HAL Id: ird-00142188
https://hal.ird.fr/ird-00142188
Submitted on 17 Apr 2007
HAL is a multi-disciplinary open access
archive for the deposit and dissemination of sci-
entic research documents, whether they are pub-
lished or not. The documents may come from
teaching and research institutions in France or
abroad, or from public or private research centers.
L’archive ouverte pluridisciplinaire HAL, est
destinée au dépôt et à la diusion de documents
scientiques de niveau recherche, publiés ou non,
émanant des établissements d’enseignement et de
recherche français ou étrangers, des laboratoires
publics ou privés.
Phenology and reproductive eort of cultivated and wild
forms of Pennisetum glaucum under experimental
conditions in the Sahel : implications for the
maintenance of polymorphism in the species
Jean-François Renno, Thierry Winkel
To cite this version:
Jean-François Renno, Thierry Winkel. Phenology and reproductive eort of cultivated and wild forms
of Pennisetum glaucum under experimental conditions in the Sahel : implications for the maintenance
of polymorphism in the species. Canadian Journal of Botany, National Research Council Canada,
1996, 74 (6), pp.959-964. �10.1139/cjb-74-6-959�. �ird-00142188�
.3
959
Phepslogy and reproductive effort of cultivated
and wild forms of P~NI~~&J~ g/arua=um wnder
experimental conditions in the Sahel: impllieatisns
for the maintenance of polymorphism in the species
J.-F. Renno and T. Winkel
Abstract:
In the Snhel region of Africa. the wild and the cultivated forms of pearl millet, Pe~~~~i.sc~frr~~r g/tr~rc~r~r (L.)
R.Br., are syrnpntric and interfertile and yet have rcmaincd distinct for millenia. Rcproductivc harriers are not sufficient
to explain this situafion. To elucidate other possible mchanisms, the two forms were coniparcd under cxperiniental
conditions in the S:~hcl for their phrnology and rcproductivc effort. The length of the flowering period of each type ~‘3s
much longer than the average individual flowering period. When the last cultivated plants were finishing knvering,
65% of the wild plants were still tlowcring and 30% were just starting to tlowcr. Thus,
tlic last group wils conipletrly
isolated from ctIltiv;lted pearl millet gene now (endog;mic reproduction). The two forms of pearl millet also differed in
tht: distribution of aboveground biomass among different plant pnrts,
except for the number uf seeds per plant. Both
phrnological behaviour and ruproductivr: effort contribute to the maintenance: of distinct forms of wild ;rlld cultivated
pCild
millet.
Introduction
Archi~eologicnl traces of inqxints of both wild and cultivntcd
pcnrl niillct (Pcrrrri.rc~urn gln~rr~r) seeds rind hrisrlcs on pot-
tery, found in southcast Mauritania and dating from around
3000 BP,
arc
evidence of a neolithic agriculture with culti-
valcd and wilcl pcnrl n1illct populations in synipatry (Amblarcl
and
Pcrni3 1989).
The current wild fomm is distributed across
the northern Sahel
and
is phylogenctically close to the anccstol
of the cullivakd pearl niillct (Harlan 1975: Portkrcs 1976).
Van der ZOII (1992) recognized three subspecies in Pcrl-
rrisc~~cm ,yhcm (L.) R.Br. : P. gkzucrtnr ssp.
gk7wrrur. the
cultivated pearl inillct; P. glfrrrc~otl ssp. I.I’~cILY~IIII~ (Lam.)
A. Rich., the \viltl and weedy forms: and P. gk~rcrrrn ssp.
sidcrkrrltr~ (Schlccht.) Slapf C! Hubh., lhc inlcrnlcdiatc
forni bctwccn the wild nncl the
CulliViltNl IlXnls, known 35
shibra in Niger. These three I;CGI hclong to rhc satne gene
pool of the allr)gamic
and polyinorphic annual species.
P. g/(~rr~~ttr (L.) R.Br. (f-iarlan rind de Wet 197 I).
Rcgnrding the cultivated pearl niilicr. die farnicr plays an
csscntial sclcctivc role through agriculrurnl practices: choice
of scccIs from spikes typical fur n local cultivar, batches of
seeds (cnllcd pockets) sown into holes Ihvoring the growth 01
the biggest seeds, hoeing to elin~inole
~ccrls
and pearl millet
plants outside the pockets, thinning out to preserve the most
vigourous plants. and elin~ination of the shibra. Ncverthe-
Icss, in the field. the frequency ol’thc rcnxiining shibra varies
bctwccn 5 and 30% in the :lI’CilS whcrc the wild and cultisxd
fornis
arc
in contact (Rcy-Hcrnx 1983). These wild popu-
lations arc: not hubjcct LU conscious scloctiun prcssurc by
huinaiis. f’or lwo saniplcs of sccrls tukcn from wiid pa-1
millet populations in Scncgul and Niger, the proportion 01
dcsccndants of the hybrid phenolypc (shihrn).
supposcdl~ the:
producl
of fertilization by CllltiVil~Cd
pdlcrl, \Sxi cstimared at
3 I iId I9%, rcbpcclivcly (MnKhiliS anil Tosrain 19911
Fonds Docurngmtaire ORSTOM
9GO
3
I-low
is
the polymorphism
of
I'.
glwcrrtrr
maintained'?
How
is
it
t1i;tt
wild
pearl
niillct
docs not rlis;ippcar
in
contact
with cultiv;itcd
pearl
niillct
under
the
cI'l'clcts
of
gene
flow?
Many
studies
have
addrcsscd
thcsc
questions. Studies
of
gcnclic
dist;ince indicated strong intcrniixing of genes
hctwccn
wild and cultivatcd
pearl
millet
(Tostain
1993).
Pcrnk
(1983,
1985)
underlines
the
fxt
that
the
two forms
interact
but
cai cxchnngc alleles without
a
modification
of
phtilotypcs
because
the
genetic control of clonicstication is
supposed
to
dcpcncl
on
a
few
linked
genes.
Moreover, gene
flow ciin be controllecl by mechanisms
that
act
at
the pre- and
post-zygote
stages
in
I'iivour
of
the niaintennnce
of
differ-
entiation between
the
wild and cultivutecl fornis. Pollen
competition was observed as a prezygotc
barrier,
giving
the
advantage
to
sel(-pollination (Sarr
et
al.
1988; Robert
et
al.
1991),
while
the
reduction
in
viability
of
hybrid grains was
olmrvctl
;is
n
postzygotc brake (Anioukou and Marchais
1993). Genomic
stratcgies
in
wild
pearl
millet
in
contact
with
cultivated pearl
niillet
were mentionccl by Joly-Ichenhauser
(1984)
and Prrnhs
(I
986). who assiiiiic the existence of
two
linkage groups of predomrstication genes that would have
the effect of favouring the maintenance
of
plicnotypes
in
spite
of
hybridization. However,
if
these
reprocluctive barriers
liniit
gene
flow, they
do
not
entirely cliniinutc
it
and are not
sufficient
hy
theiiisclvcs
to
explain
the niainteniince
of
two
distinct
interfertile
forms
in
contact for
riiillcnin.
Isolated
rcprodiiclioii
in
sp;icc
aiid
iilso
in
liille.
LIS
wcll as selective
~~rcsstrres,
cc~rltl
phy
iiii
cssenti:il
role
ill
scriicluring
[tic
poly-
IllllrpllisI11
ol'
pciir1
Illillcl.
WI1cI.L'
wiltl
ill111
cullivirlcll
fc)rllls
01'
p!itl~l
Illillet
yc
ill
~togi~iil)ltic;\l
CtilitiIcI.
wl~itt
is
[IIC
l)I~O1)(iIIiOiI
of
pli\t1ts
ill
cli~(iliic
Ilt)tvcrilig
I)e~iotls?
Wliitt
itre
IIIC
tlil'l'c~t~lccs
in
rc171'0-
c;1c41
limi cst.qirtg
ilic
geite llow ol'lttc
(iflier
liwrii
by
asyn-
tluct
ive
c'1'1i)rt
I)ctwecli wild
niid
otiltivalcd
pc:irI
inillet?
I
Iotv
:I~C
IIIC
cIiI'Ii.l.cllccs
~tl~ti~ltiti~i~d
~~[IVCCII
syliil);itriC
Wild
~11111
c~ll~iv~ttcd ~lo~~~ll;ltiotls
t1tilt
1l;Ivc
I)CCll
cxcll~ulging
genes
lo
r
I
Il
i
I
IC
II
ia?
Ill
;I11
;tllcrtl~ll
to
:ldrlrcss
IlluSc
c~ucstiolls,
wc'
compared,
tilitlcr
SiillcIiaIl
cxl)criIticiititI coiitlitiolis.
tl~c
l~hcnology
of
tillcring
ond
Ilowcrillg
of
¿I
wild
pc:"'I
Illillcl
:Illcl
a
cultivatecl
pearl
riiillct
fixmi
tltc
wiic
gcogrnpliicul region,
;IS
wcll
as
the
distributiwi
nt'
abovcground bionioss proiluction bctwccn
vcgct:itivc
and
rcprocluctivc organs
in
citeli
oftlicse
two forms.
Materials and methods
1
lilt
c'
ria
I
The
seeds
uscil
liir
tlic
stiidy
saiiiplcs
wcrc
collcctcd
from
hundreds
of
spikes
Ihr
each
rnriii,
in
six
sites
hctwccri
Uclhcji
(14"43'N.
X'OS'E)
aiitl
Taiiout
(I4"S7'N. 8"49'E)
iii
Nigcr.
whcrc
wild and
culiivntctf
pearl niillct
(cv. Ankoutcss)
;ire
syiiipatric.
'Ankoutess'
li
adapted
to
a
short
rainy
season
(3
inontlis
~'IUIU
June
to
Septem-
kr)
aiid
thus
cilri
be
considercd
;IS
miniiii;illy
or
riot
at
all
photo-
F+rioJ
sensitive
(Cleiiiciit
1985:
Burtori
niid
I'owcll
1968
in
Skcrnian
and
Riveros
IY9Uj.
Thc
sccds
01'
tvild
pl;lr)t\
were
Iiarvcstcd near
pari
Illillet
t'iclds.
Can.
J.
Bof.
Vol.
74,
1996
hctwccn
FeImiiIry
arid
bliiy
.
The tliiily
iiiciin
tciiipcr;ittire
varicd
lict\vccn
26
and
37"C,
;incl
the
diiily
niciin
incident
riidiiitioii
bctwccn
17
and
27
MJl(m'
.
clay).
Irrigiitioii
\viis
rcgul;itctl
to
simulate
tlic
;ivcr;ige
rainfall
rcginic
during
rlic
rainy
sciison
in
the
region
of
Tanout.
It
was
started
on
Fc1iru;iry
9.
one
d:iy
tierore
sowing,
then
stopped 74 days
after
emergence
(DAE,
date
at
which
SOO/,
of
the
pockets
had
emerged). There
was
one
niituriil
r;iinfiill
at
71
DAE
(7.5
mm).
Weeds
were
controlled
manually. l~crtili~ers
wcrc
supplied
at
30
kg
.
ha-'
N.
P,
;ind
K
1
thy
More
soWiiig
and
IS
kg.
ha-'
N
at
16
DAE.
The
expcrinicnt;il
ficlil
wiis divided
into
five repliciitions
with
two
subplots
each.
The
suhplots
measured
4
x
10
ni,
sown
either
with
cultivated
or
wild
sccds.
;ind
were
randomized
per
replication.
Seeds
were
sown
in
pockets
(1
pockct/m2)
and
the
development
of
eight
pockets
in
the
center
of
c;icIi
subplot
was observed. The
first
DAE wiis the
sanie
for
wild
pearl
niillet
Lind
cultiviited pearl
millet,
3
days
after
sowing.
At
15
DAE,
the
cultivated
plants
were
thinned
out
to
three
planta
per
pocket,
ii
planting
density considercd
to
be
optimal
under
Salieliun
conilitions (Institut
National
de Recherches
Agronomiques du
Niger
1987). In riich
pocket
one
randomly
chosen
plant
wiis observed
until
the
ciid
of
its
cycle.
For
tlie wild
plants,
no standard
density
can
be
represcnt;itivc
ofthc
lack
of
homogeneity
in
density
of
natural
poptil;itions.
For
the
convenience
of
observii-
tions,
taking
into
:recount
the
very
iniportiint
tillcring
of
each
plant,
the wild
plnnts
were thinned
IO
1
pl;tnt/m2.
This
experimental
iippronch
resulted
from
ii
comproiiiisc
between
111c
nerd
to
collect
precise
inforniiltioli
on
e;icli
inilivitlu;il
plant
iinil
the
iiced
to
observe
tlie
phenotypic expressitrii
of
e;ich
pl;int
form
as
it
is found
untler
niituriil
conditions.
l'lier
rolo,q>*
OJ
/i11criirg
ut
rtl
.//(I
i
I
*c*t.ir
rg
Every
2
or
3
d;iys during
the
rlcvclopriiciit
of
tlic
plants,
the
fol-
lowing
obscrv;itioris wcrc
iiwlc.
(i)
The
iiiiiiibcr
of
plants
in
the
primary
tillcring
phase:
ii
plant
WIS
ccwsitlcrcrl
to
I,c
in
the
primary
tillcring
phase
during
the
wliolc
period
hctwccri
the
tlcvcloprncnt
of
the
first
aiid
the
last
priiiiary
tiller
(the
priiiiary
tillers
arc
those
growing
out
of
the
first
stcni
th;it
lias
ciiicrgcd).
Tlic
young
pearl
riiillct
plorits.
particularly
the
wild
plants.
arc
too
fragile
to
be
frc-
clucntly
ninniptilotcd.
so
the
ohscrv;itioris
bcgnn
at
I7
DAE (2 days
:ifter
thinning)
for
the
ciiltiviitcd
torin
xrid
;it
30
DAE
(15
days
artcr
thinning)
for
thc
wild
form.
-1'11~
obscrviitions
were
niade until
harvest.
(ii)
lhe
nuiiihcr
ol'
plorits
in
tlic
Ilowcring
phase:
a
phnt
wiis
corisidcrcd
to
hc
in
tlic
Ilowcriiig
pli;isc
during
the
\\.hole
period
bctwccn
tlic
lint
and
tlic
Iiist
Ilowcr.
The
llowcring
phase
for
the
cultivatctl
li)riii
hegins
with
tlic
Il~i\vcriiig
of
the
spike
of
the
niain
stcni
(the
lïrsi
stciii
ciiicrgcd)
aiid
crids
with
;I
spike
of
a
priniary
tiller.
Siricc
thc
iiumbcr
of
priniary
tillers
\viis
low
and secondary
tillers
were
wry
rare
aiitl
;ilways
stcrilc,
the
Ilmvcring
phase
could
be
f~llowctl c.xliaiIstivcly
for
each
pliiiit.
With
the
wild
forni,
the
Ilowcririg
phax
nlso
begiris
with
tlic
Ilowcririg
of'
the
spike
of
the
but
cnds
with
ttic
Ilowcriiig
01'
tlic
y)ikcs
ol'
the
I~SI
additional
tiller
lïrst
btClll
(Ille
cr~"ivalcllr
ul'
the
II1;lill
StClll
01
Ille
culli\atcd
plant^
Renno
and
Winkel
¡Jig.
1.
Phenology of tillcring
(ti)
;ind flowering
(h)
in
cultiviited
(O)
and
wild
Iffiirl
millet
(O).
1
O0
u)
90
-
(o
80
0
70
-
U
60
Q-
50
40
c
30
u
C
Q
C
.-
L
-
.-
4
O
(o
a,
a,
Q
c
i2
20
10
O
4
15
25 35 45 55
65
75 85
Days
after emergence
ìü
a
80
-
o,
'C
a,
O
O
o,
C
a,
a,
C
70
3
GO
50
a,
40
2
30
IC
2
20
Q
10
O
35
45
55
65 75
85
Days
after emergence
(derived
from
the
rith
rank
of
a
primary
tiller).
For
each wild
plant,
the
flowering
period
of
the
first
stcni.
of
the
first
priniary
tiller.
and
of
the
last
tiller
were
noted carcl'ully. Thus, the whole
of
the
flowering period of each plant was covered.
Aboivgroi~ttcl biottrciss prodirciioti
rit
Iiciriwt
At
harvest, thc
plants
were
scpiiratctl
out
into
stenis,
leaves,
arid
spikcs
iind
dried
;II
gooc,
until
ii
constant weight was reachcd
(72
II).
For
each
plant,
the
nie;isiircniciits
ol'
(tic bioniass and the
following counts
were
carried
out:
(i)
vcgctiitivc organs:
mass
of
96
1
'priiniiry
tillers
(nhTi):
(i;)
rcprotluctivc orgiiiis:
numhcr
of
spikes
(nhSp),
nius
of
spikes
bcforc
tlircshiiig
(niSp),
iintl
;ifter
threshing.
ni;iss
of
seeds
(niSd)
and
nuiiihcr of
seeds
(nhStI).
The
reproductive effort wiis
cstirniitcd
by
(i)
the
reproductive
riitio
(RR),
which
is
the
niiiss
of
the
rcproductivc organs (spikes
including
seeds)
divitlctl by
tlic
total
hioni;iss
and
(i;)
tlic
secd
riitio
(SR).
which
is
the
miss
of
seeds
divided
by
the lotiil
bicm
An
cstiniution
WIS
niailc
of
the potential nuniber of
scctls
per
pliint
(pStllPI),
the
potential
number being
the
niiixinitiiii
nunihcr
of
seeds
produced
ifdl
the ovules (one
per
spikelet
in
pearl
millet) hat1
produced
seeds.
A
sample of spikes
(40
and
63
spikes for
the
culti-
vated and wild forms, respectively) was taken
at
rmlom from the
total nuniber
of
spikes
produced
by each forni of
pearl
millet. The
number of involucres on
the
central portion
of
each spike
was
counted for
a
length of
1.5
to
3
cm depending on the
size
of
the
spike. On each spike, the nuniber
of
spikelets per involucre was
counted
(I
for wild,
1-3
for cultivated pearl
millet).
After
nieiisur-
ing
the
total
length
of
the spikes, the mean potential nuniber
of
seeds
per spike
was
calculated and
niultiplied
by
the
number of spikes
per
plant
to
obtain
the
potentiiil number
of
seeds
per
plant
(pSd/PI).
The differences observed between cultivated and wild forms
were tested by
a
I
test
iis
a
5%#
probability level.
I'llcnology
or
~lowcrillg
'llic
nvcriigc llowcritig
period
ol'
:I
ctiltiv;itcd
pliiiit
w;is
strong
inclividual
vari;ltiolls
l'or
lXllI1
(CV
ol'
39%
l'or
lllC
wild
7.5
~~IYS,
wll~~~iis
tlicit
of
;I
\vild
IJI~IIII
\viis
21
tiiiys
\vitIl
s:liliplc
:IIKI
70%
lor
tl~
cultiv;Itcd
s;iIilplc)
(Tíiblc
I).
HOW-
cvcr, the
total
Ilowcring tinic
ol'thc
siirill>lcs
WU
45
d:iys
l'or
cultivated
pc¿irl
millet
and
iiiorc
thcin
49 days
for
wild
pcarl
niillct, sincc the
latter
population had
not
finishcd Ilowcring
at
harvest (Fig.
IO).
The distribution over
tirrie
of
thc pcrcentagc
of'
cultivatcd
plants
in
thc flowering
phasc
followcd
a
noriilal
curvc. Flow-
cring
bcgan
at
42
DAE,
whcri thc tillcring pliasc
was
finished.
Thc iiiaxiniiiiii level
\vas
attained at
63
DAE. with
50%
of
the
plants flowering. At 77 DAE, about
IO%
of
the
plants
wcrc continuing
to
flowcr. Thcrc werc
no
morc llowcrs
aftcr
84
DAE.
Thc siinic typc
of
distribution
was
obscrvcd
for
the
wild
plunts.
Thcy
began
to
flowcr
at
42
DAE,
well bcl'orc the cnd
ol'
their
tillcring phasc. The tii;1xinium numbcr
of
plants
in
tlic flowcring
phase
(85%)
w;is
at
74 DAE.
when
about
30%
of
thc cultivated plnnts
were
llowcring.
When
the
last
culti-
vatcd
plants
were I'inishing flowering
at
84
IIAE,
65
%)
of
the
thcmsclvcs; thus reproduction
was
cndoganiic.
Wlicn
irriga-
sturting
to
Ilowcr
mcl
iiiorc
than
80%
01'
their
Ilowcriiig
wild
pl~lllrs
saniplc
wcrc
going
to
exchangc
gcncs
only
anlong
tion
wits
sto^@
(74 DAE). 30%
ut'
thc
wild
pliilits
*
~
962
Can.
J.
Rot.
Vol.
74,
1996
Tnblc-
I.
Distribution
of
aboveground
Iioriiiiss
production,
secd
potcnri:il,
rcproductivc
riitio.
seed
riitio,
and
llowering duration
of
cultivated
and
wild
pcnrl
millct
plants
(nieans
pcr
indivitlual).
mSt
mL
niSp
niSd
TAm
nbSp
nbSd
nbTi
pSdlPI
RR
SR
Flowering
Cult
ivnted
Mcan
63.4 43.7 38.2 22.6
146.9 1.6
3194 6.8 8365 0.25 0.14 7.5
N
17
17
17 17 17
17
17 17
17
17
17 32
CV
46
46
64
71 42
71
68 I9 71 52 60 70
SE
7.0
4.9 5.9 3.9
14.8 0.3
525 0.3
1432 0.03 0.02
0.9
Wild
Mean
252.5 94.0
79.9
4.9
426.3
89.9
4158 16.2 32312
0.18
0.01
20.9
N
19
19
19 19
19
19
19
19
19 19 19
40
CV
52 41
67 108 44
75
108
21
75 54 98 39
SE
29.9
8.8
12.3 1.2
42.6 15.4
1026
0.8 5528
0.02
0.003
I
.3
Note:
Abbreviations
are
as
follows:
nisi, niiiss
of
stems;
mL.
nias
of
Ienves;
mSp,
iiiass
of
spikes;
niSd,
niass
of
seeds;
TAni.
total
biomass;
nhSp,
numher
of
spikes;
nhSti,
niimher
of
seeds;
nhTi, nuinber
of
primary tillers;
pSdlPI.
porential
seed
niimhrr
per
plilni;
RR.
reproductive riitio; SR,
seed
riitio;
flowering,
niem
diimtion
per
pliint
(tlaya);
N,
sample
size;
CV,
coefficient of variation
(%);
SE,
standard error.
Misses
are
in
grimis
of
dry matter per
plant.
Fig.
2.
Percentage differences
in
aboveground
hioni;iss
production, seed potential, reproductive
ratio,
srcd
riitio,
and
llowering duration of
wild
pciirl
millet rclíitive
to
cultivated
pciirl
millet.
Set
Titble
I
for
subheadings.
I
5563
208
2lld
U
I
period
took
plocc bctwcctl
this
dntc
itlid
tllc I1arvcsta
At
har-
vest
(90
DAE),
58%
of
thc wild plants wcrc
still
producing
hvcrs. The flowering period
of
thc
cultivutcd pcnrl niillct
thus
took
placc cntircly within
that
of
thc wild
pearl
millet.
Rcprocinctive
ct'fort
Exccpt
for
tlic numbcr
of
sccds per
plant,
all
the rcsults
rclatcrl
to
rcproductivc
effort
wcrc significantly diffcrcnt
(P
<
0.05)
bctwccn
the
two forms
of
pearl
niillct (Table
1;
Fig.
2).
Comp:ircd
with
a
cultivatcd
plant,
a wild plant had
thc following charactcristics:
(i)
the nunibcr
of
spikes was
57
tinics grcatcr, but
it
rcprcsentcd a total niass only twice
;IS
much as thc spikes
of
a
cultivated plant because
of
the
largcr six
of
the
spikes
of
cultivatcd
plants;
(i¡)
thc nunibcr
of
seeds
per plant was not significantly diffcrcnt from that
of
cultivatcd pcarl millct
(4
158
as
oppscd
to
3
194)
but
cor-
rcspt~ndctl
to
ti
niass
4.5
tinics
less
(4.9
g
íis
opposcd
to
27.6
g
for
the
cultivated
pciirl
millct)
bcciiUsc
of
the larger
siic
of
thc scctls
of
thc cultiviitcd plants;
(iii)
thc
investment
in
the
nunibcr
of
seeds
produccd
conipnrcd
with the total
hiotiiass
W;IS
half
(IO
scccls/g
of
dv
tiinttcr
;IS
oppscd to
32
sccdslg
of
dry
ni;ittcr
for
;i
cultivavd
plant).
his
being duc
principally
to
the difference
in
the average number
of
seeds
per
spike
(46
for
a
wild plant,
1996
for
a
cultivated plant).
The average reproductive ratios
(RR)
of
wild
and
cul-
tivnted plants were significantly different
(IS
ancl
2596,
respectively)
but
wcrc
still
quite close conip:ireti with
the
scetl
riitio
(SR),
which was
14
timcs
less
for
íi
wild plant than
for
ii
cttlI¡v;itcll
I>l;l[iI
(1
ti11t1
14%,
respectively),
A
wild
011
LtVeriIgc S-O
S~>I~CS.
A
cliltivi~tctl
pl;lt1I
li;id
[)II
ílvetïigc
1)Iiint
Iii\d
011
iivcri!gc
16.2
Ilritl1i\ry
tillers,
c;irh
Ijrotl\tcitig
0.8
priiwry
tillers
íintl
protlucctl
0.3
spikes
Ibr
ciich
of
tlicm.
'rllcrcrorc,
il
hrgc
proportion
of
its
tillers
did
not ~"'c"l1Icc
any
spikes.
Wie
cttltiviitd
IXXI~I
tilill~t
itttíiitlcd
40%
of
its cstitlli\tctI
potctltiitl
it1
IC~IIIS
of
secd
1)rotItlctiotI
(reiil
autiibcr
of
SCCC~S
Altholtgll tlic
tvild
1>ciirl
tllillct prodttcctl
i\
qit;intity
of
only
13%
of
its potc~itiill, which
W:IS
3,s
titiics
highcl
thiiti
per
plont
(nbSd)
tlivitlctl by
tlic
polcntial number (pSd/l'l))Q
cquivolcnt
to
ltial
of
cultivatcd
pcarl
tiiillct.
il
hacl
uttaincd
that
of
tlic cultivatcd
pearl
millct.
Discussion
This study shows that undcr cxperinicnlal conditions, thc
cultivatcd pcarl millct was always influenced by
the
wild
pearl millet pollcnic cloud, whereas the wild pearl millct
population escapcd
from the
cultivatcd pollenic cloud
for
a
large part
of
its flowcring
period.
If
thcrc had not becn
a
harvest, the Icngth
of
the endogamous reproduction
of
the
sample
of
wild plants would have depcndcd
on
the availability
of
water
in
the soil. According to Marchais
(1994),
undcr
natural conditions
in
Niger,
the observation
of
the flowering
of
a samplc of
tillers
in
a
population
of
wild pearl
millet
rcvcalcd that
40%
of
the
spikes
were still flowering when thc
plants
of
the cultivated and
shibra
phcnotypcs had finished,
The
latter
obscrvation docs
not
provide
any information
on
tlic
fraction
of
individual plants
that
participate
in
the endog-
atiioiis system but supports the
idea
of strong endogamy
in
ii
wild
pearl
millct population.
For
cultivated
iis
for
wild
pearl
millet,
tlic average tlower-
ing pcriod
of
a
single
plant was
much
shortcr
than
the
total