Plant and Soil 189: 33–48, 1997.
? 1997 Kluwer Academic Publishers. Printed in the Netherlands.
Architecture and development of the oil-palm (Elaeis guineensis Jacq.) root
Christophe Jourdan and Herv´ e Rey1
D´ epartement des Cultures P´ erennes et Unit´ e de Mod´ elisation des Plantes, Centre de Coop´ eration Internationale
en Recherche Agronomique pour le D´ eveloppement (CIRAD), B.P. 5035, F-34032 Montpellier Cedex 1, France.
Key words: development, Elaeis guineensis, growth, oil palm, root architecture, root system
The growth dynamics and architecture of the oil-palm root system are described. Following a transitional juvenile
phase, eight differentmorphologicaltypes of roots have been distinguished accordingto their developmentpattern
and state of differentiation: primary vertical and horizontal roots, secondary horizontal roots, upward growing
secondary vertical roots and downward growing secondary vertical roots, superficial and deep tertiary roots and
quaternary roots. The relative position of these types of roots determines a morphological and functional unit of
the root system called "root architectural unit" of the oil palm. This root polymorphism enabled us to define a
morphogeneticgradient, which reflected the oil-palm root-system ontogenesis.
Root architecture is a fundamentalaspect of plant pro-
ductivity (Lynch, 1995) through its functional impor-
tance in the efficient acquisition of soil resources. It is
therefore a subject of considerable interest in agricul-
ture and ecology. Root system architecture is directly
linked to the various functions ensured by the roots.
Many authors have demonstrated links between archi-
1993; Fitter, 1986), or with water and nutrient uptake
(Barley, 1970; Bosc and Maertens, 1981; Habib et al.,
The architecture of a branched system can be per-
ceived through architectural analysis. This analysis,
begun by Hall´ e and Oldeman (1970) on the aerial sys-
tems of tropical trees, is based on studying how the
meristems of each axis function and how hierarchical
relations are established between these axes. It also
makes it possible to carry out a complete diagnosis of
the different root types making up the branched sys-
tem, and show the growth, branching, differentiation
and mortality processes, which are characterized for
each category of axes identified. This analysis led on
FAX No: +33467593858. E-mail: email@example.com
to the concept of the root architectural unit (Atger,
1992). Architectural analysis is therefore an appropri-
ate tool for measuring the root architecture of plants.
With practical applications in mind, it can also serve
as a back-up for quantification, modelling, then simu-
lation of the various processes identified.
The oil palm, Elaeis guineensis Jacq., has been
cultivated on a large scale since the turn of the cen-
tury for its high oil yield (triglycerides). According to
Oil World Annual (1995), palm oil production world-
wide in 1995 amounted to 17.2 Mt, placing it second
in the world rankings with 23.7% of edible vegetable
oil production, just behind soybean (27.3%). Howev-
er, knowledge of its root system is still only sketchy.
Paradoxically, the surroundings of the oil-palm root
system have been studied more than the roots them-
selves. In fact, earlier studies have often concentrated
on characterizing soil evolution under oil palm (Cal-
iman, 1990; Hartmann, 1991; Tinker, 1976) and the
hydric properties of these soils (Dufrˆ ene, 1989; Olivin
and Ochs, 1978). Only a few studies have been made
opment (Agamuthuand Broughton,1986; Ruer, 1968;
Tailliez, 1971). Moreover, the isolated observations
carried out were more to solve problemslinked to cul-
tural practices than to increase overall knowledge of
root system development. For example, oil-palm root
system distribution and the position of the absorbing
parts were studied specifically to determine optimum
1967a, 1968; Tailliez, 1971) or solve water supply
problems (Dufrˆ ene, 1989; Ruer, 1969).
Although some authors (Fr´ emond and Orgias,
1952; Purvis, 1956; Ruer, 1967b, 1968; Wright, 1951;
Yampolsky, 1922) have endeavoured to describe in
varying degrees of detail the different axes making up
the oil-palm root system, only a few (Purvis, 1956;
Ruer, 1968) have described its overall structure or
architecture and nobody has touched upon growth and
Our initial aim was therefore to carry out a precise
analysis of oil-palm root system architecture. We shall
then show how architectural analysis enabled us to
reveal root polymorphismin the oil-palm root system,
ontogenesis. Lastly, knowledge acquired recently in
these fields (Atger, 1992; Hall´ e et al., 1978; Kahn,
1983; Kubikova,1967)will enable usto repositionthe
specificities of the oil-palm root system in relation to
those of other plants.
Material and methods
Planting material, study site and climate
The oil palm, Elaeis guineensis Jacq. (Arecaceae) is a
perennial monoecious monocotyledon that originated
in the Gulf of Guinea in West Africa. The variety cho-
sen comes from the family usually known as C1001F,
commonly used in commercial plantations.
The optimum conditions for oil palm cultivation
(IRHO, 1978) are (i) at least 2,000 mm of rain per
year,with a minimumof150mmpermonthandthere-
fore no water deficit, (ii) temperatures of between 20
Our study was conducted at the La M´ e experimen-
tal station in south-eastern Cˆ ote d’Ivoire, where the
humid, subtropical climate with marked seasons is
characterized by (i) average annual rainfall of 1,400
mm with a 380 mm year
1,800 hours of sunshine per year. The study site is
located on a low altitude plateau resulting fromfluvio-
lagoon sedimentation in the tertiary era, at the origin
sand (Hartmann, 1991). These formations, commonly
?C and, (iii) 1,800 hours of sunshine per year.
?1water deficit, (ii) average
knownas "tertiarysands" (OlivinandOchs, 1978),are
and giverise to deep, uniformsoils with neitherappar-
ent interruptionsnor constraints to a depth of at least 6
Architecture and development analysis
Our study was based on concepts developed in plant
above-ground architecture studies (Barth´ el´ emy et al.,
1989; Edelin, 1977, 1984; Hall´ e and Oldeman, 1970;
Hall´ e et al., 1978). The architecture of a plant or root
system can be defined by the type (or category) and
relative position of the different axes making it up. It
results from the functioning of the above-ground and
below-ground apical meristems (Hall´ e and Oldeman,
The architectural analysis of an above-ground
branched system (Hall´ e and Oldeman, 1970) or of
a root system (Atger, 1992) is based on observing
and characterizing the way in which the different axes
(1992): (i) identification and characterization of the
various elements making up the system, (ii) character-
ization of the sequence in which the different compo-
nents of the system occur, and how they develop. The
architectural analysis that led on to the concept of the
architectural model (Hall´ e and Oldeman, 1970) does
not take plant dimensions or their spatial occupation
into account, whereas these criteria are crucial if plant
growth dynamics are to be accurately characterized
and especially if any agricultural applications are to
be found. In our approach, the notions of plant size
and spatial occupationare includedin the architectural
The results are summarized in a table and dia-
grams, which together constitute an architectural dia-
gram (Edelin, 1977), now called an architectural unit
(Barth´ el´ emyetal.,1989).Anarchitecturalunitisbased
on a complete diagnosis of all vegetative axis cate-
gories in the plant. The axes are grouped into types,
egoryofrootaxes, thecharactersused toidentifythem
growth, nature, topological order, presence or absence
of lignification, deciduousness, diameter, growth rate,
Figure1. Photos ofthe oil-palm root system. a. Seedling 36days after germination; b. 3-month-old root system; c. 1-year-old plant; d. Close-up
of 20-year-old palm root system in situ.
Given the absence of morphological traits as dis-
buds, scars, etc.), we had to characterize the root axis
using simple biometrical parameters: length, diame-
ter, inter-lateral lengths, insertion angle, direction of
growth and growth rates.
We studied 4 growth phases: juvenile phase (0-1
year); field establishment phase (3 years), adult phase
(11 years) and the limit of economic viability phase
during these different phases: a static approach based
on a morphological and geometrical study of the root
system, and a dynamic approach based on a study
of development kinetics for the different axes making
up the system, along with their growth and branching
processes (Table 1).
The study techniques were based on partial or total
excavation (Table 1). Static analysis was only car-
ried out for a given stage of development, since it
destroys the roots observed, but a reconstruction of
growth dynamics was possible by studying the differ-
ent development stages on several individuals from a
The investigation was carried out in relatively
homogeneous soil for the juvenile stages and the crop
management sequence was in accordance with IRHO
(1981) recommendations. Ten plants were observed
every fortnight for this juvenile phase. For the adult
stages, under plantation conditions and on homoge-
neoussoils, a representativesample (30 to 40 roots) of
each root order was taken and analysed. The direction
of growth, angle of emission on the palm and reori-
entation (final direction of growth) were measured.
These observations were made on lateral profiles, i.e.
by digging a trench perpendicular to the planting row
and carefully removing the soil from around the roots
found. Root samples, were taken from the palm to a
distance of 2 m, down to a depth of 1.5 m. These mea-
surements were completed by frontal profiles (B¨ ohm,
1979) carried out in the planting row from 0 to 2 m
down, in order to obtain vertical root density maps.
Spatial distribution of the roots was observed, along
with their number, type (young, growing, dead, etc.)
and their diameter. A few surface roots were also fol-
lowed from the stem to their apical tip, primarily to
ly, total root-soil plate excavations were carried out to
makean exhaustivecountof the rootsemitted, charac-
terize the mortality rate and locate new emissions.
For juvenile root systems aged 0 to 3 months, the
dynamic analysis (Table 1) mostly involved the use
of miniature root-growth chambers fitted with Nylon
mesh (Le Roux and Pag` es, 1994) and measuring 100
cm high by 60 cm wide by 1.6 cm thick. Seven minia-
grown from germinated seeds, were inclined at an
angle of 30
side under slight shading, the substrate used was poor
forest topsoil (Jourdan et al., 1995), and the soil tem-
perature varied between 24
during the day.
For juvenileroot systemsaged 3 to 12 months, two
wooden miniature root-growth chambers were built:
(i) one was 47 cm high, 59 cm wide and 8 cm thick,
containing three 3-month old plants and (ii) another
74 cm high, 54 cm wide and 26 cm thick, contain-
ing one 7-month old plant. The two miniature root-
growth chambers had 2 glass sides and were placed in
The Nylon mesh was not used in this case, due to the
volume occupied by the roots. The substrate and soil
temperatures were the same as before.
For root systems in the field, 20 "field rhizotrons"
were set up in two positions, near vertical and near
horizontal, and at different distances from the palms.
The vertical "field rhizotrons" were square, with 1 m
sides, and sloped 20
window was held against the soil by a metal frame
propped up by bamboo poles. The horizontal "field
rhizotrons" were rectangular (0.5 m ?1 m) and were
inclined 20 to 30
The progression of the roots was recorded using
differentcolouredindelible felt pensat 3-day intervals
for the minirhizotrons and every 7 days for the "field
rhizotrons". Once these multicoloured records were
software developed and described by Colin-Belgrand
et al. (1989)),varioustypesofprocessescanbecarried
out (growthkinetics, rhythmratios, dateson whichthe
different roots appear, etc.).
?from the vertical. They were placed out-
?C at night and 32
?from the vertical. The Plexiglas
?from the horizontal.
Numerous root samples were taken, at all stages of
oil palm development, in order to characterize cell
differentiation in the various tissues, and validate the
root typology already discovered. The samples were
fixed in a solutioncomprisinga phosphatebuffer,25%
glutaraldehyde, 10% paraformaldehyde and caffeine.
After embedding in resin, they were sliced with a 3
acid and aniline blue-black (Fisher et al., 1968).
Table 1. Overview of the various methods of investigation of the root system for the four growth phases studied in this work
Static approach Dynamic approach
Vertical and horizontal
growth chambers <<field rhizotrons>>
Root system Lateral and frontal Root-soil plate
Growth phases studiedexcavations
Juvenile (0-1 year)
Field establishment (3 years)
Adult (11 years)
Limit of economic viability (20 years) partial
Table 2. Quantitative parameters of different root types of oil palms in the juvenile phase. RI1?5-primary root emitted for a 1 to
5-month-old oil palm, RI5?12-primary root emitted for a 5 to 12-month-old oil palm
Medium roots RadicleRI5?12
Long rootsShort roots
0.440.88 1.020.39 0.30 0.27
(cm)50 100 180 50101.5
(cm)0.10< ?<0.200.20< ?<0.30 0.30< ?<0.400.05< ?<0.100.03< ?<0.060.02< ?<0.05
vascular bundles 10 to 1515 to 25 25 to 408 to 105 to 8 3 to 5
lateral roots (cm)
Juvenile phase from 0 to 1 year in the nursery
During the juvenile phase, from 0 to 1 year, we differ-
entiatedthe radicle(RAD) fromthe set of adventitious
roots emitted (RI).
The radicle is downwardly oriented, slightly undulat-
(Figure1a). Its growthis continuous,with no apparent
rhythm, limited in time and the average growth rate
is around 0.44 cm day
than 50 cm long with an average diameter of 1.5 mm.
Its ramification is diffuse, dense and covers the entire
axis, apart from the apical zone, which is of relatively
?1(Table 2). It is rarely more
of lateral formations can be identified (Jourdan et al.,
1995); (i) "long", (ii) "medium" and (iii) "short", all
roots are mostly located at the base of the radicle and
are emitted very soon after germination. They vary in
length and are usually over 10 cm long, growing at an
average rate of 0.39 cm day
roots bear two types of roots: (i) "medium" and (ii)
"short". The so-called "medium" roots do not exceed
10 cm in length, they grow at an average rate of 0.3
above. Irrespective of their topological order within
the branched system, these "medium" roots are mor-
phologically similar to each other. They bear "short"
lateral roots with the same density as above. The so-
called "short" roots do not exceed 1.5 cm in length,
?1(Table 2). These "long"
?1(Table 2), and are distributed over the entire
Figure 2. Variation in the cumulated average number of roots ( ?)
and leaves (?) emitted per palm during the 0-1 year period. The
emission rate values for roots (R
period are expressed as roots or leaves emitted per year.
r) and leaves (R
l), during each
grow at an average rate of 0.27 cm day
and are non-branching. They differ from one another
by their origin, i.e. the type of the axis by which they
are borne (Figure 1a).
Adventitious primary roots (RI)
On average, the first adventitious root is emitted a
month after germination (Figure 2). During the first
year, the subsequent RIs are emitted at an accelerating
rhythm, whereas leaf emission remains constant (21
leaves per year). During this period, the RIs take on
a branched structure similar to that developed by the
radicle, insofar as three branching orders are observed
(Figure 1b). Their direction of growth is vertical and
downwards. The average growth rate evolves from
0.88 cm day
(RI5?12). The final length varies from 1 m at 5 months
to 1.8 m at 1 year. They are white when young, but
rapidlyturnlight todarkbrown(Figure1c). Thediam-
eter remains constant for each root emitted but varies
from 0.2 to 0.4 cm depending on the date of emis-
sion between 1 and 12 months, the roots emitted at
12 months having a larger diameter. The number of
vascular bundles depends on root diameter and ranges
from 15 to 25 at 1 month and from 25 to 40 at 1 year.
Branching, which is comparable to that of the radi-
cle, is diffuse and dense. The lateral roots produced,
which are also of several types, are distributed along
?1(Table 2) for 1- to 5-month old palms
the whole length and around the entire circumference
of the RIs, inserted at an angle of 90
distance between laterals increases from 0.24 cm for
RI1?5to 0.84 cm for RI5?12along with an increase in
um" and "short" roots described above are also found
here, but in different proportions. The proportion of
"long" roots increases, with a gradual disappearance
of "medium" and especially "short" roots. The "long"
roots are mostly distributed in the proximal quarter of
the RIs. The characteristics of these lateral formations
are similar to those borne by the radicle, but with a
slight increase in diameter.
?. The average
Field growth phases from 1 to 20 years
Primary roots (RI) and secondary roots (RII)
New RIs are mostly emitted on the periphery of the
inserted at soil level. Two distinct types of RIs have
been described in the field: vertical, downward grow-
ing RIs, RI VDs, and horizontal RIIs, RII Hs (Figures
1D and 3). Their number increases steadily up to 11
years, then the number of RI VDs stabilizes where-
as that of the RI Hs continues to increase (Figure 4).
These two types of RI therefore differ in number and
rootsthey bear. The RI VDs, with an averagediameter
of 0.45 cm, mostly bear horizontal RIs (RII H) and are
only slightly ramified (Table 3). These RIIs are quite
short (less than 20 cm), highly branched and distribut-
ed around the entire circumference of the RIs. The RI
Hs, with an average diameter of 0.68 cm, bear both
upwardly growing vertical RIIs (RII VU), of medium
length (0.5 to 2 m) and highly branched, and down-
long (more than 3 m) and only half as branched as
the RII VUs (Table 3). The very precise direction of
these RIIs, either upward or downward, gives the RI
from the radial symmetry of the RI VDs (Figures 1D
and 3). The average diameter of the RII Hs, RII VUs
and RII VDs is more or less the same, approx. 0.2 cm
The RI VDs and RII VDs have the characteristics
of orthotropic axes (vertical growth, radial symme-
try, lateral formations emitted radially at an angle of
to be positive orthogeotropic axes. Likewise, the RII
VUs, which are also orthotropic axes but with nega-
?), with positive geotropism; they are therefore said
Figure 3. Oil-palm root architectural unit. Partial drawing of the root system on a 10-year-old oil palm with the 8 root types described. Apex
death indicated by
?. The harvesting pole measures 3.5 metres.
Figure 4. Variation in the average number of RI H ( ?) and RI VD (?) emitted per tree over time.
tive geotropism,are said to be negativeorthogeotropic
axes. On the other hand, the RI Hs are plagiotrop-
ic axes (horizontal growth, bilateral symmetry, lateral
formations emitted at 90
reactions to gravity, and are said to be plagiogeotropic
?in the same plane), without
axes. RII Hs retain generally horizontal growth, they
are ageotropic. These roots have radial symmetry and
the lateral axes (RIII) are emitted radially in all direc-
Figure 5. Anatomy and morphology of oil palm roots. A. Histological cross-section of an RI; B. Histological cross-section of an RIV; C.
RIV seen under a scanning electron microscope; D. General view of a pneumathode under a scanning electron microscope. Abbreviations:
a–aerenchyma, c–cortex, e–endodermis, h–hypodermis, p–pith, rh–rhizodermis, s–sclerenchyma, v–medullary vessel.
Table 3. Quantitative parameters of different root types of oil palms in the << adult >> phase of field cultivation
<< Adult >> phase
RI VDRI HRII VU RII VDRII HsRIIIdRIII RIV
0.30 0.30 0.200.200.200.080.08-
(cm) 6002500 200 6002020101.5
(cm)0.41< ?<0.50 0.50< ?<0.80 0.15< ?<0.25 0.20< ?<0.25 0.13< ?<0.22 0.05< ?<0.14 0.07< ?<0.15 0.03< ?<0.07
vascular bundles 34 to 4534 to 4515 to 2515 to 25 15 to 258 to 128 to 123 to 5
RI H,RI VD andRIIVD growthisoftheindefinite
type, theirfinallengthrangesfromtwentyorso metres
for the RI Hs to over 6 m (limit of observations) for
the RI VDs and RII VDs. RII VU growth is limited by
3). The average growth rate of RIs in the field evolves
( ? 0.11) at 11 years, remaining stable thereafter. For
RIIs, the rates decrease from 0.27 cm day
A few centimetresfrom the apical region they are usu-
ally white, becoming brown then purple and final-
ly black a metre from the apex. The aerenchyma of
RIs and RIIs is highly developed and may take up
almost 70% of root volume (Figure 5a). The differ-
ence between RIs and RIIs lies mainly in the size of
the stele, which is larger in RIs. An average of 20
vascular bundles were observed in each RII and 40 in
a few dozen centimetres from the apex. Such lignifi-
cation first affects the xylem vessels, then the sub-
hypodermic region and the pith, then the endodermis
poles and the aerenchyma remain non-lignified.
Under oil palms, at less than 50 cm from the root-
soil plate, the RIs (especially RI VD) bear numerous
?1( ?0.2)at 3 yearsto 0.3cm day
?1( ? 0.1) at
relay axes emitted after the death and self pruning of
successive bearing axes (Figure 3). This is not a rare
occurrence, indeed it affects 12.7%, 41.3% and 58.8%
of primary roots emitted by palms aged 3, 11 and 20
Tertiary roots (RIII) and quaternary roots (RIV)
All RIIs in the field emit RIIIs at an average angle of
size and branchingcriteria. The RIIIsborneon RII Hs,
and the RIIIs borne on RII VDs are generally shorter
and much less branched than those borne by RII VUs.
The least branched RIIIs mostly found deep down,
are designated dRIII; the highly ramified RIIIs, found
mostly near the surface, are designated sRIII (Table 3
and Figure 3). These two types of RIII show definite
growth, with dRIII rarely exceeding 10 cm in length
and sRIII 20 cm. They are beige to brown in colour
and never turn very dark. They are around 0.1 cm in
average diameter and have ten or so vascular bundles.
There is no preferential growth direction; these roots
The branched system formed by RII VUs and sRI-
IIs sometimes forms a veritable "root mat" more than
20 cm thick. Numeroustraumaticincidentscan trigger
the emission of relay axes, which increase the branch-
ing potential of these roots (Figure 6). RII VUs and
sRIIIs even come out of the ground when conditions
?, which could only be distinguished according to
Figure 6. Drawing of an RII VU taken from the 0-10 cm horizon. Abbreviations: pn–pneumathode, r–reiterated RII VU.
and may penetrate the remnants of decomposing peti-
oles and leaves left in the windrows. The existence of
numerous pneumathodes (Yampolsky, 1924) on these
roots (Figure 6) is indicative of saturating humidity.
When conditions become unfavourable, such as in
the dry season, the roots located on the surface die
and decompose. Thereafter, when abundant rainfall
ting numerous relay axes again, which reach the sur-
face horizons that are rich in organic matter. All RIIIs
bear identical RIVs (Figure 3): ageotropic, white, less
than 1.5 cm long with an average diameter of 0.05 cm,
RIIIs and RIVs do not reveal any tissue lignification
similar to that occurring in RIs and RIIs, and they are
therefore classed as non-woody.
an average increase ranging from 0.14 cm day
0.07) for 3-year old palms to 0.08 cm day
for 11-year olds, stabilizing thereafter. RIV growth in
the field was not measured because their small size
prevented reliable values from being obtained with a
weekly observation rate.
Some characteristicsare commonto all typesof oil
palm roots, such as the existence of calcium oxalate
precipitates in bundles of raphides in cells on the
periphery of the cortex, and endomycorrhizae fila-
ments. The epidermal cells (rhizodermis) located near
?1( ? 0.04)
the root apex have a specific shape, similar to that of
a grain of maize and are regularly distributed on the
surface of each root (Figure 5c). Many pneumathodes
have also been observed, especially on surface roots
in the field (RIII and RIV), but also on the radicle, RIs
and RIIs in the juvenile phase. They are characterized
the internal cells of the cortex. On lateral roots, such
splitting only occurs in the few millimetres near the
zoneofinsertion, wherea sleeve ofintactcellspersists
To conclude, 6 types of roots were found in the
juvenile phase and 8 types in the adult phase using
simple morphological and geometrical criteria. These
results are shown in Table 4 which summarizes the
ing and differentiation, but also the layout of the dif-
ferent axes making up the adult oil-palm root system.
unit of the oil palm.
Discussion and conclusion
Diversity of differentiation states
The oil-palm root system revealed the existence of
diversity in root diameter, growth rate and direction,
things. Despite this apparent diversity, our architec-
tural analysis tools enabled us to establish a typology
of different root axes. We detected 8 stable root types
Each root type has morphological,morphogeneticand
functional properties which differentiate it from the
otherroottypes(Atger, 1992).It thereforerepresentsa
population of homogeneousroots with a set of similar
characters (structure, way of functioning, etc.). Then
it characterizes the specific state of differentiation of
the set of meristems of the plant (Hall´ e and Oldeman,
1970), which helps to maintain the polymorphism of
root systems (Le Roux and Pag` es, 1994).
Relations between differentiation criteria
Close relations exist between the various differentia-
tion criteria of root system axes. In the literature, there
cal diameter of roots and their growth rate, such as in
and Wilson, 1964), the pine tree (Wilcox, 1968) or
some cereals (Hackett, 1969a, 1969b), and with their
internal anatomical structure, particularly the number
of xylem poles (Charlton, 1967) or the diameter of
the primary xylem (Horsley and Wilson, 1971). Other
relations bringing into play root length and diameter,
branchingdensity and the action of external factors on
root development are summarized by Coutts (1987).
As far as the oil-palm root system is concerned, roots
that have high growth rates, long life spans and low
density branching (such as primary roots), have both
a large diameter and a large number of woody poles
(Tables2 and 3). On the other hand, unbranchedroots,
with a low growth rate that is limited in time (such
as quaternary roots), have a very small diameter and
number of woody poles. Likewise, we showed (Jour-
dan, 1995) that there is a close relation between root
diameter and branching parameters (case of RIIs in
the field). Other correlations have also been found,
notably between the diameter of juvenile RIs and the
the growth rate and the length of the unbranched api-
cal zone of radicles. The diameter is therefore one of
the main criteria for differentiation of root types and
should be taken into account in architectural analysis
of the oil-palm root system.
Root system hierarchy
The existence of several levels of differentiation in
a root system was described at turn of the century in
Since then, several attempts at root axis classifica-
tion have been published and were summarized by
Kubikova (1967): anchorage roots and feeding roots,
long roots and short roots (Wilcox, 1964), pioneer
roots and feeding roots, or even woody roots and
non-woody roots (Lyford and Wilson, 1964). This
terminology has been clearly defined by Sutton and
Tinus (1983). Root systems were thus made up of 2
types of qualitatively different axes: macrorhizae and
brachyrhizae (Jenik and Sen, 1964). Macrorhizae are
the large roots that explore the soil and extend the rhi-
zosphere (Kahn, 1977); they are vertical with marked
positive geotropism in which case they represent the
"orthotropic state", or horizontal in which case they
represent the "plagiotropic state". These axes persist
within the system, their secondary growth can be con-
siderable, they have a large number of woody poles,
a highly developed pith, a long and pointed apex of
large diameter and, lastly they show strong reiteration
ability (Oldeman, 1974). According to many authors,
these roots also ensure the anchorage and the stabil-
ity of trees in the soil (Coutts, 1986; Quine et al.,
1991). Brachyrhizae are fine roots that are not part of
the root main structure but form the ephemeral roots
at the summit of the plagiotropichierarchy,like leaves
in the above-ground system (Kahn, 1983). They are
characterized by a small number of woody poles, the
absence of pith, a substantial cortex, a short apex with
a small diameter, a limited life span and no reitera-
tion ability. Although their respective roles are clearly
established, conductionin macrorhizaeand absorption
in brachyrhizae (Kubikova, 1967; Lyford and Wilson,
1964), the distinction between these two root cate-
gories is not always so clear. In fact, all the interme-
diate axes between macrorhizae and brachyrhizae can
be found within the same root system and a given axis
may switch from one morphological and functional
structure to another (Wilcox, 1964).
woody roots, have mostly been described for woody
Dicotyledons and Gymnosperms. In the oil palm, the
macrorhizae would be the RIs and RIIs observed in
the field, and the brachyrhizae would be the RIIIs and
RIVs. Indeed, the RI VDs and RI Hs along with the
RII VDs, play a full role in the anchorage of the tree
and in exploring the volume of available soil, whereas
the RIIIs and RIVs mostly exploit it. Nevertheless, oil
brachyrhiza criteria, since, most importantly, they are
of pith in the fine roots and by reiteration ability in the
The oil-palm root system, and more generally that
of the Monocotyledons, reveals other analogies with
the root system of Dicotyledonsor Gymnosperms.For
ic tap root, or radicle, from which horizontal lateral
roots are emitted. The main difference is the absence
of radicle perenniality: it self-prunes 3 to 4 months
after germination. Contrary to the so-called primary
root system (Kahn, 1983), in which the single primary
root is orthotropic and governs the plagiotropic dif-
ferentiation of its lateral formations, the "orthotropic
of vertical primary roots (juvenile RIs, then RI VDs),
which succeed each other whilst maintaining the same
behaviour. The “plagiotropic state” in the oil palm is
characterized by a set of horizontal primary roots (RI
H), emitted more than a year after germination, which
succeed each other in time and mostly bear vertical
lateralroots. This"plagiotropicstate" wasobservedby
Granville (1974) in another palm, Mauritia flexuosa,
which has 40 m long horizontalroots that are flattened
at 30 m and beyond! The "plagiotropic" roots of the
oil palm do not seem to be directly under the influence
of the apical meristems of the orthotropic roots since
they do not grow from them, whereas that is the case
in most woody plants, such as the cocoa tree (Dyanat-
oak (Beissalah et al., 1988; Champagnat et al., 1974;
in a root axis is defined according to Kahn (1977) by
the "orthotropic state"; RI Hs do not therefore seem to
correspond to this definition. Lastly, the "plagiotropic
state" of the oil palm is complex insofar as the RI Hs
give rise to two clearly distinct types of lateral roots,
ly branched "sRIIIs + RIVs" combination making up
the root mat, and the other growing downwards (RII
in the root system, after a transitional juvenile phase
(Granville, 1974) lasting roughly a year.
acterized by the succession in time of three phases: (i)
a very brief phase where the radicle reveals a marked
"orthotropic state", similar to the tap root of woody
acterize (still by analogy with the tap root system) an
"orthotropicstate" and lastly, (iii) a muchlongerphase
where the RI VDs represent an "orthotropic state" and
the RI Hs a "plagiotropic state". The oil-palm root
system thus appears as a structured set of differen-
tiated and hierarchized roots, with clearly identified
corresponding functions: exploration and exploitation
of the soil. In fact, it constitutes a veritable morpho-
logical and functional unit that we have called "root
architectural unit" (Table 4) in reference to the work
by Edelin (1984)on the above-groundsystem, then by
Atger (1992) on the root system of trees.
Root growth and development
minationto the adultphase, showthat the formationof
disordered, but in accordance with a well determined
programme. Completion of this programme, which in
fact reflects development of the different meristems
according to their specific differentiation sequence,
governs root architecture development.
The radicle of the oil palm, and of Monocotyledonsin
general, has a limited life span, since it is a low diam-
eter axis, with exclusively primary growth (Hall´ e et
al., 1978). According to these authors, as the absorb-
ing zones are extended through branching, the point
of attachment of the radicle to the above-ground part
becomes a veritable "bottleneck" and the radicle alone
is unable to meet increasing demand from the above-
ground organs. The plant then emits numerous adven-
titious roots, which thus by-pass the "bottleneck".
Throughout the life of an oil palm, each root emit-
ted keeps the same morphologicalcharacteristics, par-
ticularly the same diameter, up to its death. More-
over, during the juvenile phase these characteristics
the emission date (Jourdan, 1995). The roots emitted
on a certain date have, among other things, a larger
diameter, faster growth, a lower branchingdensity and
greater life span than those emitted before this date.
It is only when all the different root types inventoried
have been expressed that the characteristics of each
new primary root emitted remain identical to those
of the previous primary root. This growth process,
Table 4. Oil-palm root architecture unit. W–woody, NW–non woody, I–indefinite growth, D–definite growth,
be exceeded, DS–diffuse and sylleptic branching, NB–non-branched, R–radial symmetry, B–bilateral symmetry, L–long-term self-pruning, more
than 6 months, M–medium-term self-pruning, from 1 to 6 months, S–short-term self-pruning, less than or equal to 1 month
?–maximum observed value that can
parameter RI VD RI H RII VURII VD RII HsRIIIdRIII RIV
Woody axisWWWWW NWNW NW
0.530.63 0.180.23 0.150.10 0.10 0.05
where the organs develop more and more up to the
perduring ultimate stage, is similar to the "establish-
ment growth" defined by Tomlinson and Zimmerman
(1967) for the above-ground part of a palm, Rhapis
excelsa, and described since in other Monocotyledons
(Tomlinson and Esler, 1973) and certain Dicotyledons
considered to be "primitives" (Blanc, 1986).
The dynamics of space occupation
Kahn (1980, 1983) defined 4 phases, which cover all
the rooting processes of forest plants: (i) establish-
ment of the orthotropic axis which produces a few
root mats, (ii) production of plagiotropic macrorhizae
from the initial axis, (iii) considerable development of
these axes which extend the exploitation zone from
the trunk and thereby delimit a distal exploitation ring
and (iv) production of new plagiotropic macrochizae
from the base of the tap root or trunk, which exploit
previously abandoned proximal space. However, cer-
tain plants do not complete all 4 phases, e.g. single-
stem plants in the understorey, most shrubs and a
few trees in which adaptive reiteration is non-existent
(Kahn, 1983). When observing the different develop-
ment phases of the oil-palm root system, the different
phases described by Kahn (1983) are generally found.
Indeed, we have shown that the juvenile root system
(from 0 to 1 year) was characterized by an "orthotrop-
ic state" in the radicle and juvenile RIs. The following
than a year after germination, corresponding to pla-
giotropic macrorhizae. These horizontal roots explore
and exploit superficial regions distant from the palms,
of new exploitationof the proximalspace bycontinual
the palm (Jourdan, 1995). Consequently, the proximal
space is permanently exploited by the oil palm, unlike
the case of forest trees described by Kahn (1980) and
Atger (1992). The dynamics of space occupation by
the oil-palm root system can thus be expressed in dia-
grammatic form as in Figure 7.
Synthesis of root development
During its ontogenesis, the oil palm forms different
types of roots in an ordered sequence. The "transient"
radicle is rapidly replaced by numerous adventitious
roots that gradually produce increasingly developed
lateral formationstowardsthe peripheryofthe system,
along the lines of growth by intercalation (Atger and
Edelin, 1994; Edelin, 1984). Eventually, this system
Figure 7. Various stages in oil palm rooting. The radicle stage (A) is rapidly replaced by the juvenile RI stage (B), then by the adult stage
(side view: C and view from above: D). The RI VDs and RII VDs, along with the RI Hs extend the oil palm’s exploitation zone at depth and
superficially respectively. Continuous root emission and the RI total reiteration process result in continuous exploitation of the proximal space
comprises a finite number of root types that charac-
terize the root architectural unit of the oil palm. The
oil-palm root system, along with those of herbaceous
and other woody monocots presents a relative rigidi-
ty in its developmental patterns. The continuity in the
development characteristics of the different root types
inventoried means that a morphogenetic gradient can
be defined betweenthese differenttypes, independent-
ly of their branchingorder. They are mainly character-
ized, throughthe architectural analysis, by a whole set
of various morphological and morphogenetic parame-
ters (Table 4), essentially the root diameter.
malization is possible and quantitative architectural
describe various architecture building processes, such
as branching (Jourdan et al., 1995) and to summarize
all the knowledgeacquiredthroughmorecomplexfor-
malization(Reffyeet al., 1991).Anappropriatedigital
processing sequence (modelling, simulation, display,
postprocessing on digital 3-D mock-ups) can then be
envisaged to achieve practical applications of agro-
nomic interest. This will be made possible under the
condition of both acquiring supplementaryknowledge
in the environmental effects on the root system devel-
opment and in the complex phenomena of intra- or
interspecific root competition.
We would like to thank Kouam´ e Brou, Director of the
IDEFOR-DPO station at La M´ e in the Cˆ ote d’Ivoire,
for his hospitality and for the facilities made available
locally for carrying out the observations.
Agamuthu Pand Broughton WJ1986 Factors affecting the develop-
ment of the rooting system in young oil palm (Elaeis guineensis
Jacq.). Agric. Ecosyst. Environ. 17, 173–179.
Atger C 1992 Essai surl’architecture racinaire des arbres. Th. Doct.,
Univ. Montpellier II, France. 287 p.
Atger C and Edelin C 1994 Premi` eres donn´ ees sur l’architecture
compar´ ee des syst` emes racinaires et caulinaires des arbres. Can.
J. Bot. 72, 963–975.
47 Download full-text
Barley K P 1970 The configuration of the root system in relation to
nutrient uptake. Adv. Agron. 22, 159–201.
Barth´ el´ emy D, Edelin C and Hall´ e F 1989 Architectural concepts
for tropical trees. In Tropical Forests: Botanical Dynamics, Spe-
ciation and Diversity. Eds. L B Holm-Nielsen, I Nielsen and H
Balslev. pp 89–100. Academic Press, London.
Beissalah Y, Amin T, El Hajzein B and Neville P 1988 Formation
des racines de r´ eg´ en´ eration chez Quercus ilex L. Bull. Soc. Bot.
Blanc P 1986 Edification d’arbres par croissance d’´ etablissement
de type monocotyl´ edonien : l’exemple des Chloranthaceae. In
L’ Arbre. Ed. C Edelin. pp 101–123. Naturalia Monspeliensia,
B¨ ohmW1979Methodsofstudyingrootsystems.Ecological Studies
Springer-Verlag, Berlin. 188 p.
Bosc M and Maertens C 1981 Rˆ ole de l’accroissement du syst` eme
racinaire dans l’absorption de divers ´ etats du potassium du sol.
Agrochimica 25, 1–8.
Caliman J-P 1990 D´ egradation de propri´ et´ es physiques condition-
nant la fertilit´ e des sols sous culture de palmier ` a huile en Cˆ ote
d’Ivoire. Essai de correction. Th. Doct., Univ. Dijon. 219 p.
Champagnat P, Baba J and Delaunay M 1974 Corr´ elations entre
le pivot et ses ramifications dans le syst` eme racinaire de jeunes
chˆ enes cultiv´ es sous un brouillard nutritif. Rev. Cytol. Biol. V´ eg.
Charlton W A 1967 The root system of Linaria vulgaris Mill. II.
Differentiation of root types. Can. J. Bot. 45, 81–91.
Colin-Belgrand M, Pag` es L, Dreyer E and Joannes H 1989 Analysis
and stimulation of the architecture of a growing root system :
application to a comparative study of several tree seedlings. Ann.
Sci. For. 46, 288s–293s.
Coutts M P 1983 Root architecture and tree stability. Plant Soil 71,
Coutts M P 1986 Components of tree stability in Sitka spruce on
peaty gley soil. Forestry 59, 173–197.
Coutts M P 1987 Developmental process in tree root systems. Can.
J. For. Res. 17, 761–767.
Dufrˆ ene E 1989 Photosynth` ese,
mod´ elisation de la production chez le palmier ` a huile (Elaeis
guineensis Jacq.). Th. Doct., Univ. Paris-Orsay. 156 p.
Dyanat-Nejad H and Neville P 1972a Etude exp´ erimentale de
l’initiation et de la croissance des racines lat´ erales pr´ ecoces du
cacaoyer (TheobromacacaoL.).Ann.Sci.Nat.Bot.13, 211–246.
Dyanat-Nejad H and Neville P 1972b Etude sur le mode d’action du
m´ erist` eme radical orthotrope dans le contrˆ ole de la plagiotropie
des racines lat´ erales chez Theobroma cacao L. Rev. G´ en. Bot.
Edelin C 1977 Images de l’architecture des conif` eres. Th. Doct.,
Universit´ e Montpellier II. 255 p.
Edelin C 1984 L’architecture monopodiale : l’exemple de quelques
arbres d’Asie tropicale. Th. Doct. Etat, Univ. Montpellier II. 258
Ennos A R, Crook M J and Grimshaw C 1993 The anchorage
mechanics of maize, Zea mays. J. Exp. Bot. 44, 147-153.
Fisher D B, Jensen W A and Aston M E 1968 Histochemical studies
of pollen: storage pockets in the endoplasmic reticulum (E.R.).
Histochemie 13, 169–182.
Fitter A H 1986 The topology and geometry of plant root systems:
influence of watering rate on root system topology in Trifolium
pratense. Ann. Bot. 58, 91–101.
Fr´ emond Y and Orgias A 1952 Contribution ` a l’´ etude du syst` eme
radiculaire du palmier ` a huile. Ol´ eagineux 7, 345–350.
?33. Eds. W D Billings, F Golley, O L Lange and J S Olson.
consommation en eau et
Granville JHde1974Aper¸ cusurlastructure despneumatophores de
L.etEuterpe oleracea Mart. (Palmae). G´ en´ eralisation ausyst` eme
respiratoire racinaire d’autres Palmiers. Cah. Orstom, S´ er. Biol.
Habib R, Pag` es L, Jordan M O, Simonneau T and S´ ebillotte M 1991
Approche ` a l’´ echelle du syst` eme racinaire de l’absorption hydro-
min´ erale. Cons´ equences en mati` ere de mod´ elisation. Agronomie
Hackett C 1969a Quantitative aspects of the growth of cereal root
systems. In Root Growth. Ed. W J Whittington. pp 134–147.
Butterworth and Co. Ltd., London.
Hackett C 1969b A study of the root system of barley. II. Relation-
ships between root dimensions and nutrient uptake. New Phytol.
Hall´ e F and Oldeman R A A 1970 Essai sur l’architecture et la
dynamique de croissance des arbres tropicaux. Masson, Paris,
Hall´ e F,Oldeman RAA,and TomlinsonPB1978Tropical trees and
forests. An architectural analysis. Springer-Verlag, Berlin. 44 p.
incereals andgrain legumes: Howwell arethey correlated? Aust.
J. Agric. Res. 38, 513–527.
ferrallitiques sousculture depalmiers ` ahuile. Casdelaplantation
R Michaux ` a Dabou (Cˆ ote d’Ivoire). Th. Doct., Univ. Paris VI.
of the root system of Betula papyrifera. Am. J. Bot. 58, 141–147.
IRHO 1978 Rapport d’activit´ es 1976-1977 Palmier ` a huile et cocoti-
er. Etude du milieu naturel. Ol´ eagineux 33, 29–36.
IRHO 1981 La culture du palmier ` a huile. Fascicule 1. Les stades
juv´ eniles. Rapport interne CIRAD, La M´ e. 108 p.
Jenick J and Sen D N 1954 Morphology of root systems in trees :
a proposal for terminology. In 10th Intl. Bot. Congr. (Edinburg),
Abstr. of Papers. pp 393–394.
Jourdan C 1995 Mod´ elisation de l’architecture et du d´ eveloppement
du syst` eme racinaire du palmier ` a huile. Th. Doct., Univ. Mont-
pellier II. 243 p.
Jourdan C, Rey H and Gu´ edon Y 1995 Achitectural analysis and
modelling of the branching process of the young oil-palm root
system. Plant Soil 177, 63–72.
Kahn F1977 Analyse structurale dessyst´ emes racinaires desplantes
ligneuses de la forˆ et tropicale dense humide. Candollea 32, 321–
Kahn F 1980 Comportements racinaire et a´ erien chez les plantes
ligneuses, de la forˆ et tropicale humide (sud-ouest de la Cˆ ote
d’Ivoire). Adansonia 19, 413–427.
Kahn F1983 Architecture compar´ ee desforˆ ets tropicales humides et
dynamique de la rhizosph` ere. Th. Doct. Etat, Univ. Montpellier
II. 426 p.
Kubikova J 1967 Contribution to the classification of root systems
of woody plants. Preslia, Praha 39, 236–243.
Le Roux Y 1994 Mise en place de l’architecture racinaire d’Hevea
brasiliensis. Etude compar´ ee du semisetde lamicroboututre. Th.
Doct., Univ. Aix-Marseille III. 295 p.
Le Roux Y and Pag` es L 1994 D´ eveloppement et polymorphisme
racinaires chez de jeunes semis d’h´ ev´ ea (Hevea brasiliensis).
Can. J. Bot. 72, 924–932.
Lyford W and Wilson B F 1964 Development of the root system
of Acer rubrum L. USDA Forest Service, Harvard For. Paper,
Harvard Univ. 10, 1–17.
LynchJ1995Rootarchitecture andplant productivity. Plant Physiol.