
defined genotype
of them as viable to be used in
experiments described
(1987a) proved
millet. However, such differences
1987b). They developed
index
Using
from the
flowering and
time to
each genotype. The
al.
screening
identified to select
the
et al.
of 50% of the varieties,
grain yield
drought response
some
have the
breeding process. By withholding irrigation
origin Bidinger
a
to be a result of the combination of
(Bidinger
viable
of interactions of many parameters in stress environments.
selection
conditions),
a
breeding program. Postflowering drought tolerance is a phenotypic
a
mostly millet
large
in response
genetic variation
identifying plants which
and also to find reliable means of
desired attributes. This necessitates the
varieties
new
millet. However before this can
be started it is necessary to establish that there is
to
into
discussed.
origin.
as
well
et
Further more,
as
the
stability
14
4.2. Materials and Methods
4.2.1. Location and Soil Characteristics
Three field
during
experiments
the hot dry
were
seasons
conducted at the ICRISAT Sahelian Center (ISC)
(February-May)
is rainfree and characterized with
high potential evaporation
The trials
rates
carried out
were
depth. ISC's soils belong
high
1988, 1989 and 1990. The period
in
mean
temperatures (30°-32°C) and
air
(6-15 mm) (see Annex I for detailed information).
on a
yellowish loamy sand soil of
characterized by 88-91 % sand, 4.7 % silt, 4.2-7.6 % clay and
4.9. The laterite
capacity
to 1 m
of the fields
occurs at
depth
are
more
than 3
to the Alfisols (Labucheri series after West era/.
2-4
is 120
provided by
m
mm.
depth.
an
m
1984),
average
pH
of
The water content in the soil at field
Typical basic chemical analysis for pH and
the trial 1990 with soil
samples
P
taken at harvest
(Table 4.1). After the recommendations of Bationo etal. (1991) phosphorous was
available in
mg K
excess.
kg"1.
following
The average of ISC's soils have 123 mg N
The trials
any rotation
were
planted
on
the
same
kg"1
soil and 39.1
field without any cover crop
plan.
Table 4.1. Chemical soil properties, Sadore 1990.
Depth (cm)
0-20
20
pH (H20)
5.4
4.7
4.7
P(ppm)(Bray 1)
24.2
13.3
7.4
-
40
40
-
60
4.2.2. Plant Material
A total of 42
They
(1988) and 45 (1989 and 1990) genotypes
included advanced
Burkina Faso, Mali,
incidence
Niger, Nigeria
(e.g., downy
were
planted
in the trials.
breeding lines, released varieties and landraces from
mildew
on
and
Senegal (Annex II). Due
the dwarf
types),
to
high disease
bird attack and extremes in
15
34 genotypes in 1988 and 1989 and 32 genotypes in 1990
flowering date, only
included in the final
were
not exceed 10
in
analysis.
days. This interval
The range of
was
flowering of these genotypes did
considered to be the
likely maximum range
flowering among breeding materials intended for a specific climatic
The
zone.
entry composition of the three experiments varied between years. Six genotypes
consistent in all three
were
experiments (Table 4.2).
Table 4.2. Common genotypes ot the trials in 1988,
1989 and 1990.
4.2.3.
The
Genotype
Origin
ICMVIS 85321
ISC
ICMVIS 85332
ISC
HKP
INRAN1
SYNTH-1
Mali
CIVT
INRAN
P3Kok3
INRAN
Experimental Design
experimental design
treatments
was
replicated thrice
plot. Location
main
of
and
and
modified
a
m
sating
actual
of
evapotranspiration)
(crop
Sub-plots
and
regular irrigation
development
calculated
with main
varied between years,
a
on a
a
well
postflowering drought
4 to 7
by summing up the pan evaporation
11nstitut National de Recherches Agronomiques du Niger
error
m
due
length,
stress.
day cycle depending
was
done
Both
on
applied per inigation
of the number of
unity). Irrigation
any
of 5
rows
irrigated control (compen¬
of the crop. The amount of water
coefficient assumed to be
plots irrigation
twice within each
excluding
consisted in 4
apart. Irrigation treatments consisted of
treatments received
stage
split-plot,
sub-plots genotypes replicated
main-plots
to residual soil-water in the field.
0.75
Irrigation Treatments
the
was
days preceding
by
a
line-source
16
overhead
drought
4.2.4.
stressed
Crop
Farmyard
320
sprinkler.
just before flowering, affecting both irrigation
of
manure
kg organic
pH (H20)=8
at the rate of 40
P
kg
ha"1,
K ha"1 were broadcast
to
planting
tractor drawn cultivator. A further dose of 26
kg
of N
kg
sowing (DAS)
(Nematicide)
and 1989
was
and
The resultant
incorporated
applied
only. Seeds
ridges
1990 on
by
high plant population
kg
ha"1
a.i.
plants per
were
ha"1
N ha'1 and
kg N ha'1, 19.6 kg
and
incorporated with
was
applied
18
days
plough. Carbofuran
drawn
at the time of
hill 0.4
of 100'000
sowing
in 1988
no
m
apart at 11 -14 DAS.
plants ha"1 (traditionally 30'000
enhanced
controlled both
cultivator) and manually. Birdscaring apart,
after
donkey
depletion and
soil water
Weeds
a
kg
machine in 1988 and 1989, and by hand in
and thinned to three
withholding irrigation.
with
at the rate of 4
were sown
plants ha"1) hastened
required
treatments.
120
matter and mineral fertilizer at the rate of 45
prior
after
was
Treatments
P ha"1 and 37.4
a
problems in irrigation, the 1990 trial
Due to technical
drought
stress after
mechanically (donkey-drawn
other disease
or
pest control
was
planting.
4.2.5. Observations and Measurements
Only the
and 2.8
middle two
m
(4.2 m2)
rows
of 2.4
m
length (3.6 m2)
in 1990 were harvested at
utilized for other observations
plots
were
plots
between the three years. Time to
when
stigmas
had
emerged
on
in 1988, 2.0
to four
plots simulating
an
(3.0 m2) in 1989
maturity (yield plot). The
resulting
flowering
rest of the
in different sizes of the
was
determined
on a
days before
early ending
the
for the stress treatment
likely 50% flowering average
were
yield
was
of the stress
of the rains. At harvest the number of
panicle weight, grain yield and above-ground biomass
yield
plot basis
50% of all main shoot inflorescences in the
plot (Maiti and Bidinger 1981). The last irrigation
applied three
m
panicles,
recorded per
plot.
All
17
crop and
grain samples
were oven
weighing. Triplicate samples
of 100
harvest to determine the 100
mass as
4.2.6. Statistical
The method
al.
1987b)
grain weight. Number of grains per panicle and per
at the ICRISAT Center in
classify
irrigated conditions
Ya
bY,,
a +
=
where E is random
et al.
+
cFL,
error
(DR,
+
of
c
+
with
(1987b) found
be attributed to variation in
a,b and
on
the
stress conditions
(YpJ,
time to
equation (1)
rm
=
a +
b\
+
DR,
+
is
a
the
function of
flowering (FL,)
and
et
Drought
a
assumption that
(YJ
a
grain
potential
drought
can
(1)
zero mean
and variance
(remaining
(Ya-Yg
=
DR,
and time to
therefore be estimated
can
then be estimated
flowering.
The parameters
by minimizing
the residuals
as:
(2)
cFL,
terms of
+
lY.-rj/o-
yields
is then
equal
to the
equation (1)):
E.
significance of the drought response (DR,)
z-
o.
that about half of the variation under stress could
The difference between the actual and estimated
A test of
according
(Bidinger
to
E,
potential yield
Yield under stress
E).
residuals
India
(DR,):
response
Bidinger
to total
grain yield
Hyderabad,
millet genotypes
This method is based
(DRI).
yield of a genotype under drought
under
the ratio of
analysis
used to
was
as
well as the harvest index were derived from the data.
developed
Index
Response
yield
grains were taken randomly from the bulk plot
panicle yield, threshing percentage
unit area,
panicle
dried at 70°C for at least 24 hours before
(3)
can
be calculated
as:
(4)
18
where
is the standard error of
o
selecting those genotypes
selected to be 1.3,
normal distribution of
Ys.
If Z is
<
A threshold value for Z is
Y'a (equation (2)).
1.3, DR,
in the upper and lower 10% of the
is considered to be
zero.
that the absolute value of the difference between the measured
predicted yield (Y'J
(YJ
and the
Y'a.
In this case
The above derived estimates of E and
*
0. Therefore,
better estimate of E
a
only those genotypes
Y"a
DR,
can
The
=
a +
then be
drought
bYp,
if
| Y.
+
negative
(tolerant
to
susceptible
(DRI)
DRI,
o', then
>
a', then DRI,
o
(o')
E' in
Genotypes
as
with
=
0
=
(Ys
-
and genotypes with
drought
cases
where
be calculated
can
DR,
using
on
(3),
of o' and may have
DRI have
positive
where E' is estimated
a
DRI of
for all
zero are
were
stress treatment to
o'.
positive
a
or
positive drought response
a
with
DRI
negative
a
average in response to
independent of both yield potential
genotypes
by
DR and is defined as follows:
Y'J/a'
multiple
a
a
is based
postflowering drought stress), genotypes
flowering. The DRI
control and
and of
(5)
stress. The DRI as calculated here is
to
drought response (DR^O).
1.3, i.e., for which DR,= 0:
<
<
DR, is, thereby, expressed
value.
(E')
no
cFL, +E'
response index
Y-. |
yield under stress
affected by those
a are
expressed by substituting
(')if |Y„-Y'J
(ii)
for which Z
means
less than 1.3 times the standard error of
was
is considered to have
genotype
a
This
are
drought
and time
yield components in the
correlated to
identify parameters related
to
drought
response.
Measured and calculated
further correlated across
yield parameters
genotype
means
associations between the parameters.
parameters
measured in
independent of the
ters
only
stress
the
to
irrigation
treatments
yield
treatment
under stress with
were
assumed to be
(constitutive), while associations with parame¬
under stress were assumed to reflect the response of the
(adaptive).
These
were
under stress to examine
grain yield
Relations of
well-irrigated
stress effects
in both
adaptive parameters
were
genotype
to
correlated to both time to
19
flowering
and to DRI to separate those
(correlated
to time to
(correlated
to
A
flowering)
from those
expressing
stability analysis
for the six common
showing promising correlations
Eberhart and Russel
a
a
drought escape
drought response
DRI).
genotypes in all three experiments
carried out for the determinants of DRI and also for the
of
due to
relationships
to DRI. The method
was
adaptive yield parameters
applied
(1966). They used the following model
developed by
was
to
study the stability
parameter of T varieties under's' different environments:
Y„
where
=
B,l, +5, (i=1,2,...t;j=1,2,...s)
m +
Y,,
Mean of the parameter of i'h
=
m
=
variety
in
jlh
environment,
Mean of the parameter of all the varieties
over
all the
environments,
B,
=
The
of the i,h
regression coefficient
mental index that
measures
variety
on
the environ¬
the response of this
variety
to
varying environments,
I,
8,
=
=
The environmental index. It is defined
mean
of all the varieties at
mean
with the
deviation from
sum
of all
regression
I,
a
given
=
0.
of the im
as
the deviation of the
location from the overall
variety
at the
jlh
environ¬
ment.
Two parameters of
stability
are
calculated:
a) The regression coefficient, which is the regression of the performance of each
variety
under different
parameter over
This
stability parameter
(1963).
environments
on
the
all the genotypes and is estimated
was
environmental
as
follows:
also described and used
B,
=
by Finlay
means
sum
of the
Y.J/suml2.
and Wilkinson
20
b) The
the
A
square deviations
mean
unexplainable deviations
variety is
significantly
genotype is defined
trait
a
they speak
poorly adapted
For statistical
on
as
the environmental index.
below average
is
one
(b=1)
and the
b>1
resulting
stability for
trait due to small environmental influences. For b<1 the
stability
general adaptability
with above average values
low values in better environments.
possibly
with above average of the
trait, and
with below average values of the trait.
analysis
as
ANOVA, correlation, linear and multiple regression,
MS DOS Computer with the Genstat 5, Release 1.3
Experimental Station, 1988) package
Rothamsted
which represents
(s2d =0).
zero
to have above average
of
regression
regression coefficient
of the trait in poor environments, but
For b=1
regression
different from
Finlay and Wilkinson described
large changes in the
from the linear
from the
said to be stable, if the
deviation not
in
(s2d)
(by
Lawes
was
used.
a
Agricultural Trust,
4.3. Results
Substantial
genotypic variability for yield-related
irrigated conditions (control),
analysis
of variance for
was
observed
This
genotypes.
as
proves
a
ability
wide
of genotypes to
potential
of
significant F ratios
variability
postflowering drought stress. The broad ranges
the individual
traits under controlled well-
of
grain yields
produce grains
under
genotypic expression that
can
from the
maintained under
was
under stress show
drought
be
stress and
exploited
in
a
breeding program (Table 4.3).
The average time to
drought
flowering
drought
flowering
was
significantly reduced
stress, which must be due to
heads. The number of
or
did not
produce
probably
any
in 1988 and 1989 under
accelerated
panicles per m2
stress. This indicates that
flowering stage
an
was
development
significantly
of late
reduced under
the later tillers did not achieve the
grains. Grain yield
under
drought
stress
21
Table 4.3. Means and F ratios of genotypes for growth and yield components measured
the irrigated control (c) and drought stress (s) treatments Sadore 1988 -1990
Variable
Means
Time to
flowenng (days)
m2)
Biomass (g
Stover (g
Panicle
m2)
(g m2)
m2)
Grain yield (g
m2
Panicle No
Panicle
yield (g)
No grains
Grain
panicle'
(g 100')
mass
m2(* 103)
Harvest Index
Threshing Percentage
**
P<0 01
*
Means
F ratio
Means
F ratio
64
5
80"
66
5
92"
67
1190"
s
62
5
40"
65
5 99"
68
15
60"
c
621
3
02"
858
179"
662
2
49"
s
437
2
49"
585
3
28"
598
2
76"
c
398
3
29"
482
191"
431
6
50"
s
300
4
38'
359
4
57"
423
3
91"
378
1
77"
230
0 97
228
2
19"
175
1 41
c
223
2
54"
s
137
2
05"
c
156
2
30"
271
1
66'
158
1 07
s
83
2
48"
140
2
64"
121
0 78
c
98
1
58'
115
3
61"
10 3
1
s
82
147
10 0
2
33"
84
23 7
3 81"
15 0
2
0 72
02"
50'
151'
46"
162
4
s
98
3
00"
13 7
3
23"
142
c
2440
4
47"
3090
3
75"
2260
196"
s
1840
190"
2300
3
16"
2330
0 72
c
0 67
10
39"
077
3
96"
066
5
90"
0 53
7
30"
0 53
2
89"
0 61
2
55"
3
34"
35 3
1
93"
23 5
1 16
c
c
23 6
s
15 4
1
88"
23 3
2
43"
19 6
0 81
c
25 4
4
04"
32 0
3
20"
23 0
2
s
18 6
4
38"
24 0
3
81"
20 0
065
61"
71
1
90"
67
2
60
3 76"
69
0 61
c
70
3
s
58
3 44"
01"
30"
P<0 05
F ratios for irrigation are
significant at 1% level for all parameters except in 1990 for time
dry weight, panicle yield, grain number per panicle, harvest
and threshing percentage (not significant)
flowenng (5%)
index
F ratio
c
s
No grains
1990
1989
1988
in
and for stover
to
22
was
The relative reduction of
(Table 4.3).
1990
compared
to
grain yield
under control,
grain yield under control
(r=0.25).
was
47% in 1988, 48% in 1989 and 23% in
significantly (P<0.01) reduced by
The
yield
loss
in 1988
panicle.
This resulted in
a
an
significantly
a
stress
(P<0.05)
correlated
but not in 1989
(r=-0.05)
nor
panicle yield
reduced individual
again
lower
threshing percentage (ratio panicle dry weight
to technical
problems,
level
grain
to
in 1990
that
number and reduced individual
unsatisfactory irrigation
Therefore the relative
grain yield under drought
grain
panicle grain weight). Due
1990 trial had
(r=0.38)
due to
was
due to both reduced
was
mass
per
to
the control treatment in the
resulting
impact of the drought stress was
slight
in a
not as
high
water stress.
as
in the other
years.
The reduction of the total
primarily
above-ground biomass
due to the reduced
dry weight
flowering
of 2-25%
was
less than for
a
reduced
panicle weight (between
index declined between 13% and 27%
Correlation analysis of yield under
in the
drought
flowering,
as a
drought
harvest index and
significantly
on
40%).
result of the reduced
stress to
threshing percentage
dependent
drought
under
24 and
stress,
as
stress after
The harvest
grain yield.
parameters in the control and
(Table 4.4).
stress treatment were consistent over years
traits and therefore less
was
plant growth
stress was
The reduction of stover
panicle weight (Table 4.3).
showing
drought
under
were
Time to
found to be constitutive
grain yield
under
drought
stress
correlated to these parameters in the control. Total above-
ground biomass and
stover
dry weight in particular were
under stress neither in the control
nor the
drought
not related to
grain yield
stress treatment.
Other
parameters e.g., panicle dry weight, panicle yield, grain number per panicle and
except in 1989 grain number per
stress and are therefore
panicle yield
under
area, turned out to have a
predominantly adaptive
drought
the strongest correlations
stress to
over
grain yield
high response
to
parameters. The correlation of
under
drought
three years and is shown in
tress was one of
Figure
4.1.
23
Table 4.4. Correlations of grain yield measured under drought stress versus yield
parameters in the irrigated control and drought stress treatments Sadore 19881990
Correlation Coefficients
Yield
Yield
(stress) vs para(control)
ter
meter
1989
1988
Time to Flowering
Stover (g
(days)
-0
m2)
Biomass (g
m2)
(g m2)
Panicle
59"
-0
48"
0 17
-0 31
1990
-0
36'
-0
parame-
1990
1989
46"
0 31
-0
35'
-0 28
033
0 23
-0 02
-010
-0 41'
0 06
-0 32
-0 01
0 34'
0 32
0 94'
0
96"
0 72"
013
0 27
0 22
018
0 74
0
51"
0
46"
Panicle
0 02
0 26
029
0 89
0
87"
0
86"
-0 23
019
0 20
056'
0 61"
0
84"
029
0 06
0 26
0 61
054
014
0 24
0
79'
0
85'
0
0 41
087'
0
69"
0 93"
0 33
0 81'
0 80'
yield (g)
panicle'
No grains
mass
No grains
(g 100')
m2(x103)
-014
Harvest Index
Threshing
"
0
*
43'
0 30
043'
%
P<0 01
0
64'
0
50"
the 100 genotypes
response)
and 28
analyzed twenty
genotypes
had
remaining genotypes had
a
a
showed
a
DRI of 0,
more
than +2 1
indicating
yield
yield potential and
time to
adequately
estimated with
drought stress gram yield, panicle yield, gram
average
an
consistent
for all
parameters significantly related
drought
escape
e
g
,
three years
drought
stress could
drought
be
flowenng
number per
panicle and per
(all adaptive parameters) showed the strongest relationships
were
(Annex II) Of
positive value (tolerant drought
under
response
correlations
78"
negative DRI (susceptible drought response).
For latter, genotypes grain
Under
0
96"
P<0 05
Individual DRI values vaned from less than -1 8 to
area
1988
-0 22
vs
Panicle No m2
Grain
The
(stress)
(stress)
(Table
4
5)
to DRI These
A
number of
to DRI were at the same time expressions of
related to time to
flowering (Table
4
5) Biomass and
stover
24
Panicle yield
25 r
(g)
a
l
b)
20
-
**
*#
*
5
t
-
c
*
*
*
0
**
0**
1989.
4.1. Correlations of
*
*
#
%
150
100
50
Grain yield
Figure
stress)
****
%.
200
(g m-2)
grain yield m'2 to panicle yield (both under drought
(a), 1989 (b) and 1990 (c). Sadore 1988-
of genotype averages in 1988
25
Table 4.5. Correlations of yield parameters in the drought stress treatment to
flowering under drought stress ancfDRI Sadore 1988-1990
time to
Correlation Coefficients
1988
Time to flowering (days)
Biomass (g
Stover
0
m2)
Panicle (g
76"
m2)
46'
0
56'
0
63"
0
77"
0 26
-0 21
-0
35'
0
-0 35
-0 28
-0
0 20
0 06
0 74"
0 31
0 69
0 71"
0
0 50
011
010
73"
-0 47"
-016
0 06
0 13
0 22
0 04
0 55'
063'
-0 47'
-0 57"
0 04
0 28
0
36'
62"
yield (g)
panicle'
(g)
mass
(x103)
Harvest Index
Threshing
%
*
P<0 01
dry weight
as
mass as
0
62'
0
88"
0
92"
-0 01
0 22
-0 08
0 28
065
0
0
85"
-0 69'
-0 44
044
0 25
0
82"
-0 60
-0
-0 09
0
044
0
91"
57"
48"
P<0 05
indifferent, harvest index and threshing percentage
rather
adaptive parameters
(within the drought
stress
flowenng
not related to time to
flowenng
grain number
panicle1
a
and to
in
all correlated to time to
to DRI
1988 and 1989
in
significant,
therefore data
are
only
not
attnbutes
presented
in
ones
Grain
yield
parameters panicle
extent grain number
yield
flowenng
As per definition DRI
stress were found to be the best expressions of individual
response No relationship of any of the
rather
Selecting genotypes after
all three years The
some
as
lesser extent individual grain
probably identify rather early
related to
significantly
were
treatment) and also
of these parameters would
DRI was
78"
0
83"
-0 74"
constitutive, and number of panicles m"2 and to
yield,
59"
0
-0 29
66'
1990
-0 02
037"
-0 55'
No grains m2
was
56'
Panicle
No grains
one
1989
0 07
Panicle No m2
Grain
was
1988
-0 07
-
0 47"
0
-0 31
-0
1990
-
053'
m2)
(gm2)
Grain yield (g
1989
-
m2 under drought
genotype's drought
the control treatment to
26
In
Figure
common
that
4.2 the correlation of
although overall
(years),
six genotypes
common to all
and years
(Table
the
Table 4.6.
an
attempt
expressed
as
to
analysis
were
a
be
can
seen
very consistent
drought
over
responses of the
high interaction between genotypes
II).
Drought Response Indexes (DRI) of the
genotypes. Sadore 1988-1990.
No.
Genotype
1
ICMVIS 85321
2
ICMVIS 85332
3
HKP
4
SYNTH-1
1.38
-2.51
0
5
CIVT
0
.0
1.24
6
P3 KOLO
0
-0.05
0
1988
1989
0.69
2.49
0
-0.55
0
0.9
1990
0
-0.29
improve comparability
0
flowering
of the three years, the
date
was
growing degree days (base temperature 10°C, optimum tempera¬
This however did not
stability analysis
as
trials indicated
numbers. It
by
common
ture 33°C and maximum
interaction
to DRI is shown. The six
of the individual
comparison
4.6 and Annex
six
indicated
are
results of the correlation
all three trials
In
panicle grain yield
genotypes of all three years
was
well
as
temperature 45°C;
improve
the
therefore
the
see
consistency
required
consistency
to
5.2.3)
and DRI
of DRI
(data
was
not
parameter panicle
parameters grain
number per
since their utilization
screenings and
for
as a
panicle
presented). A
investigate the genotype
of the three determinants of DRI
flowering, potential grain yield, and grain yield under stress)
recommended
recalculated.
yield under drought
and per unit area
are
as
stress.
x
(time
well
The
not further
year
as
to
the
other
analyzed,
selection parameter is too labor intensive in large
grain number
per
area not
very viable.
27
Drought Response Index (DRI)
4
o
d2
g—BOB—3—Effi-CEJ
3D°
8
S
B-B
o
b)
D
—BIBB—B-Q—fh
rawn
1
gg-
c
—B—BB—B8 1-CBD
16
18
Relationship of panicle yield under drought
stress to
10
12
14
Panicle yield
Figure
4.2.
20
(g)
drought
response index (DRI) in 1988 (a), 1989 (b) and 1990 (c). The vertical line
representing the trial average. Numbers indicating common genotypes between
years (see Table 4.6). Sadore 1988-1989.
28
The
analysis of variance of
grain yield
calculated
drought
genotype
of these
over
Further
x
drought
the three years
stress and
variability.
the parameters
under control and
Time to
(Table 4.7).
grain yield
under
panicle yield
and
panicle grain yield,
time to
stress of the six common
drought
genotypes
flowering, panicle yield
stress all showed
grain yield under drought
year interaction confirming the
flowering,
was
under
high genotypic
stress had a
high
high variability and different expression
parameters of each genotype in the different climatic environment of the
three years. Grain
variable between
yield
under
well-irrigated conditions (potential yield)
was not
genotypes.
Table 4.7. Mean squares of ANOVA for panicle yield (stress), time to flowering, grain
yield (control) and grain yield (stress) over three years for six genotypes. Sadore 19881990.
Source of Variation
d.f.
Panicle
yield
Flowering
(stress)
genotype
Grain
Grain
yield
(control)
yield
(stress)
5
44.2"
82.7"
3040
6207"
2
108.1"
44.7"
176322"
20065"
genotypexyear
10
27.5'
14.8
1106
3097"
Residual
90
8.1
8.3
3000
1570
year
Total
*
P<0.05
107
**
P<0.001
Stability parameters (b=regression coefficient [shaded]; s\=mean square
regression) of panicle yield (stress), time to flowering, grain yield (control)
grain yield (stress) of the six common genotypes. Sadore 1988-1990.
Table 4.8.
deviation from
and
Panicle
yield
Grain
Flowering
yield
(control)
(stress)
Genotype
s2,
b
ICMVIS 85321
s%$?°,,
ICMVIS 85332
;j>.3^V"
0.89
HKP
"llf;*
3.78
0.36
SYNTH-1
t).S$*"
>' 4.25
*
CIVT
%m/^:
1.66
P3KOLO
%m>;:
2.64
s2„
b
s2„
b
Grain
yield
(stress)
b
s2d
2.06
:-$Mr
416
'$$$?£
so
\P$f%ii
1-46
'-'0,9?
349
tfM0:
187
3-1*^
1-47
368
ifj§|S
153
'jjsasri
1.46
,273
%
f§j|^
288
2.55
'fj2|*",=
1.28
:Q;to£:
{''it >£*
tm
Qte'
:1,16
*
tM
0,97,
t1
,
;
--
466
44
f|;|?§|§
iMilf4f
589
360
29
The results of the
stability analysis
stability parameters
are
shown for
of the four
genotypes
suggested by Eberhart and Russell (1966)
any of the parameters.
yield
under
one.
none
for
According to Finlay
well-irrigated
coefficient of about
are
s2,j=u
required conditions of b=1 and
the
a
parameters including the
of the six genotypes achieved
stable
genotypes' expression
and Wilkinson
flowering
and
very unstable between environments. The
between these parameters. For
grain yield
stability
of
a
in
(1963) however, grain
conditions is stable for all genotypes with
Time to
two
in Table 4.8. After the method
under
regression
a
drought
genotype
was
stress
different
panicle yield under drought stress the genotypes
ICMVIS 85332, SYNTH-1, and P3Kolo showed above average
ICMVIS 85321, HKP, and CIVT showed below average
stability
while
stability.
4.4. Discussion
Drought
during grain filling
stress
This was
mainly due
individual
grain
of
Bidinger
et al.
number per
maize
mass
(Grant
et al.
grain filling phase
and
in
resulting
in
a
in a low
stress to
mean a
panicle yield.
was
a
et al.
mass
1981).
that under
early
The reduced
reported to
supply
be
a
to the
findings
A reduced kernel
stress was also
reported
drought
supply
grains
suggested
grain number per panicle
sequence of
grain (Reed
as a
was
result of the
physiological
and
for
stress the
to the
senescence as was
grain abortion after flowering
selective
In maize this is
due to
to an inferior
This agrees with the
the assimilate
reduced, shortening
grain
grain yield substantially.
largely Indian origin.
postflowering drought
rather than restricted assimilate
The lack of
millet reduced
grain numbers per panicle and
with material of
under
smaller
due to
drought stress.
resulting
pearl
1989). These findings suggest
sorghum (Duncan
apparently
to reduced
(1987a)
panicle
of
events
Singletary, 1989).
significant correlations of the relative yield reduction under drought
grain yield under control indicates
better
drought
response. The
yield under drought stress
years. This proves
a
were
that a
relationship
high potential yield
of
does not
yield components
to
different between treatments, but consistent
different pattern of
yield
structure under
drought
grain
over
stress than
30
under control. Therefore
be used to estimate
can
necessity
of
drought
no
parameter measured under well-irrigated conditions
grain yield under postflowering drought, confirming
nurseries to select for
drought
the
tolerance.
Separating yield parameters into groups of predominantly drought escape (related
to
flowering),
with
a
or
potential
drought tolerance (related
as
tolerance
for
screening tools
panicle and panicle yield
and therefore
parameters. However,
were
suggested
are
likely
grain
relationships
to
be
used
were
panicle yield
found
al.
where the
(1991)
stress conditions
a
large proportion
based
of
Overall
results and
relationships
analysis for the identification
as
with
all
of
a
of
yield
as a
under
panicle yield higher than
or
have
a
positive
suggests the method of
using inter-relationships
the
stress in different
performed
feasible with the
drought
as
a
genotypes
Modelling
suggested method, since time
prone environments and has to be
to
grain
genotype's cycle length
well
as to
across
of individual
of
parameter consistent for
facilities, the genotypes
physiological stages
variable.
as a
flowering, determining
irrigation
common
of the determinants of DRI identified
well-irrigated conditions (potential yield)
design required by
therefore
in Fussell ef
related parameters were constant
selection parameter
stability analysis
genotypes. However time to
drought
(1987b) for
breeding aspect is
drought
viable tool and
is very unstable for most genotypes. Due to this,
in
a
feasible. However the individual DRI varied for the six
between years. A
yield
the most
(Figure 4.2).
between years. This confirms DRI
DRI
a
tolerant genotypes could be
drought
the trial's mean show either an indifferent reaction to
response
the
is
panicle yield under drought
on
(positive drought response). Genotypes
identified
published
phenotype-yield relationships and
By selecting entries
panicle
seems
etal.
by Bidinger
were
drought
to
possible screening
as
number of each individual
Indian millet. Part of the results of 1988 and 1989
further discussed.
identify parameters
parameters related
as
labor intensive task and is not feasible. Therefore
parameter. Similar
to
drought tolerance. Both grain number per
identified
to determine
DRI) helped
to
the
were
experimental
submitted to
the three years and
genotypes is therefore
flowering
predictable.
is of
not
high importance
31
The
of
use
panicle yield
as a
tolerance is hazardous since
a
The genotypes ICMVIS 85332,
(b)
of
<
(1963)
genotypes
perform
can
x
year interaction
P3Kolo with
Synth-1, and
a
expected
to have a
postflowering drought
well under
very sensitive to environmental
drought-prone
necessary
The
areas.
complement
high genotype
x
screening-technique
stress. The other three
The results indicate that
to the
screening with
year interaction of
a
reasonable
>
an
photoperiodic sensitivity
a
probably
a
stability analysis is
of individual
large
genotypes
at the same time
on a more
a
makes this
approach
Since
a
genotype
to
x
a
it is
be
improve the
on
cycle length.
year interaction of
genotypic variability,
as
can
emphasize
to stabilize their
the genotype
viable basis
of
amount of data over years
stability analysis.
physiological parameter explaining
panicle yield, expressing
genotypes
1 are
improved irrigation facilities and also to
drought tolerant genotypes
and to
agronomic parameter.
panicle yield
requires
a
reliably characterized by its yield potential, further research has
experimental design using
and
changes. They definitely cannot be recommended
unreliable and
and/or locations to get
Further
found.
changes. These
mostly positive drought response
(ICMVIS 85321, HKP, and CIVT) with regression coefficients
for
was
regression coefficient
classified as insensitive to environmental
are
be
significant genotype
stability analysis for panicle yield suggested by Finlay
1 after the
Wilkinson
screening parameter for postflowering drought
could
suggested
so
identify
far.
32
5. Grain Growth under
Stress
Postflowering Drought
5.1. Introduction
grain yields from postflowering drought stress were primarily due to
Reduced
reduced
see
grain
4.4.).
also
abortion of
can
grain
mass
grain number per panicle
flowering
mass, in response to
due to
drought
be due either to a reduction in
of GFP. The processes of
parameters
and individual
that occurred after
grain
period (GFP),
panicle
The reduced
grains
in individual
length
numbers per
in the response of
grain growth
pearl
millet to
(Fussell
due to selective
was
drought
stress. Variation
during
stress
grain growth
the
rate
grain filling
(GGR)
and the role of these
drought
both
et al. 1991;
and/or
grain growth
have not been elucidated
in the literature.
There exists
genetic
variation in
Powell
1968),
1980).
Fussell and Pearson
period
in
pearl millet:
accumulation takes
linear
subsequent
both the
LGGP
length
an
place
single grain
millet for
(1978a) identified
initial short
and
of the
when little
grain
mass
1979)
suggesting
was
wheat
and
(Triticum
that these
aestivum
GGR with
a
relatively short
postflowering drought
The extent of
(Oryza
L) (Bruckner
were
sativa
and
(GGR)
in the
pearl
millet
found to be
L.) (Jones
a
ef al.
Frohberg 1987b),
in a selection program, if this
Frohberg (1987b) proposed that
duration was
a
high
desirable, risk reducing strategy for
stress.
genotypic differences
well-watered and
in rice
rate
between
differences
parameters could be used
warranted. For wheat Bruckner and
dry weight
found that differences in
genotypes. Phenotypic differences for these parameters
primarily genetic rather than environmental
Majmudar
cell differentiation occurs, and a
grain filling period and the grain growth
with
and
phases of the grain filling
lag phase (LP)
presumably
(Burton and
mass
grain"1 (Rachie
two
grain growth phase (LGGP). They
correlated
were
pearl
which varies from 3.5 to 16.0 mg
in individual
postflowering drought
grain growth of pearl millet
in
stress conditions is examined in this
33
chapter. Further
the
relationships
of
grain growth parameters
components under well-watered and drought
order to understand their
drought tolerance
is
importance
stress conditions are
yield determination.
in
yield
to
and
yield
studied, in
Their relation to
investigated.
5.2. Materials and Methods
5.2.1.
Experimental Design and Crop Management
Detailed
4. The
thrice and
irrigation
millet genotypes
forty-one genotypes
accepting
a
higher
were
were
twice within each main
analysis done
in
flowering
at 50%
of 50% of the
included in both field trials but
included in the
chapter 4.
plot. Inigation
postflowering drought stress where
a
flowering
date of a
susceptible
respectively,
genotype than
The genotypes excluded from the
to
plots.
forty-two and
in 1988 and 1989,
analysis
variation of differences in
dwarf types and those very
fully given in chapter
plots (irrigation treatment) replicated
well-irrigated control and
completely discontinued
was
Forty-five
of main
subplots (genotypes) replicated
treatments were a
in the
in both 1988 and 1989 are
descriptions of the trials
split-plot design consisted
analysis were
mildew. Sets of genotypes
downy
varied between years.
5.2.2. Observations
Flowering
stigmas
for individual
had
emerged
plots
on
was
determined
Bidinger 1981). Individual inflorescences of both
tagged on
of
the date when 50% of the
as
50% of all the inflorescences in the
main stems and tillers
their date of flowering. The inflorescences
grain growth based
flowering: EF)
time of the
on mean
time to
flowering
consisted of inflorescences
plot
and the second group
plot (Maiti and
of
were
a
flowering
were
grouped for monitoring
plot. The
before the
(late flowering: LF)
first group
mean
(early
flowering
of inflorescences
34
flowering after
differences of
this time.
This
separation
characters for
grain growth
genotype under both well-watered
without
Samplings
taken per
were
staggered
day resulting
in
a
one
inflorescence
in the control and from
daily sampling
in the stress
linear
was
replications, such that
the
replication
of at least 100
5 to
day
day 16
(day
6 to
day
randomly
at 25
and
days
Inflorescences
were
The
were
in each treatment.
sampled
once
grain
mass
taken
day
midway
6 to
day
14
assumed to cover the
Further,
total of 18
a
Maiti and
were
mass
were
Bidinger 1981)
after the end of
only. Grain samples
grain filling.
oven-dried for 24h
determined.
analysis
grain growth and development
development
is determined
1983) and there
were
was
based
comparisons
on
as
thermal time. The
duration of
primarily by temperature (Fussell
use
of
grain
pearl
millet
et al.
1980; Ong
variations in temperature between and within the seasons.
temperatures below the
calculated
were
panicles were tagged
17 in the control and
flowering (i.e., maturity,
thermal time was necessary for line
For
grains
replication group (6 panicles per replication)
after
was
in the stress treatment in 1988. In 1989,
day
70°C, hand-threshed, counted and their 100-grain
5.2.3. Data
sample
groups 1 to 3 and 4 to
19 after the
treatment). These sampling periods
grain growth phase (LGGP)
one
replications. For both early and
samples
in each treatment to determine final
at
drought
from
5 to
reduced
samples per genotype
taken
daily
day
a
stress environments
stem and tiller inflorescences.
set of data with two
late flowering inflorescences,
along
and late inflorescences within
and
and treatment from each
genotype
6 for each
across
early
(control)
differentiating between main
undertaken to observe any
was
optimum for grain growth, thermal time (0,)
as:
e1=
s
1=1
civr,,)
was
35
where
is
T,
mean
below which
(02)
temperature for the ith day and Tb is the base temperature
calculated
was
is
grain growth
zero.
For
as:
"
(r.-r,) (t0-t„)
hi
where
T0
1983),
day using
T0
a
Tm
(T--r<>>
temperature above which grain growth is ceased and
optimum temperature for grain growth. Thermal time
ed as °C
and
is the maximum
Tm
is the
the two
of 33°C
equations.
(Fussell
have not been
et al.
It
1980; Ong 1983)
determined for
accurately
of the crop
development processes
developmental
processes in at least
1985) and
assumed to be the
were
was
recorded at the
Linear
regressions
were
fitted to the
nearly
of 10°C
the
pearl
(Ong
of 45°C. Both
Tm
findings
in
T0
pearl
for other
1982; Ong and
ef al.
constant for a range of
(Ong and Monteith
millet
grain filling.
for
grain sample
to determine the
lag phase (LP) (after
of the
is
Tb
line of
Tb
a
Air
temperatures used
station at ISC, Sadore.
meteorological
grain growth phase (LGGP)
length
same
a
be accumulat¬
grain development
(Garcia-Huidobro
one
can
using
and
supported by
are
and the observation that
1985)
calculated
was
millet. The cardinal temperatures used
Monteith
thermal time
supra-optimal temperatures,
data taken
grain growth
during
the linear
(GGR) and
rate
the method of Johnson and Tanner
the
1972)
(Figure 5.1).
The GGR
was
by setting
the
°C
day
from
to the final
as
the
expressed
regression
micrograms per grain per °C day.
flowering. Duration
grain
period
mass
(LF) panicles
of GFP
a
grain
was
determined from the
mass
were
analyzed. Finally,
the
samples
from
LP
of zero, and
calculated
was
derived
expressed
as
by setting the regression
maturity sample.
from the end of LP to end of GFP
irrigation treatments,
grain growth
as
of the LGGP to
LGGP
(i.e., GFP
early flowering (EF)
was
derived
LP).
For both
and late
flowering
-
initially kept separate and their grain growth characteristics
the two
samples
were
combined to calculate
attributes in both environments.
mean
genotype
36
Relationships
determined
tolerance
flowering
between
using
grain growth characteristics
correlation
analysis.
(the drought response index [DRI]
was
and
yield
attributes
In addition, their correlation to
as
in
4.2) and time
analyzed
in both years.
explained
calculated for the sub-sets of 34 lines
were
drought
to
Grain weight (mg/grain)
(irairi
a
(GFP)
Filling Period
.
Linear Grain Growth Phase.9
(LGGP)
y
Maturity
Lag Phase*
(LP)
slope
*
Grain Growth Rate
Flowering
/
Thermal time from
Figure
5.1. Model of
flowering (*C d)
grain growth analysis.
5.3. Results
5.3.1. Grain Growth Under Well-Watered Conditions and
There
of
was
considerable
genotypic variability
grain filling (LP, LGGP, GFP),
mass
the
under well-watered conditions
means of these
flowering
parameters
over
grain growth
(control),
both
was
components of the duration
rate
(GGR)
and the final
grain
in both years. Table 5.1 shows the
sampling periods
inflorescences. The extent of
parameters observed under control
for the
Drought Stress
genotypic
of the
early
variation for
maintained under
and late
grain growth
postflowering drought
37
stress
(Table
phase (LGGP)
while the
1) The length of gram filling period (GFP), linear grain growth
5
and the GGR varied
of the
length
of 265-495 °C
There was
30-32°C)
no
represents 13
significant
LGGP, GGR and final gram
grain
mass
mean
mass
of all genotypes, there
LP and GFP
were
two-fold between lines in both years,
varied up to three-fold. The overall range
lag phase (LP)
for GFP
day
nearly
to 22
days (at
mean
temperature of
difference between the genotype
between the two years. For the
was
significant
vanation in 1988
slightly longer than observed
means
mean
only.
final
In 1989,
in 1988
Table 5.1. Means, ranges and results of ANOVA for grain growth parameters under wellwatered and drought stress treatments Sadore 1988 and 1989
1988
well-watered
Lag phase (°C day)
Linear grain
growth phase
drought stressed
Range
F„2
CV%
71
45-114
67
34-112
23
40 5
273
186-369
226
146-310
29
6"
31 8
343
265-422
292
214-359
65
3"
18 2
28"
18-39
Mean'
Gram growth trait
Range
Mean
(°C day)
Grain
filling period (°C day)
Grain growth rate (ug
grain'
26"
17-43
32
24 9
3 4-9 4
231"
146
F,„
cv%
56
50 0
("Cday)1)
Final grain
mass
(mg gram1)
7
1"
4 3-10 4
5 6"
1989
well-watered
Grain growth trait
Mean
Lag phase (°C day)
Linear grain
growth phase
Range
drought stressed
Mean
Range
94
30-127
83
10-123
298
231-459
262
155-387
15
0"
314
392"
332-495
345
265-402
612"
154
26"
17-33
22"
16-31
34 2"
23 9
51-10 9
5 3"
3 8-6 3
242"
13 3
(°C day)
Grain
filling period (°C day)
Grain
growth
rate
(ug
grain'
("Cday)1)
Final grain
'
P<0 05
mass
(mg grain')
7
"
P<0 001 for genotypes
4"
'"
and for irrigation
for
38
Total GFP and LGGP
when
GFP
the
compared
shortened
was
were
reduced
significantly
47-51 °C
by between
stress. Mean GGR
imposed drought
drought
stress in 1988
drought
stress
(compared
to the
day (i.e.,
(data
the
drought
not
presented).
In 1988
though
GGR
postflowering drought
stress.
1988 and 28% in 1989
Average
are
was
1989
influence
on
shown in
(LGGP)
with P<0.01
for the
final
of
of EF had,
only
early flowering
unchanged
to EF in 1989. The
not
significantly
in
1989).
was
grain
by
15%. Nonethe¬
by
associated with
grains. These trends,
enhanced under
were
got reduced by 21% in
mass
early flowering (EF) and late flowering
5.2. The
on
grain growth
of
flowering of
within
a
single
genotype had
only
a
in
average, longer lag phases (LP)
as
the late
grain growth period
flowering
grain filling period (GFP)
significantly
(GGR), averaged
samples
a
under control conditions
The average linear
the same
rate
timing
flowering
for the LF in 1988, but was
different between
second half of the
due to
such genotype in 1988
one
under control,
Figure
in both years. As a result, the mean
remained
only
grain growth parameters
(Table 5.2). Grains
(significant
GGR
mean
positively (P<0.05)
inflorescence in relation to the distribution of
significant
days)
not affected
was
(Table 5.1).
flowering distributions of samples
(LF) inflorescences
than two
significantly (P<0.05) lower GGR
a
were
always significant
not
the
which also had heavier
early flowering genotypes,
evident
The
high
more
(all genotypes)
significantly (P<0.01) reduced
stress in 1989, and there
In both years, average
control, Table 5.1). However, in 1989
less, only 3 of the 41 genotypes showed
under
by postflowering drought stress,
(Table 5.1).
with well-watered conditions
across
inflorescences
for all genotypes
shorter
all
compared
genotypes,
was
of inflorescences from the first
flowering distribution. Under well-watered conditions only
or
4 out
of 42 genotypes in 1988 and 3 out of 41 genotypes in 1989 had a
significantly
lower GGR for late
of individual
grains
was
significantly reduced
Comparison
samples
flowering
inflorescences. However the final
of
under
in the LF in both years.
grain growth parameters
postflowering drought
from
drought
grain
early (EF)
and late
stress to control showed a
of all parameters in the LF than in the EF
stress final
mass
masses were
(Table 5.2),
as
(LF) flowering
higher
reduction
expected. Under
reduced in the later inflorescences
mainly
because of shorter LGGP.
Number of Panicles
1200
1000
800
600
400
200
0
1000
800
600
400
SO
45
40
55
60
65
70
75
Days After Sowing (DAS)
i
I Early Flowering (EF)
Figure
5.2.
Flowering distribution
of
Late
Flowering (LF)
early and late flowering samples
as
averages of two years (1988 and 1989) in control (a) and drought stress (b).
Sadore 1988-1989.
However, the
with
only
having
a
mean
GGR (all lines)
5 out of 42
genotypes
was not
in
significantly different from EF
1988, and 8
out of 41
and LF
genotypes in 1989
lower GGR for inflorescences from the second half of the
flowering
distribution.
The
relationships of grain growth parameters
attributes under well-watered conditions
analysis
across
genotype
means
were
to each other and to certain
examined
using
yield
simple correlation
(Table 5.3). The length of linear grain growth
'
EF
=
mass
first half of the
Final grain
grain1)
(ug
(mg
rate
growth
'(°C day)')
Grain
grain
68
706
27
345
277
flowering distribution,
716
28
341
filling period (°C
Grain
day)
267
74
Mean
Mean
growth
phase (°C day)
Linear grain
Lag phase (°C day)
LF
EF
LF
well-watered
=
drought
605
28
303
230
73
Mean
EF
stressed
606"
5 65
9 2"
0 69
10 9'
F,
110
782
27
411
302
547"
8.01
313"
0 63
52 7"
in
samples
606
24
373
273
101
Mean
EF
*
P<0 05
458
21
316
252
64
Mean
LF
"
stressed
8"
9"
P<0 01
113
7 64
39
1 74
19
well-watered and
drought
distribution
F, = Variance between
696
25
372
294
78
Mean
LF
F.
1989
flowenng
well-watered
Mean
EF
flowering distribution,
509
25
281
221
60
Mean
LF
second half of the
91'
177
014
0 60
179
1988
Table 5.2. Means of grain growth parameters from inflorescences of the first and second half of the
drought stress treatments Sador6 1988 and 1989
o
41
phase (LGGP)
(GGR)
grain
mass
had
as well as the
grain filling period (GFP)
and
positively
were
correlated with final
LGGP and faster
longer
grain growth
and low GGR
grain growth
rate. Other
panicle yield
and
related
years to the
grain growth parameters.
Several
grain yield
stress conditions
correlation
(LGGP)
flowering
grain filling period (GFP)
genotypic response
of a
genotype through
lag phase (LP) (Table 5.3).
stress. GFP was
panicle yield
grain
final
inconsistently
under
there
was no
(data
not
was
significant
found
or
in 1989, no
flowering
yield
grain growth parameters of the
since
under stress
however indicates the
were
attributes.
six
Except
linear
genotypes discussed
grain growth phase and total grain
genotypic differences
in
grain growth parameters could
Under
drought
stress
mass, while no differences were detected in 1989.
different responses to
resulted all in
mass.
reduced final
grain
genotypic
grain filling period, grain growth
grain growth parameters showed
a
(see 4.2.6)
yield
and
not very
parameters seemed to be higher in 1989 than in 1988
under controlled conditions.
grain
trait
drought
area
was
with
was
mass
under
correlation
positive
related to these
differences occurred in 1988 for total linear
and final
grain
grain yield per
to
predominantly adaptive
to time to
flowering
parameter. Other grain growth parameters
inconsistently
4 are listed. All
identified
although
this
final
grain yield
and
The
mass.
related to the time to
was
might expect
one
growth phase
grain
association of time to
panicle yield
relationship
under controlled conditions.
filling period
a
final
correlation of GFP under control to
Its
In Table 5.4 the average
chapter
stress
character of the
unrelated
As
correlated to
drought
presented).
drought escape
negative
stress
mass
intensity. Simple
of linear grain
determining
in
consistently significantly correlated
high. Therefore GFP
be
or
were
1989). Long
postflowering drought
nature nor
importance
drought
to
the
change
not
confirmed the overall
significantly (P<0.01)
in
unrelated
were
maintained under
were
(Table 5.3) and did
analysis
and
variation in
mostly
in
rate
higher
interrelationships between grain growth parameters and final grain
found under control conditions
the
only
a
LGGP
Long
rate.
latter
yield attributes than
mass, such as
across
itself
(the
grain growth
genotypes with
mass;
phase (LP)
associated with short lag
LP therefore enhanced
grain
Genotypes
rate
Genotypes'
drought stress,
showed either
but
longer lag
yield (g)
flowering
filling period (°C
'
P<0 05
"
P<0 01
Lag phase (°C day)
growth
phase (°C day)
Linear gram
day)
Grain
(ng gram' (°C day)1)
Gram growth rate
grain')
(mg
(g m2)
Final grain mass
Panicle
(day)
Time to
Grain yield
'
0 24
-012
-014
019
76"
-0 53"
-0
-0 55"
-0 28
-011
0 67"
0 77'
0 09
040"
0 08
0 23
58"
-0
-010
-0 30
0 57"
0 70"
40"
-041"
-0
0 23
0 33
Linear grain
(C)
and
48"
68"
0 90
0 95
-0
-0 34
042"
0 45"
-0
0 29
0 04
-0 04
-0 33"
018
C
day)
0 84
0 89
-0 71"
-043"
0 33'
035"
0 23
0 25
-003
-011
017
020
S
growth phase (°C
C»well-watered control, S=drought stressed treatment
-012
012
-013
-010
012
0 08
0 09
-0 04
0 02
014
-0 20
011
0 06
-0 03
-0 05
0 23
012
78"
018
0
0 26
-011
-0 67"
-0 29
0 42"
0 07
62"
0 07
0
0 09
71"
012
0
0 69
-012
0 08
-0 07
0 03
-010
S
C
C
S
Lag phase (°C
day)
DRI
Table 5.3. Correlations between gram growth and yield attributes (in the control
and 1989 (second row) SadorS 1988 and 1989
stress
31'
50"
-0 51"
-014
0 59"
0 61"
-0
0 30
-0
-015
-017
0 24
C
35'
-0 49"
-0 09
0 57'
0 64"
0 42"
042"
-0 39'
-0
0 44"
0 42"
S
031'
0 63"
017
0 07
-0 01
-045"
0 07
0 22
C
ug
growth
rate
60"
64"
0 21
0
-0 03
0 53"
-0 08
-0
0 04
0 58"
S
(first row)
46"
38'
-0 36'
0
-015
-0
-0 21
0 44"
C
mass
50"
78"
052"
0
-038"
-0 72"
0
080"
S
(g grain')
Final grain
and DRI for 1988
gram1 (°C
day)')
(
Gram
(S) treatment)
filling period
(°C day)
Gram
drought
day)
a
25
79 a
413.9 be
76 a
P3Kok>
310.2 abc
89 a
58.4 a
103.6 a
CIVT
a
a
19
20
a
77
77 a
66 a
a
a
a
a
494.6
447.5 ab
66
404.3 a
389.1 ab
124 a
151
a
90.3
Synth-1
a
59
a
89.8
HKP
96
a
24
306.6 abc
a
27
a
88
a
93
396.4 be
a
364.4 be
a
66
a
92.9
ICMVIS 85332
103
a
271.6 c
a
85
400.0 be
85 a
a
105.5
ICMVIS 85321
294.5 be
a
27
54 a
a
78.6
P3Kolo
85 a
6.49
76 b
29a
102 ab
a
120
217.6 a
99 a
a
74.2
CIVT
296.2
74 b
6.43 a
100 b
a
24
1989
90 ab
7.38 a
104 b
a
26
77 ab
351.6
a
71
277.5 a
Synth-1
a
26 a
a
89 a
357.0 a
83a
98
a
320.7
99a
293.3
73.7 a
a
252.2 a
96 a
102 a
68.5 a
HKP
69
7.24 a
71 a
116a
77
7.40 a
104a
a
69 a
70 a
7.82 a
7.74 a
a
a
115a
75
7.11
77 a
a
8.00
75 a
83 b
85 a
a
a
66 b
83 b
6.46 a
85 ab
79 ab
7.68
77 ab
26 a
a
a
7.18 a
116a
25 a
75 b
82 ab
344.7 a
360.2
67
83a
a
S
299.0 a
C
290.2 a
S
79 a
C
116 a
S
a
C
grain mass
(mg grain1)
Final
a
S
growth
(ug
61.2
C
rate
grain'1
(°C day)"1
Grain
54.5
S
grain filling
period (°C day)
Total
stress
ICMVIS 85321
C
Linear grain
growth phase (°C
grain growth parameters for six genotypes under control (C) (real) and under drought
control). Letters grouping genotypes after Duncan's multiple range test for p<0.05.
Lag phase
(°C day)
in % to
ICMVIS 85332
1988
(S) (relative
Table 5.4. Means of
44
phases and reduced
linear
grain growth phases
not consistent between years
vice versa. The response
or
allowing any identification
of
was
grain growth
a
strategy.
5.3.2. Grain Growth and
To
explore the role
of
Drought Response
analysis of grain growth parameters
subset of 34 lines
(which
carried out for both years
final
grain
were
yield
mass
(Table 5.3).
response
drought escape,
a
(i.e.,
(Table 5.3)
consequence of the
postflowering drought stress and GGR,
for GFP
There
the
as
were
the
The
growth
with
in GFP and
(GGR)
long GFP
and
grain
between certain
mass
grain
mass
under control
reductions in GFP and
caused
GFP
grain
was
grain
grain
mass
associated
mass, as
expression
of
yield
under
same was
found
conditions had
mass
under
(r=0.79
expected.
grain growth parameters and
by postflowering drought
in the
drought
drought
stress.
Genotypes
significantly (P<0.001) larger
drought
were
stress.
grain filling period (GFP), grain
under control and
GGR under control conditions, however,
reductions in
an
of both
early flowering. The
Table 5.5 shows the correlations of differences in
rate
relationship between grain
thus appears to be
only consistent adaptive parameter.
significant relationships
changes
grain growth parameters and
positive relationships
with
to DRI was
postflowering drought conditions
DRI).
to
treatments of a
irrigation
days after sowing)
None of the
under both well-watered and
drought
related to
measured in both
flowered from 59 to 71
in stress and GGR in 1988
drought tolerance, correlation
in
grain growth parameters
stress.
not more
Genotypes
likely
with
to have
high
larger
stress. The extent of the reduction of
in 1988 and r=0.70 in
1989) with
the reduction in
45
Table 5.5. Correlations between certain parameters of grain growth and the relative
reduction (%) of grain filling period, grain filling rate and grain mass due to postflowering
drought stress Sadore 1988 and 1989
Relative reduction
Year
Grain
in
Grain
filling
period
growth
Grain
mass
rate
Well-watered conditions
Grain filling period
(°C day)
1988
(ug grain'
1988
-0 23
1989
-0
1989
Gram
growth
(°C day)')
Grain
mass
Drought
Grain
(mg grain1)
filling period (°C day)
growth
(°C day)')
*
-017
0 68"
-0 59"
0 44"
-0.13
0 04
37'
0
56"
-013
1988
0 27
-017
0 25
1989
0
46'
-014
0 47"
stress conditions
Grain
Gram
rate
0.44"
0 62
mass
P<0 05
rate
(ug
grain1
(mg grain')
"
1988
-0 52
0
32'
-0.43"
1989
-0 72"
0
43"
-0
1988
010
-0
65"
-0 14
1989
0
-0
74"
0 06
53"
1988
-0 30'
-018
1989
-0 42"
0 11
56"
-0 47"
-0
72"
P<0 01
5.4. Discussion
Pearl millet is known to vary
Rachie and
material
greatly
in
grain
mass
Majmudar 1980). This study corroborates
mostly
of Sahelian
is consistent with
previous
origin.
(Burton and Powell 1968;
these
The overall range of
studies
(Bishnoi
findings, with genetic
grain filling period (GFP)
et al. 1985; Fussell and Pearson
1978a; Fussell etal. 1980; Ong 1983) but much shorter than reported for other
cereals, i.e., wheat, rye and triticale (Stamp
growth
curve as
it was used in the studies of
be reduced to a linear model.
Figure
et al.
Stamp
1982).
etal.
The
(1982)
5.3 shows the average
sigmoid grain
could therefore
genotype grain
46
growth
growth
in
for each year and treatment
curves
curve
grain
mass were
phase (LGGP),
1980).
mostly
and to
a
due to differences in
These differences
stress treatment. The
length
lesser extent to differences in
the results with
confirming
the
confirming
viability
of
a
linear
high R2 for the regression. Differences between genotypes
with the
a
of linear
grain growth
smaller set of genotypes
magnitude
so
rate
(GGR),
and Pearson
(Fussell
evident in the control but less
were
grain growth
in the
of the variations in the various
drought
grain growth
parameters (more than two-fold in both years and both moisture environments)
appears sufficient for effective utilization of
millet
breeding program
Qrain
mass
grain growth parameters in
a
pearl
for all environments.
(mg/grain.)
8
R-Squares for linear fitting
6
control
stress
1988
0.994
0.970
1989
0.996
0.991
-
z8* -d'
-*
'*
200
150
100
50
*
Figure
D
Control
5.3. Grain
growth
curves
Stress
450
(°C)
1989
1988
under control and
400
350
300
250
Thermal time from flowering
drought
stress in 1988 and
1989. Sadore 1988-1989.
The
principal effect of postflowering drought stress on individual grain growth was
to reduce
grain
than GGR for
have been
mass.
a
This occurred
large majority
reported
for maize
primarily through
of the lines tested
(Zea
mays
a
reduction in GFP rather
(Figure 5.3).
L.) (Ouattar
et al.
Similar
findings
1987a). Truncation
47
of GFP could have resulted from a limited assimilate
In maize, restricted current
grain storage capacity.
caused
under
by
implicated
water deficit was
as
the
supply
(and)
or
a
reduced
photosynthate production
for reduced
reason
grain
mass
postflowering drought stress (Ouattar et al. 1987b). Water deficit during the
lag phase (LP)
division
could have reduced the
grain storage capacity,
during that phase of grain development,
a
LP
length of
Because the
1987a).
imposed,
was
restricted current assimilate
for the reduced
grain
mass
and
found in maize
as
little affected
supply appears
of
early cessation
by cell
determined
by the drought
a more
etal.
(Ouattar
stress
likely explanation
grain filling than
a
decreased
storage capacity.
The
sources
of assimilate
photosynthate
drought
supply
are
to ensure seed
available to maintain
cause
stomatal closure in
before
pearl
millet
Squire 1979; Henson etal. 1981) and, presumably,
production (Boyer 1976). However,
by
grain filling: first,
a
and Pande ef al.
grain filling
water deficit
flowering. Severe
water
(Squire 1979; Black and
reduction in
photosynthate
lack of current assimilation can be buffered
in
(1983) observed that soluble sugars stored in
millet
peari
were
mobilized
during grain filling.
conditions, the assimilates stored before anthesis
contribute 26 to 30% of final
grain weight (Ouattar
et al.
under either well-watered
pearl
Although
millet.
performance
were
Consistent
in
relationships
were
yet
were
grain
et al.
mass
to be done for
consistent, the
grain growth characteristics
varied.
due to the date of onset of stress that had affected
slightly
different
development stages
imposed
for each genotype
stress could not be
growth parameters
conditions has
prior
estimated to
1987b; Jurgens
the overall results of the two years
mainly
genotypes
drought
drought stressed
of individual genotypes in their
This is attributed
individual
or
stems
In maize, under
1978). Quantification of the contribution of stored stem sugars to final
of
current
secondly,
mobilization of sugars stored in the stems in many cereal crops. Fussell etal.
(1980)
to
a
Two
production.
flowering and,
after
carbohydrates produced
translocation of stored
implies that pearl millet
stress
photosynthetic activity
from
stress is known to
under
adaptation mechanisms
robust
possesses
GGR
of
maintenance
between
found
grain yield
only
under
for GFP and
since the
individually.
drought
grain
beginning
mass.
stress and
There
are
grain
similar
48
reports for wheat (Bruckner and Frohberg 1987b; Sayed and Gadallah 1983).
Furthermore, there
were
parameters and DRI, suggesting that
measured
selecting
in
pearl
were
for
expressions
grain growth
consistent associations between
no
of
none
drought
grain growth parameters to
of the
tolerance
enhance
grain growth parameters
susceptibility. Therefore,
or
postflowering drought tolerance
millet appears not to be feasible.
However, exploiting certain drought escape features observed for the grain
growth parameters
does appear
postflowering drought
a
possible approach
Lines
stress.
with
to
improving yield
shorter GFP
conditions had smaller reductions in both GFP and
under
under well-watered
mass
under stress.
Moreover, lines with higher GGR in well-watered conditions
had smaller
reductions in GFP. Therefore,
postflowering
relatively
under
GFP
a
viable
stress would be to select lines with
short GFP. Bruckner and
postflowering drought
by direct
risk-reducing
strategy for
large
Selecting
number of
too many observations and not be cost effective.
areas
prone to
high GGR combined with
Frohberg (1987b) proposed
stress in wheat.
measurement in a
grain
for
high
a
similar pattern
GGR and shorter
pearl millet lines may require
49
6.
Carbohydrates
Reserves
During Postflowering Drought Stress
6.1. Introduction
Previous results indicated that the reduced
drought
for
carbohydrates
et al.
(Ouattar
plant parts
grain filling
cereals:
major
maize
grain yield
to
have been
postflowering
matter. He
the
as
supply.
source
of
for the
1987a) and the tillering C3-species rice and wheat (Borrel
et al.
explained
proximity
well
non-tillering C4-species sorghum (Goldsworthy 1970) and
assimilatory surface,
5% of the total
as
thoroughly investigated, mainly
that the
1989). For pearl millet Egharevba (1981) found
part, its
under
mass
stress conditions could have been a result of limited assimilate
The contribution of different
four
grain
but contributed to 32% of the total
of the
high efficiency
the
flag leaf had only
flag leaf
as a
vantage position
to the sink and the
grain dry
relative young
to
plant
intercept incoming
radiation.
stress on mobilization and transloca¬
Some studies report the effects of
drought
tion of assimilates, e.g., in maize
(Ouattar
1983; McCaig and Clarke 1982) stem
importance
for
showed that
a
grain yield
low leaf water
completely
inhibited
entirely
the
on
under
stress.
of maize
so
that
of assimilates.
Dry
photosynthesis
reserves
1987a).
of
reserves
drought
potential
et al.
as
In wheat
carbohydrates
Westgate
temperature (Fussell
matter
consisting mainly
contribution of almost all
plant parts
(Rao
ef al.
1978). Long and thick
(Pande
ef al.
high
in
carbohy¬
mobilized after
drought
stress
might
dependent
1980). With increasing drought
on
stress the
decreases. However the contribution of the
assimilates in millet stems increases
of assimilates
a
were
In millet translocation of assimilates is
ef al. 1978b and
of
reproductive development depended
anthesis. The amount of stored assimilates before
important.
are
Boyer (1985a)
it occurs under drought stress,
drates, which had accumulated in maize stem and leaves
therefore be
and
et al.
(Blum
as
the contribution from leaves decreases
stems are therefore
1983). Bidinger
contribution of stored assimilates in wheat of
ef al.
advantageous for storage
(1977) suggested
that the
preanthesis photosynthesis
can
50
determine between 12% and 98% of the
Campbell (1972)
Hume and
described
grain yield
an
under
optimum conditions.
accumulation of assimilates in
com
stalks until 2-3 weeks after anthesis which then declined due to mobilization and
translocation into the
In
grains.
plants carbohydrates
sugars and their
(a) simple
(b) storage
be classified in three
can
conjugates
sucrose
polysaccharides, e.g., pectines, hemicelluloses
irretrievable to the
generally
nonstructural
plant parts
from
plants in
were
The effect of
and
as
maturity
were
celluloses,
which
are
study
total
reserves
in
in the amount of stored TNC
investigated for the
postflowering drought
1987):
fructans; (c) structural
storage and energy
analyzed quantitatively. Changes
up to
and
times of stress. In the present
carbohydrates (TNC) serving
flowering
Soest
intermediary plant metabolism;
active in
compounds, e.g., starch,
reserve
categories (Van
shoots of three genotypes.
stress on the TNC content was studied.
6.2. Materials and Methods
6.2.1.
Experimental Design
An off-season trial
plot
with
established in 1991. The
main-plots irrigation
replicated
a
was
treatments
five times within each main
well-watered control and
irrigation
was
withheld at 50%
Irrigation was
done
in four
of 5m
rows
same as in
0.75
m
a
by
a
linear
experimental design
replicated
plot.
The
twice and
irrigation
postflowering drought
flowering (Maiti
moving
overhead
et al.
the trial of 1990
(see 4.2.4).
m
enable the observation of individual
taken from the two middle
rows.
plants.
sub-plots genotypes
stress treatment, where
1981) of 50% of the plots.
irrigation system. Plots consisted
The trial
apart. Thinning
was
treatments were the
planted by hand
was done to one
All
split-
treatment consisted in
length. Fertilizer application and crop
apart, in hills 0.25
was a
plants
plant
on
ridges
per hill to
for the observations
were
51
6.2.2.
Sampling
nonstructural
to
of Plant Parts
Processing
genotypes (ICMVIS 85332, CIVT and P3Kolo)
Three millet
expected
and
carbohydrates.
represent
a
These genotypes have similar
range of
difficulties in
consisted in
differentiating
one
plant,
were
main stems and tillers in
been removed
early after their appearance
One genotype
(CIVT)
for
yield analysis
main stem
panicles
flowering (FL),
stage)
of a
at
panicle
The
date.
all
plants
main stem
Samples
immediately
of the
the
was
9
of
days (FL+9)
at
plants
and at
was
cut
the tiller
done at the stage of
maturity (M) (black-layer
was
sampled
were
stem at the base of the
dried
into three pieces equal of
as one
taken between 8.00 h and 9.00 h. All
samples before processing
and
samples
Dry weight
in a hammer mill and
sample
each.
length including their
sub-samples per plant
after harvest at 60°C for 48 hours.
at each of the
plant height of the
the total
panicle) and the diameter of the
panicle and
as a
samples.
control. Plants
genotypes. Flowering date for
ground level and
leaves. This resulted in five
were
for all
Sampling
all other tillers had
as
in order to have uniform
designed plants per plot
were cut
measured. The
corresponding
+
were
later stage. The hills
a
left with all its tillers
normally
recorded.
flowering
(including
was
additionally
left to tiller
was
plant. One
three dates. The
main stem
were
phenology and
tagged before tillering, excluding
with one main stem and one tiller
was
analyzed for
drought responses (see 4.3). Of the designed
for this study the main stems
plants
were
was
subsequently
sampling
were
dried
determined for
in a
high speed
mill to pass a 1 mm sieve.
6.2.3. Chemical
Analysis
of Nonstructural
Total water-soluble nonstructural
tively after
the method
consisted in
a
contents
hexose
boiling
as
suggested
colorimetric
250 mg of
carbohydrates (TNC)
for
phenol
-
were
a water
analyzed quantita¬
sorghum by Guiragossian
sulfuric acid
test tubes,
bath. After
cooling,
et al.
(1979).
It
procedure determining TNC
equivalents. Sugars (including starch)
plant sample in big
during 40 minutes in
Carbohydrates
adding
were
extracted
by
15 ml of distilled water
10 ml of each acetate buffer
(pH
52
4.45), amyloglucosidase (0.5%),
and
incubated for 24 hours at 60°C. It
was
into 100 ml volumetric flasks and
included for each
1 ml of
on a
run.
phenol (5%)
One ml of
was
amylase (0.25%) solutions
then filtered
were
added and
through Whatman
#42 paper
to volume.
brought
aquatic solution
was
Enzyme blanks
pipetted
added. Five ml of concentrated sulfuric acid
test tube shaker. The
highly exothermic reaction
was
were
into test tubes and
added
were
accompanied by
a
change
to a reddish color. Seven ml of distilled water was added for dilution and
stirred
again. Readings
Spectronic 501)
color
The
fully developed.
were
calculated
were
taken with
a
spectrophotometer (Milton Roy,
at 490 nm after 30 minutes when the
Glucose standards
TNC
were
as
analysis
of each
samples were cool and the
plant sample
repeated thrice.
was
used in concentrations of 0.2, 0.4, 0.6, 0.8 and 1
% TNC
as
well
as
g TNC per
plant
g/liter.
part and per
plant.
6.3. Results
6.3.1. Yield and Yield Related Parameters in the Yield Plot
Some
are
important parameters
presented
significantly
CIVT but
at 62
nor
the same
DAS, equivalent
yield
was
85332,51.4% for
a
reduction
stress was
neither between
to a water
supply adequate
supply
to all
for
received
a
last
approximately
four
plants during their period
(lag phase). Compared
and
in all three
drought
irrigation
to the two other
genotypes. This
was
due to
a
days.
of cell
genotypes the
minimal for P3Kolo. The reduction in biomass under
significant
dry weight
significantly
physiological period. Stress plots
sufficient water
was
not differ
between genotypes. Therefore all lines suffered
differentiation in the grain
stover
stress of 42.4% for ICMVIS
days after sowing (DAS) and did
of harvests
This allowed
yield
drought
yield plot
observed in the
as
in Table 6.1 in order to bias the chemical data. Grain
reduced under
during
stress
grain and biomass
only of 25.1 % for P3Kolo. Flowering date of plant samples was between
58 and 60
samples
for
drought
reduction of both
panicle dry weight. The decreased biomass under drought
stress was less in P3Kolo than the other two
genotypes.
53
Table 6.1. Yield parameters of the
yield plot. Sadore
1991.
ICMVIS 85332
CIVT
Parameter
C
S'
C
S
C
S
Flowering (day)
60
60
59
60
58
58
Biomass
(g m'2)
m2)
Stover (g
m'2)
Panicle (g
Grain yield (g
Panicle
1
'
C
m'2)
yield (g)
control
well-irrigated
=
P<0.01
678"
1191
717"
1095
782"
17
421"
754
459"
706
472"
17
422
257'
436
258"
390
310'
24
260
137'
248
121"
237
178'
34
27.0
17.3"
25.8
14.2"
26.8
20.6"
26
S
=
drought
stressed
for irrigation treatment
Total shoot biomass
significantly
as
Carbohydrates
derived from the total
An average of 35.5 % of the total biomass at
produced during the grain filling phase.
proportion
biomass instead of
Figure
between the accumulated
dry
matter accumulated
matter to the
panicle
genotypes with
of the
Nine
panicle
days
among
a
final
was
after
or
height
8.8 mm at
under
filled.
of 234
flowering the
plant parts
the
calculated
was
flowering
major
plant parts
to
as
maturity
the ratio
to the total
period. Partitioning
same
growth after flowering
produced,
were
matter from
plant during
All other
of
dry
after anthesis was 1.0 for ICMVIS 85332 and 0.74 for both
CIVT and P3Kolo. Tiller
panicles
the
35.9 %
after anthesis. From data
dry weight significantly
panicle dry
by
grain.
in
to
maturity was
Therefore the
flowering.
was
partitioning coefficient
the
6.1
dry weight
plant parts
flowering
panicle comprised
At harvest the
13.4 % at
of the increase of shoot
did not increase nor decrease
shown in
of all five
dry weights
differ between genotypes from the control from
maturity (Figure 6.1).
of the total
7
1104
6.3.2. Total Nonstructural
did not
CV%
682
"
P<0.05
P3KolO
Plant
cm at
flowering
total
drought
height
was
flowering.
for all
hardly any
nil and
was
not
tiller
different between
Stem diameter at the base
genotypes.
plant dry weight
as
well
as
stress did not differ from the
its distribution
well-irrigated
54
dry weight (g)
CIVT
Figure 6.1. Distribution of dry weight in shoots of single plants of three
genotypes (ICMVIS 85332 (IC 32), CIVT and P3Kolo) at flowering (a), flowenng
+ 9 days (b) and maturity (c) under control (C) and
drought stress (S). Standard
errors
of
means are
given
in
parenthesis. Sadore 1991.
55
conditions, indicating
total
dry weight
for the
filling
even
the
panicle
to
a
plant growth during the early grain filling (Figure
similar
period dry weight accumulation
6.1).
After this
declined
dry weight
plant
dry weight in
stem
increase from
dry
dry
was
flowering
dry
matter from
average rate of
genotypes. The
flowering and
weight
decreasing
dry weight
stem
panicle weighed
new
dry
matter
with translocation of
to the
plant from flowering
to
well-irrigated treatment for all
biomass at
above-ground
the ratio of
panicle weight
to stover
remained the same as in the well-watered treatment.
harvest dates
(Figure 6.2)
carbohydrates (%TNC) expressed
significant (P<0.001)
in the control. As
plant increased significantly from flowering
days
later and 16.2% at
contents in the
panicle from
maturity.
7.4% at
differences could be found under
a
above total
was
12.4% of the total
maturity. At maturity
relative contents showed
stem
(1.76
For this reason the
parts and vegetative plant parts. The
compared
The concentration of total nonstructural
nine
weight. P3Kolo
partitioning of dry
The
P3Kolo).
accumulation in the whole
to about half
31.1% at
mid-grain
of
period
for all genotypes above 1.0
maturity, compensated partially
to
matter
maturity got reduced
panicle
matter in the
period.
same
for ICMVIS 85332, 1.93 for CIVT and 1.20 for
accumulation of
of CIVT. In CIVT except
parts. A significant reduction in dry weight occurred
drought stress
under
panicle
case
to increase its tiller
also in the tillers of CIVT and P3Kolo for the
matter to the
stress was nil and
drought
parts declined in the
maturity. ICMVIS 85332 tended
maintained
in the
significantly
of all
under
parts
did not
higher level
change
from
differences in
expected average
to
This
maturity:
was
plant parts
TNC contents per
11.5% at
mainly
flowering, 13.9%
increasing TNC
due to
flowering to 28.2% at maturity.
well-irrigated conditions. TNC
flowering
to
No
one
higher
genotypic
contents in the
maturity. The lower stem
than the middle stem, and the latter
as
for all
always had
than the upper
stem.
Average TNC
content per
flowering (12.1 %) to
FL+9
plant under drought
slightly
from
(13.1 %) and decreased significantly at maturity (9.5%).
Relative TNC content declined for all
to
stress increased
plant parts except the panicle
from anthesis
maturity for ICMVIS 85332 and CIVT. P3Kolo maintained its TNC levels in the
stem
parts up
to FL+9 after that values declined for all stem
parts and tillers
56
(Figure 6.2).
The TNC content
genotypes from flowering
decreased
declined
(CIVT)
or
significantly
the
in
panicle increased significantly
to FL+9 and was then maintained
further increased
from FL+9 to
(P3Kolo).
physiological
plant parts
exclusively
were not
place
panicle while
panicle
0.63 for P3Kolo in the
significant (Figure 6.3). Only
was
of 0.97 for ICMVIS 85332,0.81 for CIVT
period from flowering
of total TNC
well-irrigated
Genotype
FL
ICMVIS 85332
12
CIVT
12
P3Kok>
13
Corresponding
to
maturity.
then decreased to
physiological stages (FL, FL+9, M)
partitioned to panicle Sadore 1991
control
drought
FL+9
M
2(7 9)
164(40 9)
26
5(7 5)
213(310)
6(6 7)
16
371
to relative
8(38 8)
In
stress
FL+9
FL
M
13 3(59
7(561)
13
6(71)
141(313)
355(551)
16
4(91)
19
7(315)
10
(44 8)
12
9(7 6)
23
6(26 2)
20.1 (58
TNC, total TNC increased from flowering
maturity under drought stress. ICMVIS
amount of TNC in the
8)
0(535)
3)
to FL+9 and
85332 showed
a
stable
plant throughout grain filling (Table 6.2). CIVT accumulated
TNC in the panicle up to FL+9, after that values decreased in all
except the panicle which kept the
TNC in all plant parts between
same content
flowering
decrease in all stem parts and tiller
panicle. Partitioning
in
in the lower stem and also in the
Table 6.2. Total g TNC per plant at three
parenthesis Percentage
panicle.
in all other
grain filling. The corresponding partitioning coefficient
of TNC after anthesis to the
only
found for the absolute amount
were
increased in the
increase of stored TNC's took
tiller in the second half of
and
generally
in the control. Table 6.2 shows total TNC at the three
plant parts, including the tiller, changes
an
In tillers TNC content
stages and the relative amount of TNC allocated in the
After anthesis TNC almost
P3Kolo
(ICMVIS 85332),
maturity.
Results similar to the relative TNC content
of TNC in the
for all
and FL+9, followed
accompanied by further
of absolute TNC to the
plant parts
(Figure 6.3). P3Kolo increased
panicle
from
by
a
dramatic
accumulation in the
flowering to maturity was
57
Figure 6.2. Concentration total nonstructural carbohydrates (TNC) in shoots of
at
single plants of three genotypes (ICMVIS 85332 (IC 32), CIVT and P3Kolo)
flowenng (a), flowering + 9 days (b) and maturity (c) under control (C) and
Sadore
drought stress (S) Standard errors of means are given in parenthesis.
1991.
58
TNC
(g)
Legend
(015)
mmm
IO
Tiller
B^a
Panicle
CD Upper
item
EMS
Middle stem
Hi
Lower stem
iMsiia
1
IC 32
Figure 6.3. Distribution of total nonstructural carbohydrates (TNC) in shoots of
single plants of three genotypes (ICMVIS 85332 (IC 32), CIVT and P3Kolo) at
flowenng (a), flowering + 9 days (b) and maturity (c) under control (C) and
drought stress (S). Standard errors of means are given in parenthesis. Sadore
1991.
59
much
higher
than
for ICMVIS 85332, 2.68 for CIVT and 1.47 for
unity (1.04
P3Kolo).
6.3.3.
The
Carbohydrates Under Unlimited Versus Limited Tillering
variety CIVT
unlimited tillering
irrigated
a
weight
growth
tendency
weight.
was
flowering
to
maturity
irrigated
of
(Figure 6.4).
was
There
plant parts
in all
generally
growth
well.
plant parts
nil.
During
or
parts under well-
was a
as
only: C)
tiller
the
treatments
tendency for
By
that time
in UC.
same
period
observed that the main stem parts under UC increased their
between C and UC under
stress was
between the two
non-significantly higher
was no
was
did
in all other
maturity there
weight
At
from
(one
matter of the main stem
significantly
slightly higher dry weights
was
studied with limited
flowering dry
tiller
course
relative TNC content
Tiller
additionally
At
conditions did not differ
C and UC, but of
UC for
was
(UC).
significantly
conditions. Under
sample dry weight
in
drought
stress the tiller
decreased in C. The tendency of the response to
for UC
equal
significant difference
as
to the
higher dry weights for UC
was
plant
with controlled tiller
observed for all
treatments. The TNC content tended to be
Under drought stress this difference
was
drought
A
tendency
in both
inigation
growth.
plant parts
dry
higher for C, especially for the panicle.
much smaller and not consistent
(Figure
6.5).
Partitioning
but
this
only
dry
represented
partitioning
dry
of
matter to the
0.45 for UC under
0.96 for
coefficients
panicle after flowering
well-irrigated
C, but
more
were more
matter and absolute TNC.
than
was
0.81 for C
(see 6.3.2)
conditions. For absolute TNC however
than 1 for UC. Under
unity for both growth
drought
stress the
treatments and both
60
dry weight (g)
200
150
100
150
-
100
Figure 6 4. Dry weight of CIVT with one (C) or all (UC) tillers under control (left)
drought stress (nght) at flowering (a) and matunty (b) Standard errors of
means are given in parenthesis Sador6 1991
and
61
TNC
(g)
Figure 6.5
(UC) tillers
carbohydrates (TNC) of CIVT with one (C) or all
and drought stress (nght) at flowenng (a) and
of means are given in parenthesis Sadore 1991
Total nonstructural
under control
matunty (b) Standard
(left)
errors
62
6.4. Discussion
High temperatures (Squire 1989)
a
major
influence
on
respectively. Ferraris
(Powell
was
etal.
probably
and
partitioning
Charles-Edwards
and
be the
might
presented study
inter-plant competition
is assumed to have
supply for grain filling,
(1986a) found for sorghum
linear function of cumulative radiation and
genotypic specific
content. This
and
the assimilate
relatively
are
The
1991).
reason
lower. No
low
the contents of
why
compared
planting density
was
to similar
high
carbohydrate
carbohydrates
experiments
and therefore
in
intercepted light
measurements were taken in the canopy.
light
and TNC contents in individual
more
plant parts
matter and TNC to the
stress treatment. The
higher
tillering
additional panicles under
under normal and
For cereals the
panicle
in the
amount of biomass
can
change
the
well-irrigated
nor
the
produced by
have the dual purpose to
optimum conditions
weight
of
strategy
the
drought
plant due
to
changes
and to serve as a
transitory
with
source
and translocation of assimilates after
production, partitioning
is therefore
presented describe
under
supply the plant
suboptimum conditions.
flowering determines grain filling.
a
Its
for
responsible
capacity
grain
and
mass
flexibility
and final
under environmental
grain yield. The results
detailed pattern of assimilate distribution of three millet lines
well-irrigated conditions (potential yield) and under postflowering drought
stress. The observation of
between 65% and
partitioned
to the
plant parts
100% of the
panicle
under
for
dry
dry weight accumulation showed that
matter
well-irrigated
produced after flowering
conditions. This
was
corresponds with
study of Muchow and Wilson (1976) and Goldsworthy (1970) that partitioning
of assimilates in
conditions
higher than
a
but did not
tillers also increased the amount of carbohydrates available for mobilization.
This indicates that
the
in the
sorghum
The removal of all but the first tiller influenced the absolute amount of dry
partitioning dry
a
sorghum
partitioning
partitioning
of
was
dry
about 1
matter to the
after anthesis.
Under
panicle after flowering
well-irrigated
was
of absolute TNC for ICMVIS 85332 and P3Kolo
transformation of nonstructural
carbohydrates
into structural
slightly
indicating
carbohydrates.
63
A
high partitioning coefficient
period
better
and
grain filling
about 50% of the
higher grain
panicle weight
while
flowering
after
reserves
drought
conditions for the
under
was
drought
under
same amount as
grain filling
necessarily
stress. In this
mean
genotype
attributable to remobilization from
the rest
was
translocated from current
stress. CIVT reduced in the same
biomass
P3Kolo with the lowest
mass
increase
only
photosynthesis
by the
well-irrigated
under
genotype ICMVIS 85332 however did not
for the
as
period
its total
panicle weight increased (partitioning 1.93).
the
partitioning coefficient
under
well-irrigated conditions
(therefore the highest vegetative growth after flowering) needed only 15% of its
panicle weight
strategies
In the
the
from remobilization under
exist to sustain
present study there
panicle
under
grain growth
was a constant
current
photosynthates
their level of TNC
even
are
with their
while the rest is translocated to the
of assimilates in millet with
after
flowering
under
C,4 that
current
found that TNC in wheat
irrigated conditions
reserves
function
the
photosynthate production by
the
through respiration and remobilization,
No
place
in TNC level in the stem
change
panicle
could be
photosynthate.
to the
were
buffer for a
photosynthates
grain filling. After
that the sink demand of the
findings
of
McCaig
and Clarke
(1982).
accumulated in the stem even under well-
to a maximum at
as a
exploited
for
mobilization of stem assimilates took
no
satisfied with the
They
plant parts except
This suggests that
panicle. Jacquinot (1970) showed by labelling
means
correspond
Seemingly, different
stress.
distributed among stem parts to maintain
use
parts during grain filling also
This however does not
drought
were not
optimum conditions.
partitioned
stress.
level of TNC in all
well-irrigated conditions.
stored in stem parts before anthesis
flowering,
drought
under
anthesis, whereafter they decline. These
phase difference
plant
between time of maximum
and the time of maximum
requirement by
developing grain.
Jacquinot (1970) found that maximal translocation in millet of
thesis into the
found in this
maturity
was
panicle
study.
actually
takes
place about ten days
after
current
photosyn¬
The
same was
flowering.
The absolute amount of TNC accumulated from FL+9 to
much
higher than during
the first
period
of
grain filling. This
64
however seems not to affect
grain growth potential
is almost linear
1992) but
(Bieler
ef al.
as
its
is very small.
panicle
confirmed for the three genotypes. The
TNC to the
panicle
after
photosynthesis
unimportant.
of the
was
importance
leaves is
a
major part
corresponds
of current
under
genotypes had
was
reduced
they
occur
matter to the
in maize due to
drought
is
more
confirmed that
a
likely to
of assimilates is limited. In
grain yield
confirmed in
contribution
of
decreased with
Genotypic
very
ability
to
under
important
that low leaf water
potentials
in
large reductions
extent
to
by
the storage
by Rao
et al.
photosyn¬
on
current
(1975a and 1975b)
capacity
assimilates to the
important
as
compensation by translocation
Fischer and Wilson
is not limited
some
drought
was
grain filling, expressed
exceeds the rate of total
panicle,
be reduced when
sorghum
carbohydrates
differences
photosynthates
was
of
net redistribution of stored assimilates.
stress result in
transport system moving
was
senescence
grain yield which largely depends
assimilates should therefore be very
This
a
by maturity. Boyer (1976) found
photosynthesis
by
was
because
similar level of TNC in all stem parts at flowering that
a
He concludes that
the
grain filling, late
stress. The rate of
necessitating
matter accumulation
thesis.
vegetative growth
as
preanthesis nonstructural carbohydrates
postflowering drought
by the partitioning coefficient of dry
All three
for
photosynthesis
Mobilization and translocation of
dry
panicle
very desirable trait.
a
important
partitioning
for ICMVIS 85332 and CIVT
sorghum (Goldsworthy 1970),
to
to the
coefficients of
of the available assimilates of current
translocated into the
The same
lag phase when
Genotypic differences of partitioning of TNC
flowering showed that
almost all and for P3Kolo
matter accumulation
could be due to the initial
grain growth
were
dry
of the
panicle.
for millet under
grains,
drought
(1978). They concluded
grain filling by
various
nor
Accumulated
millet
stress.
that the
plant parts
stress while the contribution from stem sugars increased.
were
found in this
study
postflowering drought
for P3Kolo and in
in the
stress. The
some
ability of using
panicle
as a
current
sink for TNC
extent for ICMVIS 85332 while its
compete with other sinks (respiration, translocation into roots, transfor¬
mation in structural
dry matter from
carbohydrates) was
flowering
to
maturity
rather low for CIVT. The reduction in total
for CIVT, but the
changes
of absolute TNC
65
in all
plant parts
drought
shows different responses to reduced
stress. All these observations are
drought response expressed
best
ability
the lowest
to
in
supported by
loss.
genotypic
under
different
grain yield. The panicle of P3Kolo showed the
compete with other sinks under drought
grain-yield
photosynthesis
the
stress and therefore had
66
7. Further
Aspects
Development After Flowering
of Plant
7.1. Introduction
Additional
potential
but not due to
drought
role. The later
a
stress, the relative
yield
produced.
is
drought tolerance
a
Bidinger 1986). For
of tillers takes
Their relative
an
inverse
grain filling
development compared to
study of the developmental strategy of
necessary. Therefore the individual
The
development
flowering compared
by
vegetative growth
panicle
yield performance
the
availability
of
millet
genotypes is
of genotypes has to be
of tillers, their number and
to the main stem
to the total
influenced
be
development
and
yield
Bidinger, 1985a),
panicle is therefore important. To understand genotypic differences
the main stem
importance
(Mahalakshmi
reduction in
a
and
tiller flowers, the less time and water is available for
and therefore the less
observed.
compensate for
can
stress
postflowering drought stress (Mahalakshmi
postflowering drought
in
panicles
set of tiller
grain
due to mid-season
synchronization of
has to be monitored to
a
judge their
genotype. Their performance
of soil water and its utilization
can
millet
by
genotypes.
The
extract soil water for
ability to
of the root system. The water
plant growth
supply
is
of the crop
dependent
on
throughout
the
the
development
growing period
influences the vertical distribution of the root system. Blum and Ritchie
found that in
is
sorghum
restricted
penetration
of
crown
compensated by enhanced development of existing
length is comparably little affected. Chaudhuri
that millet has
to
roots in the
a
higher
root
sorghum. They observed
depth
of 300
weight
cm.
roots.
Thus, total root
and Kanemasu
(1985) describe
spread, and higher
that millet is
water use
down to
180
efficiency compared
capable of extracting
uptake takes place and that
cm
only.
in
beyond
a
sandy
The amount of roots
top soil layers where the
soils millet root
produced
ability
to access the
remaining
growth goes
under well-watered
conditions and the location of the root system in the soil
indications of the
water
Nevertheless, Chopart (1983) reported that most of the root dry
of millet, like that of other field crops, is in the
main nutrient
dry
(1984)
soil surface
profile
can
give
water in the soil once the rains
67
stopped.
have
This
cal
and discusses the results of
chapter presents
phenological
observations for five genotypes under
development
a
and
and
a
above-ground
stress treatment. Observations include
postflowering drought
morphologi¬
well-irrigated
development under well-watered and drought stressed conditions, estimation of
yield potential
well
as
measurements are
description
as a
complemented by
Problems and constraint of millet
of the root system of four genotypes. Root
neutron
measurements of soil water.
probe
drought research
in the Sahel
are
discussed.
7.2. Materials and Methods
Experimental Design
7.2.1.
The trials
were
repetitions
irrigation
drought
of 5
stress treatment.
50% of the
drought
0.4
measurements
a
well-irrigated
Sub-plots
apart. For the
50% of the
plants
split-plot design
treatment and
a
were
In 1991 five
postflowering
were genotypes which consisted in 4 rows
stress
plots, irrigation
was
stopped
flowering, simulating
reached
planted by
hill in 1990 and to
only
with six
plots represented the
hand
on
ridges 0.75
one
plant
when in
a
postflowering
m
apart, in hills
apart in 1990 and 1991, respectively. Thinning
m
plants per
Synth-2.
m
stress. The trials were
and 0.25
m
three
0.75
plots
in
consisting
treatment
in a
in 1991. The main
repetitions
in 1990 and ten
length,
m
planted during the hot off-season
was
done to
per hill in 1991. In 1990 root
carried out for the two genotypes ICMVIS 86327 and
genotypes
were
planted (ICMVIS 85327, HKP, ICMVIS
85332, CIVT and P3Kolo) for vegetative growth observations, root measurements
were
done for the first two
for the trial in 1990
chapter
are
genotypes. Fertilizer application
further
explained
in
chapter
and crop treatments
4.2 and for the trial 1991 in
6.2. The climatic environments are shown in Annex I.
68
7.2.2.
Vegetative Growth Observations
To determine the
sampled starting
up to
growth of five genotypes (in
the 48th day after
maturity. The plants
Green leaf
biomass
area was
was
measured
determined after
For the observation of the
recorded. Tillers
chronology
drying the samples
was
taken
tagged
weekly
two middle rows.
and total
above-ground
at 70°C for 48 hours.
thinning
after
were
stems of 14
flowering date
and their
and monitored at their flowering date and numbered
of appearance per
plant.
stems, first, second, third, and fourth tillers
and 100
plant per plot
sampled randomly from the
(Li-Cor 3100 Area Meter)
were
tagged
were
sowing.
only), individual plants
development of individual plants the main
plants per plot (yield plot)
after their
were
1991
One
grain weight determined
and
At
were
maturity, the panicles of main
harvested
panicle yield
and
separately, their yield
grain number per panicle
derived.
7.2.3. Root Growth Observations
In both trials 1990 and in 1991
flowering
soil
samples
cutting
described above root
depth
in the
row.
depth
and
plants)
and
were
Samples
deep-frozen.
on
were
the hill
(after cutting
taken at 10
dry weight
was
length
plot, i.e.,
was
cm
diagonal
the
one
and at 20
were
on
drying
taken at
diameter
(after
steps from 20-260
were
taken at three
cm
and 38
cm
steps from
20-100
cm
plant),
cm
carefully
cm
the hill itself
cm
19
washed
measured with sectional
determined after
were
to the next hill in 1990.
depth. In 1991 samples
kept separate. Samples
True root
half the
taken at 10 cm and at 20
and mixed for each
plot, i.e.,
on
samples
aluminum tube of 7.8
an
taken at two locations per
both locations
locations per
root
were
the hill of three
Samples of
cm
as
of two genotypes each year. With
and the
roots
tracing paper. The
for 24 hours at 70°C.
69
Neutron Probe Measurements
7.2.4.
(ICMVIS 86330, 86313, 85321, 86327, Synth-1, Synth-2,
For nine genotypes
CIVT, P3Kolo and Sadore local) in 1990 and for the five genotypes planted in
1991
(see 7.2.1)
stress
neutron
plots only
steps of 20
cm
1991.
For 15
from 30-210
cm
tubes
installed after
planting
probe (Troxler) readings
the water content
derived
was
in the
taken in
were
the last
in 1990, and twice in
calibrated and the
was
were
depth three days after
weekly intervals
depth
cm
probe
access
and covered. Neutron
afterwards thrice at
neutron
probe
irrigation and
weekly intervals in
gravimetrically. The
readings converted
to
relative water
content.
7.3. Results
7.3.1.
Vegetative
Leaf Area Index
and Generative
(LAI)
was
Development,
highest at flowering
filling (Figure 7.1). Drought
and
and Yield Performance
rapidly decreased during grain
stress enhanced leaf
senescence
from
flowering
onwards while under well-irrigated conditions leaf senescence occurred about one
week later. There
senescence
were no
neither under
differences between genotypes in the rate of leaf
well-irrigated
average values of genotypes
The results in the present
(r=0.98)
in the
irrigated
conditions
and 18
days
above-ground
after
The
period
after
are
study
from 50
nor
stress
show
days
an
after
almost linear
flowering (59 DAS).
dry
sowing (DAS)
(Figure 7.1). The interval covered
biomass under
conditions, therefore
presented.
The
buildup
ten
of
matter accumulation
to 78 DAS under well-
days before flowering
dry weight of the
drought stress conditions stopped
flowering, then dry weight possibly
dry
drought
even
declined
during
about
late
matter accumulation could therefore be described as linear
to 64 DAS
(r=0.99).
one
total
week
grain filling.
only
from 50
70
At
a
tiller
plant density of
panicles
with
panicles per plant
in the
conditions and under
a
well-irrigated
drought
produced
produced
stress conditions
a
a
high
average of 0.75
variation
no
fourth tiller with
respectively.
significant difference
a
first tiller
for the
two
irrigation
ability
no
grains under
of
producing
irrigation
such
significant
treatments.
a
LAI
analysis could be
g/plant
£OU
=
0175
-200
•>.
o
-150
-
\
j/
-100
1
*
Flowering
-
-50
-
1
SE (DM)
(3
*
7 94
=
i
i
i
i
i
50
57
64
71
78
0
85
DAS
control
*
LAI
°
DM
first
treatments.
carried out.
SE (LAI)
a
well-irrigated
There were
tiller of the five genotypes could be identified in either of the
Due to the low rate of second, third and fourth tiller
only
second tiller, 6% and 3%
(P<0.001) differences for all tiller numbers between the
Due to
was an
treatment and 0.55 tiller
drought stress conditions respectively, produced
third tiller, and 1% each
and
irrigated
stress treatment. Table 7.1 shows that
Sixteen percent and 10%
grains.
produced
drought
in the
of 48% and 37% of the main stems under
genotypic average
with
panicles m'2, there
5.3 main stem
grains per plant
stress
Figure 7.1. Green leaf area index (LAI) and dry matter (DM) during flowering
early grain filling, average of five genotypes. Sador6 1991.
and
71
flowering date for main
Mean
Having
no
significant differences
all lines suffered
drought
distributions of all
son
of the
averages
only
as
the
an
in average
during
were not
the
was
days
59
flowering date
same
for both treatments.
nor
physiological period. Flowering
sizes
lagging
maturity,
are
compari¬
presented
different
the
as
The
(Table 7.1).
behind most, with an average of 64
second and third tillers flowered earlier than first tillers
days
average of 62
date of
different between genotypes. The
respective sample
of first tillers was
under control and 61
panicles
distribution of first to fourth tillers is
days (Figure 7.2), while
with
stress
panicles
flowering
flowering
average
stem
days
in both treatments. Fourth tillers flowered at 65
after
sowing
under
drought
days
stress.
production of tillers with grains first, second, third, and fourth grade
T4) per genotype and inigation treatment. Sadore 1991.
Table 7.1. Relative
(T1
-
well-irrigated
drought stressed
Main stems with:
Main stems with:
T1
T2
T3
T1
T2
T3
ICMVIS 85332
0.48
0.17
0.08
0.28
0.10
0.04
HKP
0.46
0.13
0.05
0.36
0.11
0.03
ICMVIS 85327
0.50
0.16
0.08
0.37
0.09
0.03
CIVT
0.55
0.18
0.06
0.48
0.13
0.02
P3KOIO
0.41
0.15
0.04
0.36
0.08
0.05
CV%
29
.
.
37
_
_
Total
grain yield
was
significantly
reduced under
drought
stress
by 47% for
ICMVIS 85332, 44% for HKP, 53% for ICMVIS 85327, 51% for CIVT and 25%
for P3Kolo.
irrigation
Figure
7.3 shows
treatments
panicles contributed
was
79% under
to total
according
grain yield
stress conditions
decreased
of individual
genotypes under both
to main stem and tillers. While the main stem
67% to the total
drought
grain yield
grain yield
under
(Figure 7.4).
irrigated
conditions this
The contribution of tillers
dramatically (Table 7.2).
First tillers made
an
72
average of 23% and 15%, second tillers 7% and 4% of the total
well-watered and
mass
and
drought
stress treatment
respectively (Figure 7.4). Both grain
grain number per panicle of tillers decreased
the main stem
(Table 7.2).
lowest loss due to
drought
genotypes. Since there
representing only
a
P3Kolo had the lowest
stress in any of the
small
were
only
a
sample size,
few
grain yield under
to a
larger
extent than of
yield potential but also the
parameters compared to the other
plants
a mean or
with third and fourth tillers,
relative reduction
in Table 7.2 was not calculated. Their total contribution to the total
unimportant (1.8%
and 0.4%
respectively
under
as
presented
grain yield is
well-irrigated conditions).
Days after Sowing
Figure 7.2. Flowering distribution of five genotypes according to main stem (MS)
(T1-T4) under control (c) and drought stress (s) conditions. Sadore
and tillers
1991.
The
grain yield
grain
loss due to
number and less to
first tillers, but inverse for
a
drought
stress was
reduced
panicle
mostly
grain weight
due to
a
reduced
for main stem and
of second to fourth tillers
(data
not
panicle
panicles
of
presented).
73
50
100
200
150
Grain yield
Main Stem
EH
Tiller 1
Tiller 3
SU
Tiller 4
300
250
(g m-2)
Tiller 2
Figure 7.3. Grain yield of genotypes (G1: ICMVIS 85332; G2: HKP; G3: ICMVIS
85327; G4: CIVT; G5: P3Kolo) under control (C) and drought stress (S)
conditions. Sador6 1991.
MS 79%
MS 67%
Tl 15%
T1 23%
Figure 7.4. Average grain yield distribution according to main stem (MS)
(T1-T4) under control (left) and drought stress (right). Sadore 1991.
tillers
and
74
yield, grain weight and number of grains per
panicle, of five genotypes of the main stem, first and second tiller
panicle (T1-T2)' under well-watered conditions and the relative
reduction (%) under drought stress given in parenthesis. Sadore
Table 7.2. Panicle
1991.
Main stems
T1
T2
ICMVIS 85332
31.9(35)
21.4(53)
18.7(37)
HKP
34.1(41)
22.6(51)
18.8(45)
ICMVIS 85327
34.6(44)
21.9(66)
19.8(49)
CIVT
29.2(37)
23.2(71)
15.0(51)
P3Kolo
30.4(17)
24.3(39)
19.3(38)
Panicle
100
yield (g)
grain weight (mg)
ICMVIS 85332
896(12)
771 (27)
725
(22)
HKP
930
(22)
810 (35)
759
(19)
ICMVIS 85327
892
(21)
783
(31)
716(28)
CIVT
965 (27)
861
(37)
793
P3Kok)
935
834
(25)
780 (33)
Grain number
(10)
panicle
ICMVIS 85332
3520(28)
HKP
3660
ICMVIS 85327
3880
CIVT
P3Kob
3200
*
(45)
2730(38)
2560(25)
(32)
2770
2490
(32)
2790 (58)
3030(20)
2700(63)
1870(38)
2920
2440
(6)
T3 and T4 have been omitted because of
(37)
(20)
a
(40)
2820 (41)
small
sample
(12)
size
(see Table 7.1).
7.3.2. Root Growth
Total above-ground biomass at
as
well
as
grain yield
under both
in the root observation
in
1990 was
drought
study
for 1991
only) and at harvest
treatments of the
genotypes included
flowering (available
are
stressed
irrigation
shown in Table 7.3. As
reported
in 4.3.2 the trial
just before flowering affecting both irrigation
75
treatments.
Synth-2,
However ICMVIS 86327 showed a better
had
higher grain yield
a
grain yield
reduction under
genotypes
in 1991
was
under
drought
the
same
stress. The biomass
until
flowering,
vegetative growth than ICMVIS 85327 only
yield
was
drought
slightly
stress
vegetative growth than
well-irrigated conditions
under
produced by
HKP had
drought
and also less
a
stress.
the two
slightly better
Although grain
lower for HKP under well-irrigated conditions its reduction under
(44%)
was
less than for ICMVIS 85327
(53%).
grain yield of the genotypes included in the root
(C) and drought stress (S).
drought stress are presented in parenthesis. Sadore 1990
Table 7.3. Shoot biomass and
observation studies under well-watered conditions
Relative values under
and 1991.
Biomass at
Genotype
Year
flowering
ICMVIS 86327
Synth-2
1991
S
C
S
-
602
569 (95%)
145
136
(94%)
-
597
541 (90%)
123
96
(78%)
HKP
642
1110
750 (67%)
251
140 (56%)
ICMVIS 85327
641
1090
690 (63%)
272
128
The distribution of root
dry weight and
root
length along the soil profile
two
x
depth interactions could be shown.
genotypes
in
depth
both genotypes to
taken. The
a
of root
depth
penetration
of 280
cm
In 1990
was
beyond
rooting density
decreasing rapidly
89% of the total root
mass
same
further
were
found for
samples
were
be
expected since the
increasing depth.
In 1990, 86% and
respective
depth (Figure 7.5).
no
were to
of ICMVIS 85327 and
observed in the top 100 cm, while the
and 72% for the
with
at anthesis
difference between the
identified. Roots
which
significant differences between depths
is
no
(47%)
statistically significant
followed similar pattern for all genotypes in both years. No
genotype
yield g/m2
g/m2
C
1990
Grain
Biomass at harvest
g/nf
The
Synth-2 respectively
values for root
same can
length
was
were 70%
be observed for the root
76
density expressed
Specific
as
mg/cm3
soil
or
cm/cm3
It differed between the two genotypes in 20
showed
a
a
higher,
soil
as
it is shown in Annex III.
increased with depth in the top soil
length rapidly
root
but very variable
very fine root system. For
specific
cm to
root
100
length
deeper soil layers
cm
layers (Figure 7.6).
depth. ICMVIS
in these
the values
86327
depths expressing
were
similar for both
genotypes.
Cumulative total root length/volume (%)
^^^~^^
a)
80
Jy/
60
40
-
if
20
6
I
I
I
i
i
I
1
I
I
i
1
10
30
50
70
90
110
130
150
170
190
210
+
+
170
190
I
1
230 250
270
Depth
Cumulative total root dry wt/volume (%)
+
i„\
+
t__i-
+.—•
*
80
60
40
20
5
i
i
\
i
i
i
i
i
10
30
50
70
900
110
130
150
210
230 250 270
Depth (cm)
—
Figure
ICMVIS 86327
7.5. Cumulative root
genotypes. Sadore 1990.
+
Synth-2
density. Length (a) and dry weight (b) of
two
77
length densities
The root
in
plant.
length density is highest
7.7. Root
Figure
Within
of the two genotypes observed in 1991
short distance from the
a
very low in the top soil. From
and between
and
a
sharp
total root
the
top
rows,
depth
dry
more
rows.
40
cm
just
presented
beneath the
length density
is
already
downward differences within
cm
row
because of some decline in values between rows
The
dry
matter contribution between rows to the
(Annex IV). The major part of
specific
root
length
root
dry weight
is in
tended to be higher in between
(Figure 7.8). Generally, values increased with
in the top soil
rapidly
depth differences within
within rows; at 100 cm
and in
disappeared.
had
rows
The
the root
plant
depth of
rows.
is minimal
mass
especially
between
disappeared
rows
decline within
soil within
a
in the top 10
are
1000 cm/g root dry weight
qI
10
1
i
i
i
i
i
i
i
i
i
1
1
i
20
40
60
80
100
120
140
160
180
200
220
240
260
Depth (cm)
—
Figure
7.6.
Specific
root
ICMVIS 86327
length
of two
—<-
Synth-2
genotypes. Sadorg 1990.
280
78
1000
cm/g root dry weight
20
0
80
60
40
120
100
Depth (cm)
—
Figure
7.7. Root
and 38
cm
(D3)
D 1
length density
for
+
D 2
-*--
D 3
at interrow distances of 0 cm
(D1),
19 cm
(D2)
genotype HKP (a) and ICMVIS 85327 (b). Sadore 1991.
79
cm/cm3 soil
a)
\
0,8
0,6
0,4
-
I
+
I
*
0,2
I
I
0
i
1,4
b)
\
1,2
1
0,8
0,6
0,4
+
*
*..._
0,2
I
0
0
20
40
60
80
100
Depth (cm)
d 1
Figure
7.8.
and 38 cm
D 2
'
D 3
Specific root length at interrow distances of 0 cm (D1), 19
(D3) for genotype HKP (a) and ICMVIS 85327 (b). Sadore
cm
(02)
1991.
80
7.3.3. Neutron Probe Results
suffered
experiment
In 1990 the
treatment due to technical
only affecting individual plots
In the
drought
over
This
was
a
in
This
showed
evaporation
a
was
large
planted
water
this
on
field and
depth
Bray
was
irrigation
probable disturbing
moving inigation
experiments
were
determined to
pH
see
whether they influence field
in water did not vary much across the
respectively. Phosphorus
as
cm
and 4.8 at 100
cm
determined after the method
1 showed an average of 24.2 ppm, 13.3 ppm and 1.0 ppm in 1990 and 18.8
available in
excess
in 1991
11.3, and 9.8
plots
plots.
the field. In 1991
influence. The two
5.4 and 4.9 in 20 cm, 4.7 and 4.5 in 60
in 1990 and 1991
(determined
-
respectively,
the same field in the years 1990 and 1991.
depth of 20,60, and 100
After the recommendations for these soils
was
measurements
at the rate of the pan
across
the direction of the linear
ppm, 2.2 ppm and 1 ppm in 1991 for the
ly.
probe
in 1990 and 1991
application through irrigation
was across
as a
again
The trial in
stress from one side to the other of the field.
micro-environment. Soil
a
missing plots.
range of different volumetric water content between
Therefore soil characteristics
in
some
visibly variable drought-intensity.
assumed to be sufficient and uniform
system, excluding
capacity
stress in the 'well-watered'
drought
all blocks and resulted in
after the last
days
gradient
the observed
not
of
irrigation. This drought occurred in patches
obvious from visual observation. The first neutron
happened although
were
a
strong gradient in
taken three and four
already
in
stress treatment, stress effects were more severe, but
unevenly distributed resulting
1991 showed
slightly
problems
-
in all
only).
13.3 %
plots. Soil
texture varied
For the three same
clay
depths
content. Plots with
genotypic differences of
no
found. The fourth neutron
probe reading
the third
reading
in 1991
(18 days
in 1990
after the last
slightly
respective¬
the field
across
it varied from 4.3
higher clay
with low volumetric water contents after the last
Due to these variabilities
cm,
(Bationo etal. 1991) phosphorous
fractions
-
6.3, 9.0
were
the
irrigation.
water extraction could be
(24 days after last irrigation) and
irrigation) however showed quite
uniform water contents over the whole field. The volumetric water content at
harvest
as
shown in
Figure
7.9 for 1991
(FL+17)
varied from 1% in 10 cm
depth,
81
and 4%
7% from 20
-
200
-
cm
assuming that wilting point
for this soil must be
in this range. The different amount of water extracted in each
to have an influence on
grain yield, but could
not be
plot
assumed
was
proved statistically.
Vol % H20
Depth (cm)
0
4
2
8
6
0
*-tr
50
*+
-
*+
100
—
+
-
*
150
-
FL-3
*
\
\
FL-10
FL.17
*
t
*
200
1
/
\
*
t
4-
I
1
one
week
*
250
Figure 7.9. Volumetric water content three days after flowering (FL+3),
(FL+10) and two weeks (FL+17) later. Sadore 1991.
7.4. Discussion
According
described
days
an
Bramel-Cox et al.
by two phases: First
flowering,
a
(1984) the growth pattern of
linear accumulation of
and then either
a
by
15
change,
or
continued increase,
study
the present data. The maximum of LAI at
of
normal leaf
Carberry
and
senescence
ing upwards,
was
Campbell (1985).
after
probably largely
made up
no
further
two
phases
could be
flowering corroborated with
The loss of
flowering, starting
millet can be
dry weight until 10 to
apparent decrease. No clear distinction between these
made
the
after
to
dry
matter because of
with lower leaves and progress¬
by the accelerated grain
mass
82
accumulation, resulting in
stress the
crops
rapid decline
being due
as
leaf senescence,
induction of
green leaf
continued increase of total
to low water
and
low leaf water
area
plants'
was
the
photosynthetic
potential caused
index in maize
can
be
(Jurgens
Under
resulted in
an
seen as
active
drought
for
grain
surface.
1978). Reduced leaf
insufficient assimilate
The
area
and
supply
dry matter, starting
for
about ten
compensatory effect.
The low tiller number per main stem is due to the
and the
mass.
almost immediate decrease in
an
et al.
basic maintenance. A decrease of total
days after flowering
dry
explained by Boyer (1976)
potentials reducing photosynthesis, enhancing
reducing
so
photosynthesis together
reduced
the
a
a
of leaf area
high off-season temperatures
that
(Ferraris and Charles-Edwards 1986b).
are
The
relatively high planting density
both constraints to
why tillers
reason
tillering ability
of third and fourth
grade showed earlier flowering date than first tillers can probably be explained by
the fact that
only early plants
tillers. Plants with late
with
appearing
however could not be confirmed
sample
size of fourth
eariy
tillers will
first tillers do not
by
correlation
produce
a
high
produce additional
analysis, probably
number of
ones.
due to
a
This
small
grade tillers.
The lower number of tiller
due to the fact that tillers
panicles
were
with
affected
grains under drought
by drought
in
an
stress is
earlier
possibly
stage than main
stems and that assimilates in the tillers were translocated to the main stem. The
relatively higher grain yield
loss of tiller than main stem
panicles confirmed the
importance of the strategy of development of genotypes. The contribution
of
one
third to the total
role of tillers in
grain yield
under
expressing yield.
determining for plant development.
panicle and
lakshmi etal.
proves.
in
panicles
1986)
Genotypic
Millet root
(1983)
tiller
but
can
is
well-irrigated
Environmental influences
A
long flowering
be disastrous for
yet
therefore be
range between main stem
stress
postflowering drought
as
(Maha¬
this
study
to be identified.
growth has been intensively studied for
Senegal. Although
can
important for midseason drought
differences however have
of tillers
conditions underlines the
one
genotype by Chopart
results of root studies are very difficult to compare
due to differences in the method of root
sampling, Chopart's findings correspond
83
quite
well with the observations made in this
environment of Sadore
depth
It
(Niger).
than 280
found in this
depth
probable.
Chaudhuri and Kanemasu
extracting
water
cm.
beyond
mass
is
higher specific
to be
reported
root
profile
confirm
length
length
fluctuating
lot
a
root mass as there were
cm.
deeper soil
in
is located in the top 40
along
exploitable
than has root
This suggests that millet
layers
the soil
than
were
profile especially
a
rows
gets
more
uniform with
water has a much lower horizontal
length density.
just
indicated
100 cm as an indicator of a
relation of biomass
It is concluded that the
as
probably
grain yield
or
tendencies, i.e., specific
more
to root
gradient in
possibility
stress
root
growth
conditions
takes
place mainly
Schmidhalter
and
generally favored
maintain
growth
over
shoot
allowing
the
an
sandy
slow
adaptation
soils as in the
adaptation
is less
(1984)
was
of the
ef al.
on
top
study.
flowering,
Under
drought
increased
an
plants
1991), since
than shoots
was
roots can
(Westgate
and
the soil characteristics,
decreasing soil
present study drought
likely.
depth.
of young cereal
potentials
to the
in the
extraction. No
found that after
largely dependent
plant
length
use
the soil
of water
Genotypic
rows.
(1987) described
growth
growth (Schmidhalter
under lower tissue water
Boyer 1985b). This however
a
in
depth.
could be indicated in this
in 90-120 cm
Oertli
assimilate accumulation in roots. Root
root
efficient water
growth
Under moisture stress Blum and Ritchie
sorghum
sorghum. Root
fact that results here confirm. The
utilization and nutrient uptake is similar within and between
differences
by Chopart. For
that the distance of the roots from each other which expresses
Chopart explains
of
capable
report
confirmed for millet
are
preanthesis development (Chopart 1983),
potential
is not very
Nagarajarao (1981)
study only for
and
distribution of roots within and between
the
penetration
flowering
sorghum Myers (1980) reported 86-87% of
found in the top 40
were
after
root
that millet is
however
Misra and
m.
to this
depth
same
49-50%
length only
has a much
(1985)
depth of 3
figures correspond
These
density
a
study
of roots to
penetration
that the
and 77-78% of the total root
found 82-85% in the
root
cm as
down to 300 cm. In
rooting depth of
the total root
reported
was
genotypes in the
two
rapidly after panicle development. Therefore,
slows down
to further
of
study for
water
potentials.
In
stress occurs fast, therefore such
84
The first neutron
probe readings after
initial soil water content among the
variability. This effect was regarded
ment, since the pattern of
irrigation neither
in the
range in
clay content
tal variabilities had
an
natural
corresponding
in 1990
cm.
of
investigated is
the crop
inigation
was
exploited
growth
phase.
in the linear
or root
moving
a
very
development,
hypothesis that
whether the soil water content at
personal communication).
In all
to the same residual content up to
genotypes and initial pattern of soil
drought
nor
for texture showed
flowering
water or even residual water from last season's rain
SAT Sahelian Center,
genotype
Whether these small environmen¬
discussed before, could not be identified. An other
further
wide range of
variability in the soil microenviron-
samples analyzed
on
a
to any
could not be attributed to any mistakes in
in the top 100
influence
inigation showed
not
sprinkler type system
overhead system in 1991. Soil
narrow
plots,
as
variability
the last
plots
maturity
as
has to be
was a
storage
(Brouwer, ICRI¬
however soil water
of the crop,
water content. Therefore all
stress had a different amount of water available for the
despite
plots under
grain filling
85
8. General Discussion and Conclusions
present study the effect of postflowering drought
In the
West Africa
a
reduced
was
characterized
panicle grain
Bidinger etal. (1987a)
the
of
of
an
during
the
imposed
was
nism of
cence
in the
that reduces or
peari
millet under
filling come largely from
being exploited.
were
The
potential
in
the
mecha¬
following
inhibits
even
photosynthesis. Therefore leaf
photosynthetic
optimum
current
active surface reduced. It
conditions the assimilates
photosynthesis,
partitioning
stress results in low leaf water
stored soluble
coefficients of
was
required
shown
for
grain
carbohydrates
not
matter accumulation in the
dry
unity corresponding
senes¬
to the
1976). However, under drought stress,
findings
current
in
sorghum
photosynthe¬
important only during early grain filling until the declining assimilate
production due
grains
same
Bidinger
to fill the same number of
opportunity
grain filling period suggested
found close to
(Muchow and Wilson
sis is
and
flowering (Maiti
present study, agronomic and physiological observations
stressed
is enhanced and the
panicle
flowering
by Boyer (1976) postflowering drought
potentials
that in
stage and since the number
genotypic individual drought response.
a
As found
at
of the control and stress treatment had the
drought
stress in
genetic material of Indian origin. As drought
with
number and therefore the same
grains. Hence,
This confirms the results of
mass.
inflorescence is determined before
1981), the plants
grain
grain
millet in
pearl
reduction due to both
significant grain yield
number and
presented experiments
grains
as a
stress on
are
to
reducing photosynthesis
equal, allowing
Nonstructural
translocation of current assimilates to the
no
carbohydrates
and the demand of other sinks than
of
preanthesis photosynthesis
and translocated from the main stem and from tillers to
to some extent
confirming
the observations of Rao etal.
of remobilization results in
an
unchanged grain filling
compared to optimum conditions (Bieler etal. 1992).
found to be
primarily
due to
a
reduced
satisfy
probably
the sinks demand
(1978). This
rate under
The smaller
grain filling period
assimilates to be translocated, but most
panicle.
then mobilized
are
mechanism
drought
grain
stress
mass was
and due to exhaust of
also because of
an
early
86
collapse
assimilatory apparatus. However there
of the
grain filling by adjusting grain
assure
ment. A
signal
must be
stress than under
and
under
source
may draw
the stored and current
of tillers under
grain yield
drought
partitioning of dry
an
partitioning
ability
partitioning
of
relatively
partitioning coefficients
system
flowering
at
under
under
drought
root
can
main stem
as a
panicle
reduced than of the
more
genotypic
differences in
flowering to maturity
carbohydrates
under low leaf water
can
be
or/and
The
potentials.
stress however were not related to the
optimum
a
large
also act
optimum conditions. The description
differences that could influence
A
suggested.
carbohydrates to the panicle.
coefficient from
photosynthetic activity
coefficients under
pearl millet.
are
indicator for the mobilization and translocation of stored
an
of Reed
from the tillers. Therefore,
be found in the
can
to
environ¬
kernels under drought
flowering
carbohydrates
matter as well as of soluble
Under drought stress the
for
specific
a
grain yield. They
the earlier
as
stress is
main stem. The second mechanism
the
for
improve grain yield under optimum
can
up to one third of the total
sub-optimum conditions
on
more
the amount of stored assimilates
developing tillers
making
panicle
plant to abort
genotypic strategy
a
in maize but could not be identified in
influencing
number of fast
conditions
within the
is
optimum conditions. This is supported by the studies
Singletary (1989)
Two factors
given
number per
conditions did
drought
indicate
not
response. However
an
of the root
genotypic
adaptation
of
growth to drought stress as suggested by Schmidhalter and Oertli (1987) can
not be excluded.
Under
postflowering drought
is different than under
does not
stress the
optimum
necessarily perform better under drought
relationship
existed of
yield
under
any parameter measured under
individual genotype affected
parameters measured
and
yield development pattern
conditions. A genotype with
screening
for
drought
drought
during
drought
by drought
conditions
stress.
tolerance
rainy
season
millet
Furthermore
Therefore, grain yield of
particular
have
nurseries. However, to achieve
the
pearl
(i.e., potential yield)
cannot be estimated
in
of
high potential yield
stress.
potential yield. Seemingly, drought
drought
undertaken in
stress
as
well-irrigated
a
a
no
to
an
by yield
related
research in
general
necessarily
natural
to
be
postflowering
in an environment where rainfall is erratic
87
and the
is
photoperiod sensitivity
extremely difficult.
dry
little
only
in
flexibility
planting
date
A rainout shelter to control rainfall is therefore necessary
experiments, allowing high
season
any
of millet allows
mean
temperatures, required
or
to conduct
experiment successfully. For the latter the lack of irrigation equipment makes
further research in national
agricultural research programs
in the Sahelian
region
impossible.
According
to these
findings postflowering drought tolerance
described in the response to the above
with the
ability
In
panicle.
yield of
a
multiple genotype
response in the
correlation
(DRI)
analysis
presented
could not be
screening
stress was
in this
study,
adequately
trial
as
x
the
well
panicle grain
positive drought
was
an
at 50%
found that
flowering
as
panicle yield
flowering which excludes
of
expression
flowering
and the
drought
of 50% of the
intensity
of the
a
plots
drought
genotype due to the
or
susceptible
drought
over
in all
stress
experimental
on
the environ¬
a
drought
microvariability. Seemingly, repeated
different pattern. This results in
yield. Together with
an
a
high
unstable vegetative
(photoperiodic sensitivity) it
years. A
and Wilkinson
was
(1963)
stability analysis
found to
stress. The
yield potential
optimum conditions.
of these
a
as
of the parameter
is therefore recommended
tolerant genotypes for environments with
be low under
a
tolerance.
date of all the genotypes included in
the identified soil
indexes
for the variable classification of genotypes after their DRI
postflowering drought
might
It
genotypic performance therefore depends
suggested by Finlay
to select
(1987b).
timing
in different environments
responsible
grains per
confirmed with the consistent
was
measured for each
year interaction for panicle
drought tolerant
as
imposed
years each genotype reacted in
cycle length
be
yield
ef al.
be
mechanisms
number of
a
can
related parameters measured under
mechanism and is therefore
An individual
genotype
physiological
high
large number of genotype specific drought response
ment that consists in the
over
years of
stress is related to DRI but not to
Although drought
design.
over
suggested by Bidinger
drought
drought escape
trials
a
above average individual
an
specific environment. This
stress to a
as
under
trial
and/or
of pearl millet
genotype measured under drought stress expresses
a
drought
agronomic
produce large grains
to
and
as a
tool
high probability
of
genotypes however
88
Drought escape
is
one
of the two mechanisms besides
drought avoidance describing
in this
emphasis
escape
as
study
a
screening option
a
selected after their
environments with
study showed
reaction to
specific
was on
drought
early flowering
date
frequent drought.
the Sahelian environment. When
a
tolerance and
Although
stress.
of
importance
the
drought
neglected. Short cycle genotypes
are a
valuable contribution to work in
However it has to be
reduced biomass for short
a
drought
tolerance, the
cannot be
drought
cycle
mid-season
kept
in mind that this
genotypes what
drought
can
be
risky
occurs, a short
in
cycle
genotype is less flexible to compensate yield loss with its tillers. Bidinger (1987a)
showed further
a
confirmed in this
study. This of course would
yield
in the Sahel where
season.
To
develop
to maintain a
in
reduced
a
yield potential for early lines which could
drought
screening
can
be
frequent
method for
high yield potential, exploiting
good years
and to stabilize
Although drought
stress is
not be in the sense to
drought
but is not
length
grain yield in drought prone
regarded
as an
improve grain
occurring every
tolerance therefore
the whole
of the
rainy
millet
in
water
production. Penning
grassland production
was more
to
Sahel,
limiting
supply
important production constraint
in the south. Bationo etal.
growing region
Niger.
in the
only,
(1989)
like
in the
they
are
constraint to
found that water is
whereas nutrient
limiting
efficiency
noted that fertilizer is needed
nutrient-poor sandy
However moisture
the response to fertilizers.
primary limiting
Dijiteye (1982)
in the northern Sahel
improve crop production
in
may not be the
de Vries and
means
season
seasons.
Sahel, Payne etal. (1990) reports that in the sandy, low input fields
usual in the
not be
availability
soils of the
has
pearl millet
significant effects
on
89
9. References
Annerose, D.J. 1988. Criteres physiologiques pour ('amelioration de
I'adaptation
a
la secheresse de I'arachide.
Oleagineux
43:217-222.
Appa Rao, S., Mengesha, M.H., and Reddy, C.R. 1988. New sources of
early-maturing germplasm in peart millet. Indian Journal of Agricultural
Science. 58:743-746.
Bationo, A., Christianson, C.B., and Mokwunye, U. 1989. Soil fertility
management of the pearl millet producing sandy soils of Sahelian West
Africa: The Niger experience. Pages 159-168 in Soil, Crop and Water
management systems for rainfed agriculture in the sudano-sahelian zone:
Proceedings of an International Workshop, ICRISAT, Patancheru, A.P.
502324, India.
Bationo, A., Baethgen, W.E., Christianson, C.B., and Mokwunye, A.U.
1991. Comparison of five soil testing methods to establish phosphorous
sufficiency levels in soil fertilized with water-soluble and sparingly soluble
P-sources. Fertilizer Research 28:271-279.
Bidinger, F.R., Musgrave, R.B., and Fischer,
pre-anthesis assimilate to grain yield in
stored
R.A. 1977. Contribution of
wheat and
barley.
Nature
270:431-433.
Bidinger, F.R., Mahalakshmi, V., and Rao, G.P.D. 1987a. Assessment
drought resistance in pearl millet. I Factors affecting yields under stress.
of
Australian Journal of
Agicultural
Research 38:37-48.
Bidinger, F.R., Mahalakshmi, V., and Rao, G.D.P. 1987b. Assessment
drought resistance in pearl millet. II Estimation of genotype response to
stress. Australian Journal of Agricultural Research 38:49-59.
of
Bieler, P., Fussell, L.K., and Bidinger, F.R. 1992. Grain growth of
Pennisetum glaucum (L.) R.Br, under well-watered and drought-stressed
conditions. Field Crops Research 30: (in press)
Bishnoi, O.P., Umamaheshwara, V., and
requirement for growth and development
zone
Singh. 1985. Heat unit
pearl millet. Annals of arid
Diwan
of
24:241-250.
Black, C.R., and Squire, G.R. 1979. Effects of atmospheric saturation
deficit
S. and
on
pearl millet (Pennisetum typhoides
groundnut {Arachis Hypogaea L). Journal of Experimental
the stomatal conductance of
H.)
and
Botany 30:935-945.
90
Blum, A. 1989. Osmotic adjustment and growth of barley genotypes under
drought
stress.
Crop Science 29:230-233.
Blum, A., Poiarkova, H., Mayer, J., and Golan, G. 1983. Chemical
plants as a simulator of postanthesis stress. I Effects
desiccation of wheat
of translocation and kernel
growth. Field Crops Research
6:51-58.
Blum, A., and Ritchie, J.T. 1984. Effect of soil surface water content
sorghum
Crops
root distribution in the soil. Field
on
Research 8:169-176.
Borrell, A.K., Incoll, L. D., Simpson, R.J., and Dalling, M.J. 1989.
Partitioning of dry matter and the deposition and use of stem reserves in
a semi dwarf wheat crop. Journal of Botany :527-539.
Boyer, J.S.
1976.
Photosynthesis
at low water
potentials. Phil. Trans.
R.
See. Lond. 273:501-512.
Bramel-cox, P.J., Andrews, D.J., Bidinger, F.R., and Frey, K.J. 1984.
rapid method of evaluating growth rate in pearl millet and its weedy and
A
wild relatives.
Bruckner,
Crop Science 24:1187-1191.
P.L.,
in
adaptation
and
spring
Frohberg,
wheat.
R.C.
1987a.
Crop Science
Stress
tolerance
and
27:31-36.
Bruckner, P.L., and Frohberg, R.C. 1987b. Rate and duration of grain fill
spring wheat. Crop Science 27:451-455.
in
Burton, G.W., and Powell, J.B. 1968. Pearl millet breeding and cytoge¬
netics. Adv.
Agron.
12:187-188.
Carberry, P.S., and Campbell, L.C. 1985. The growth and development
of pearl millet as affected by photoperiod. Field Crops Research 11:207217.
Chaudhuri, U.N., and Kanemasu, E.T. 1985. Growth and water
sorghum and pearl millet. Field Crops Research 10:113-124.
Chopart,
J.L. 1983. Etude du
sableux du
systeme racinaire du mil dans
Senegal. Agronomie Tropicale
un
of
sol
38:37-51.
1981.
Descriptive
sorghum
genotypes.
Duncan, R.R., Bockholt, A.J., and Miller, F.R.
comparison of senescent and
Agronomy Journal 73:849-853.
use
nonsenescent
Eberhart, S.A., and Russell, W.A. 1966. Stability parameters for
varieties. Crop Science 6:36-40.
comparing
91
Egharevba,
to
P.N. 1981. Contribution of assimilate from different leaf zones
developing millet grain. Samaru Journal of Agricultural Research
1:27-36.
Ferraris, R., and Charles-Edwards, D.A. 1986a. A comparative analysis
growth of sweet and forage sorghum crops. I Dry matter production,
phenology and morphology. Australian Journal of Agricultural Research
of the
37:495-512.
Ferraris, R., and Charles-Edwards, D.A. 1986b. A comparative analysis
growth of sweet and forage sorghum crops. II Accumulation of
soluble carbohydrates and nitrogen. Australian Journal of Agricultural
of the
Research 37:512-533.
Finlay, K.W., and Wilkinson, G.N. 1963. The analysis of adaptation in a
plant-breeding programme. Australian Journal of Agricultural Research
14:742-754.
Fischer, K.S., and Wilson, G.L. 1975a. Studies of grain production in
sorghum bicolor. III. The relative importance of assimilate supply, grain
growth capacity and transport system. Australian Journal of Agricultural
Research 26:11-23.
Fischer, K.S., and Wilson, G.L. 1975b. Studies of grain production in
sorghum
bicolor.
IV.
Some
effects
of
increasing
and
decreasing
photosynthesis at different stages of the plant's development on the
storage capacity of the inflorescence. Australian Journal of Agricultural
Research 26:25-30.
Fischer, R.A., and Maurer, R. 1978. Drought resistance in spring wheat
cultivars.
I
Grain
yield responses.
Australian
Journal of
Agricultural
Research 29:897-912.
Forest, F. 1982. Evolution de la pluviometrie
au cours
cultures
en zone
soudano-sahelienne
periode 1940-79 et consequences sur le bilan hydrique
pluviales au Senegal. Agronomie Tropicale 37:17-23.
de la
des
Fussell, L.K., and Pearson, C.J. 1978a. Course of grain development and
its
relationship to black region appearance in pennisetum americanum.
Crops Research 1:21-31.
Field
Fussell, L.K., and Pearson, C.J. 1978b. Effect of thermal history
photosynthate translocation
Plant Physiology 5:547-551.
and
photosynthesis.
on
Australian Journal of
92
Fussell, L.K., and Pearson, C.J. 1980. Effects of grain development and
history on grain maturation and seed vigour of Pennisetum
americanum. Journal of Experimental Botany 31:635-643.
thermal
Fussell, L.K., Pearson, C.J., and Norman, M.J.T. 1980. Effect of
temperature during various growth stages on grain development and yield
of
pennisetum americanum. Journal
Experimental Botany
of
31:621-633.
Fussell, L.K., Bidinger, F.R., and Bieler, P. 1991. Crop physiology and
breeding for drought tolerance: research and development. Field Crops
Research 27:183-199.
Garcia-Huidobro, J., Monteith J.L., and Squire, G.R. 1982. Time,
temperature and germination of pearl millet. I Constant temperature.
Journal of
Experimental Botany
Goldsworthy,
P.R. 1970. The
in tall and short
33:288-296.
sources
sorghum. Journal
of assimilate for grain
of
development
Agricultural Science Cambridge
74:523-531.
Grant, R.F., Jackson, B.S., Kiniry J.R., and Arkin, G.F. 1989. Water
deficit timing effects on yield components in maize. Agronomy Journal
81:61-65.
Guiragossian, V.Y., Van Scoyoc, S.W., and Axtell, J.D. 1979. Chemical
biological methods for grain and forage sorghum. Dept. of Agronomy,
Agric. Exp. Station, Purdue University, W.Lafayette, IN 47907,
and
USA:174-178.
Henson, I.E., Mahalakshmi, V., Bidinger, F.R., and Alagarswamy, G.
1981. Stomatal responses of peari millet {Pennisetum americanum (L.)
Leeke), in relation to abscisic acid and water stress. Journal of
Experimental Botany 32:1211-1221.
Hume, D.J., and Campbell, D.K. 1972. Accumulation and translocation
of soluble solids in
com
stalks. Canadian Journal of Plant Science
52:363-368.
Jacquinot,
L.
1970.
La
nutrition carbonee du
assimilats carbones durant la formation des
mil.
I
Migration
des
grains. Agronomie Tropicale
25:1088-1095.
Johnson, D.R., and Tanner, J.W. 1972. Calculation of the
duration of
grain filling
in
corn.
Crop Science
12:485-486.
rate and
93
Jones, D.B., Peterson, M.L., and Geng, S. 1979. Association between
grain filling and yield components in rice. Crop Science 19:641-643.
Jones, M.M., Turner, N.C., and Osmond, C.B. 1981. Mechanisms of
drought resistance. In: L.G. Paleg and D. Aspinall (ed.) The physiology
and biochemistry of drought resistance in plants. Academic Press,
Australia.
Jurgens, S.K., Johnson, R.R., and Boyer, J.S. 1978. Dry matter
production and translocation in maize subjected to drought during grain fill.
Agronomy Journal 70:678-682.
Kumar,
Outlook
K.A.
on
1989.
Pearl millet: current status and future
Agriculture
potential.
18:46-53.
Mahalakshmi, V., and Bidinger, F.R. 1985a. Flowering response of pearl
millet to water stress
Biology
during panicle development.
Annals of
Applied
106:571-578.
Mahalakshmi, V., and Bidinger, F.R. 1985b. Water
floral initiation in
pearl millet.
Journal of
Agricultural
stress and time of
Science
Cambridge
105:437-445.
Mahalakshmi, V., and Bidinger, F.R. 1986. Water deficit during panicle
development in pearl millet: Yield compensation by tillers. Journal of
Agricultural Science Cambridge 106:113-119.
Mahalakshmi, V., Bidinger, F.R., and Raju, D.S. 1987. Effect of timing
on pearl millet. Field Crops Research :327-339.
of water deficit
Mahalakshmi, V., Bidinger, F.R., and Rao, G.D.P. 1988. Timing and
intensity of water deficits during flowering and grain-filling in pearl millet.
Agronomy Journal 80:130-135.
Maiti, R.K., and Bidinger, F.R. 1981. Growth and development
peari millet plant. ICRISAT Research Bulletin No.6.
of the
McCaig, T.N., and Clarke, J.M. 1982. Seasonal changes in nonstructural
carbohydrate levels of wheat and oats grown in a semiarid environment.
Crop Science 22:963-970.
Misra, R.K., and Nagarajarao, Y. 1981. Water
varieties of
29:301-306.
pearl millet.
Journal of the Indian
use
under different
Society
of Soil Science
94
Muchow, R.C., and Wilson, G.L. 1976. Photosynthetic storage limitations
to
yield
in
sorghum bicolor. Australian Journal of Agricultural
Research
27:489-500.
Myers, R.J.K.
1980. The root system of
a
grain sorghum crop. Field Crops
Research 3:53-64.
O'Neill, M.K., Hofmann, W.C., and Dobrenz, A.K. 1987. Moisture stress
on the yield and water use of sorghum hybrids and their parents.
effects
Journal of
Agronomy
&
Crop Science
159:167-175.
Oertii, J.J. 1988. Effects of periods of drought at different stages of
growth of a spring wheat cultivar. In: Challenges in dryland agriculture a
global perspective. Proceedings of the International Conference on
Dryland Farming, August 15-19, 1988, Amarillo/Bushland, Texas, USA.
-
Ong, C.K.
Journal of
1983. Response to temperature
Experimental Botany 34:337-348.
in
a
stand of
pearl millet.
Ong, C.K., and Monteith, J.L. 1985. Response of pearl millet
temperature. Field Crops Research 11:141-160.
to
light
and
Ouattar, S., Jones, R.J., and Crookston, R.K. 1987a. Effect of water
during grain filling on the pattern of maize kernel growth and
development. Crop Science 27:726-730.
deficit
Ouattar, S., Jones, R.J., Crookston, R.K., and Kajeiou, M. 1987b. Effect
of
drought
on water
relations of
developing
maize kernels.
Crop Science
27:730-735.
Pande, P.C., Pokhriyal, S.C., and Balzor Singh. 1983. Heterosis in grain
activity in pearl millet. Indian Journal of Genetics 43:409-413.
sink
Payne, W.A., Wendt, C.W., and Lascano, R.J. 1990. Root zone water
balances of three low-input millet fields in Niger, West Africa. Agronomy
Journal 82:813-819.
Penning de Vries, F.W.T., and Dijiteye, M.A. 1982. L'elevage et
I'exploitation des paturages au Sahel. In: F.W.T. Penning de Vries and
M.A. Dijitrye (ed.) La productivity des paturages Saheliens. Une etude des
sols, des vegetations et de I'exploitation de cette ressource naturelle.
Centre for Agric. Publ. and Documentation, Wageningen, The Netherlands.
Powell, J.M.,
carbohydrate
83:933-937.
Hons,
F.M., and McBee, G.G.
partitioning
in
sorghum
stover.
1991.
Nutrient
Agronomy
and
Journal
95
Rachie, K.O., and Majmudar, J.V. 1980. Pearl Millet. Pennsylvania State
University Press, 307 pp.
Rao, D.V.M., Mahalakshmi, V., and AH, S.M. 1978. Studies on the
relative contribution of various photosynthetic plant parts to the grain
development at various moisture levels in Pennisetum typhoides. Mysore
Journal of
Agricultural Science
12:363-367.
Reed, A.J., and Singletary, G.W. 1989. Roles
phytohormones
of carbohydrate
in maize kernel abortion. Plant
Physiology
supply and
91:986-992.
Sagar, P., Kapoor, R.L., and Jatasra, D.S. 1984. Phenotypic stability
drought index in pearl millet. Annals of Arid Zone 23:207-211.
Sanders, J.H.
1989.
of
Agricultural Research and cereal technology
Niger. Agricultural Systems 30:139-154.
introduction in Burkina Faso and
Sayed, H.I., and Gadallah, A.M. 1983. Variation in dry matter and grain
filling characteristics in wheat cultivars. Field Crops Research 7:61-71.
Schmidhalter, U., and Oertii, J.J. 1987. Beitrage zur Kenntnis der
Mitteilungen Deutsche Bodenkundliche
Wasseraufnahme der Wurzeln.
Gesellschaft. 53:467-472.
Oertii, J.J. 1991. Osmotic
ecology and its practical
application a contribution to the investigation of the whole plant. 3rd Int.
Symposium, Vienna, self publishers, Verein fur Wurzelforschung, A-9020
Klagenfurt.
Schmidhalter,
adjustment of
U.,
roots
Evequoz,
M.,
and shoots.
and
In:
Root
-
Sivakumar, M.V.K. 1986. Soil climatic zonation for Weat African semi-arid
tropics
implications of millet improvement. Paper presented at the
Regional Millet Workshop, Sept. 1986, Niamey, Niger.
-
Sivakumar, M.V.K. 1989a. Agroclimatic aspects of rainfed agriculture in
sudano-sahelian zone. Pages 17-38 in Soil, Crop and water
the
management systems for rainfed agriculture in the sudano-sahelian
Proceedings of
an
International
zone:
Workshop, ICRISAT.
Sivakumar, M.V.K. 1989b. Climatic changes and crop production patterns
in the Sudano-Sahelian zone. A paper
presented
at the OAU/STRC
SAFGRAD/PANEARTH/ISRA Workshop on the effects of climatic changes
the agricultural and ecological systems in Sub-Saharan Africa, 11-15
September 1989, Saly, Senegal.[Limited Distribution]
on
96
Squire, G.R. 1979. The response of stomata of pearl millet {Pennisetum
typhoides S. and H.) to atmospheric humidity. Journal of Experimental
Botany 30:925-933.
Squire, G.R.
1989.
Response
to
Temperature in a stand of peari millet.
Experimental Botany 40:1391-1398.
10. Partition of assimilate. Journal of
Stamp, P., Rave, G., Diepenbrock, W., Herzog, H., Krippgans, O.,
and
Geisler, G. 1982. Comparative analysis of dry matter accumulation in
caryopses and seeds of aestivum wheats, durum wheat, rye, tritcale, rape
seed and pea by means of
Crop Science 151:224-234.
Ugherughe, P.O.
for the
arid and
Agriculture
1987.
a
growth function. Journal of Agronomy and
Improvement in the drought resistance of crops
International Journal of Tropical
semi-arid tropics.
1:28-40.
Van
Soest, P.J. 1987. Carbohydrates. Pages 95-117 in: Nutritional
ecology of the ruminant. Cornell University Press, New York.
West L.T.,
Wilding L.P.,
Landeck J.K., and
Calhoun, F.G. 1984. Soil
survey of the ICRISAT Sahelian Center, Niger. Soil and crop science
department/Tropsoils, Texas A&M University System College Station,
Texas, USA.
Westgate, M.E., and Boyer, J.S. 1985a. Carbohydrate reserves and
reproductive development at low leaf water potentials in maize. Crop
Science 25:762-769.
Westgate, M.E.,
and
Boyer, J.S.
1985b. Osmotic
inhibition of leaf, root, stem and silk
maize. Planta 164:540-549.
growth
adjustment and the
potentials in
at low water
97
Annex I. Climatic environment of the field
1990
(c),
and 1991
(d)
in Sadore,
February
experiments
in
1988
(a),
1989
West Afnca
Niger,
I
March
April
I
May
60
50
J Sowing
50
I
b)
Flowering
40
40
30
30
20
20
-
10
10
February'
IB
rainfall
March
~
evap
I
April
mintemp
I
May
maxtemp
(b),
98
Annex I.
(continued)
February
March
I
April
May
60
50
50
JSowing
d)
J Flowenng
40
40
30
30
20
20
10
^^\J^J^^
February
^H
rain
I
—-
^W^~
March
April
evaporation
mintemp
I
May
maxtemp
10
99
Flowering date (DAS=days after sowing) and drought response
(DRI) of all the genotypes (by alphabetic order) tested in the postflo¬
wering drought screens in 1988 to 1990. Trials were all planted in the hot
off-season (February to may).
Annex II.
index
DRI
Flowering (DAS)
Genotype
1988
1989
1990
1988
3/4 HK-B78
63
0
Ankoutess
66
-0.24
Boudama
58
0.32
CIVT
66
66
67
0
1989
1990
0
1.24
CMM 466
68
-0.67
HKB-P1
65
1.75
HKB-Tift
68
0
HKP
(Mali)
65
1.57
HKP
(Maradi)
65
58
66
-0.29
0
0
ICMLIS 82205
69
0
ICMLIS 82215
67
-1.35
ICMLIS 82270
67
0.47
ICMLIS 82298
70
-0.98
ICMLIS 82299
70
0
ICMLIS 84213
68
0
ICMLIS 84221
70
-0.41
ICMLIS 84227
68
5.37
ICMLIS 84232
66
-0.54
ICMLIS 84248
69
0.34
ICMLIS 84256
66
0
ICMLIS 84262
65
0
ICMLIS 84284
66
-0.25
ICMLIS 84296
67
0
ICMLIS 85211
69
-0.3
ICMLIS 85212
66
0
ICMVIS 85321
63
65
67
0.69
2.49
0
o>
0.
o
Ol
o>
60
63
Ol
61
64
-•
ro
SJ
61
oo
57
o
O
W
en
*.
\l
O)
Ol
o
oo
O
*
o
O
vl
CO
8
p
CIS
flo
ro
CO
O
o
CD
to
•si
sj
Ol
o>
59
Co
s|
*
853-820 842-817
AWN NMV AWN AWN AWN NMV NMV AWN NMV NMG NMG NMG NMG NMG NK2 NK2 NK2
8428 8421 8417 8271 8269 8258 8246 8210 8201
o>
O
CO
o>
CO
ro
00
CO
CO
8201834-8CO
-09
00
CD
<
<
Ol
-si
CJ>
o>
O
CO
O)
a>
o>
o>
o>
CO
CO
o
m
co
en
CO
ro
CO
CO
cr>
00
00
00
ro
00
<
CO
<
CO
CO
<
-si
CO
o>
00
o>
00
01
U
CO
CO
<
Ol
*.
O)
CO
CO
CO
Ol
00
CO
<
Ol
^
CO
CO
00
Ol
CO
<
s4
CO
CO
_i
CJ>
CO
CO
CO
<
ro
CD
o>
ro
-si
o>
oi
00
CO
<
Ol
en
o>
CO
to
00
Ol
CO
<
OOOOOOOOOOO
KMV
CO
CO
o
-
CO
CO
CO
_i
00
00
o
TJ
s-»
>
CO
CO
3
CO
S
o
TJ
a-.
§
CO
O
CD
o
o
101
Genotype
1988
1989
INMV 8429
58
INMV 8442
65
63
INMV 8456
60
64
INMV 8459
DRI
Flowering (DAS)
(continued)
1990
1989
1990
0
0
0.62
0
0
-0.3
58
ISC 851
1988
1.02
63
ITMV 8304
62
0
M2D2
60
0
NKK
63
0
P3Kolo
63
POOL-6
64
64
67
0
-0.05
0
-0.31
RS89 104128
81
0
RS 89 209103
73
-0.09
RS 89 221130
75
0
SOxSA
66
SOxTO
60
0
Sosap
-1.49
1.43
68
Soumno
72
Souna III
68
Synth-1
60
Synth-11
59
Synth-2
Synth-5
0.23
65
61
Synth-6
0.47
-0.81
59
63
62
64
1.38
61
Trial Mean
62
0
0
-0.33
0
64
0
-0.5
65
-2.81
-1.88
63
Synth-8
-2.51
68
102
Annex III. Root
(b) of
density
per volume soil
by length (a)
and
dry weight
genotypes. Sadore 1990.
two
cm/cm3 soil
5
30
10
50
70
90
110
130
150
170
190
210 230 250 270
Depth (cm)
g/cm3 soil
5
10
30
50
70
90
110
130
150
170
190
Depth (cm)
I ICMVIS 86327
I
Synth-2
210
230 250 270
103
Annex IV. Root
(D1), 19
cm
dry weight density at interrow distances of 0 cm
(D2) and 38 cm (D3) for genotype HKP (a) and
ICMVIS 85327
(b).
Sadore 1991.
g/cm3 soil
40
80
60
Depth (cm)
•D 1
+
D 2
"*--
D 3
Acknowledgement
My
sincere thanks to Prof.
Technology (ETH)
Stamp of the Swiss Federal Institute of
P.
Dr.
Zurich who gave
me
the
opportunity
to do my work in an
international research center in West Africa and also gave
support and
understanding.
Important
examiner.
valuable while
his maximum of
me
Thanks also to Prof. Dr. J.J. Oertii to accept the
discussions at the Institut of Plant Science at ETH
closing
co-
were
up.
I'd like to express my sincere thanks to Dr. R.W. Gibbons, Executive Director,
and Dr. K. Anand Kumar, Peari Millet
Sahelian Centre
(ISC)
in
Niamey, Niger to accept
Dr. L.K, Fussel and Dr. F.R.
Bidinger
This thesis would not have been
Adeyemi,
generating
and
Operations Program,
labourers. I'm
greatful
my time in
for
Niger
thanks for
the
valuable
without the
the data.
by Agra
help
of
qualified support
My special thanks
Moustapha Amadou,
Transport Unit of ISC
at ISC
to
Jimmy
Ene Nweeze,
and the
cheering
numerous
me
up and
experience.
Cooperation
and Humanitarian Aid
supporting this thesis financially.
Fond
PhD research scholar,
of my work.
everybody's contribution
a
To the Swiss Directorate of
financed
possible
analyzing
me as a
of ICRISAT
well as Dr. J.H. Williams and Dr. J.
Boubacar Mara, Monkeila Boriema,
the Farm
making
as
during difficult phases
Brouwer for their support
staff at ISC in
Improvement Program Leader,
(Switzerland).
The final
(SDC) my sincere
writing up
was
generously
Curriculum Vitae
Name:
Peter BIELER
Date of Birth:
7 March 1963, born to Kurt Bieler and Martha nee
Gartner in Zurich, Switzerland.
Place of
Origin:
Solothum
Nationality:
Switzerland
Civil Status:
Single
Education:
1970-1982
Five years of
primary school, four years of secondary
school and three and
half years of grammar school,
a
leaving with the scientific diploma (Matura C).
1983-1988
Ten semesters at the
(Swiss
Federal
Diploma
in
Faculty
Institute
Plant
"Characterization
of
of
of
Agriculture
at ETH
Technology) Zurich.
Science.
Diploma
thesis:
forms
double-haploid
of
summerwheat landraces in Switzerland".
Experience:
1989-1992
Associated
Expert of Swiss Development Corporation
at ICRISAT
(International Crop Research Institute for
the Semi Arid
Tropics)
Niger (West Africa)
Sahelian Centre in
in the Pearl Millet
Program, Physiology Department
Niamey,
Improvement
for PhD-Thesis.