Variation in the Biological N, Fixation by Tree ...
Variation in the Biological N, Fixation by Tree
Legumes in Threé Ecological Zones from the North
to the South of Senegal
<.
,j
’
M., NDIAYE
Jf ? Pj.w~ *lc;q+
Institut Sénégalais de Recherches Agricoles
,
Centre Nord Bassin Arachidier
JJ7/73-
.,
Bambey, Sénégal
,
. . ..b...1,.
I\\
i-4. ‘ii..,,. , <a “.,
&
.-. ,.” -.. . . . . * ._._,-.
F. GANRY
,_
.Iv.&
.w -.- .-4. d ,” / i-&*.,“,
Centre de Coopération Internationale en Recherche Agronomique pour le
Développement
Département des Cultures Annuelles
Montpellier, France
The utilization
NJïxing trees (NFTs) in agroforestry, in semiarid regions, is O&n
limited by weak N, fixation. This study was conducted to evaluate the N&ing capac-
ity of trees growing under natural conditions for 10 years with neither chemical fertil-
izer nor inoculum application. For this purpose, natural isotopic “N abundance (b5N)
of leaves harvested from NFTs and reference plants (zen growing in three agroeco-
logical zones of Senegal were analyzed. At each site (one site per zone), average g5N
ref was calculated N contents and b5N values increasedfiom the north in the Sudano-
Sahelian zone (Bambey and Nioro) ta the south in the sub-Guinean zone (Djibelor).
The percentage ofjixed N, (%Ndfa) was low at both Bambey (Ndfa 5 22%) and Niera
(Ndfa c; 39%) and very high at Djibelor (76% 5 Ndfa s 95%). TO explain the low lev-
els of Ndfa observed at Bambey and Nioro, the hypothesis of lack of Bradyrhizobium
strains in these soils was made. This hypothesis was confirmed by results obtainedfrom
a tria1 conducted under controlled conditions, which showed absence of nodulation on
Gliricidia sepium growing under optimal conditions of water and nutrient supp[y.
Improvement of NZ fiation of NFTs in the semiarid zone is a prerequisite for imple-
mentation of sustainable agroforestry systems.
Keywords N,-fixing trees, natural “N abundance, percent of fixed N:!, Sudano-
Sahelian and sub-Guinean zones
Received 20 August 1996; accepted 12 December 1996.
We thank P. N. Sall, S. Badiane, S. A. Ndiaye, B. Ndour, and 1. Diaïté (researchers at the Insti-
tut Sénégalais de Recherches Agricoles, ISRA) for allowing us to sample tree legumes, and 1. Dioum
and G. Deme (technicians at ISRA) for their help in sampling trees and soils. We are very grateful
for the advice in tree sampling received from Dr. A. M. Domenach. We also thank R. Oliver for his
assistance in analyzing natural ‘% abundance. Professor Yvon R. Dommergues is acknowledged for
his valuable comments on earlier versions of the manuscript. The comments of anonymous review-
ers are much appreciated.
Address correspondence to Dr. F. Gam-y, Centre de Coopération Internationale en Recherche
Agronomique pour le Développement, Departement des Cultures Annuelles, BP 5035,34032 Mont-
pellier, France.
245
Arid Soi1 Research and Rebabilitation. 11:245-254, 1997
Copyright 0 1997 Taylor L Francis
0890-3049/97 $12.00 + .oo
-.-,-l--^--l-l-l-
---
.- .-
-------
c
246
M. Ndiaye and F. Ganry
The establishment of sustainable agroforestry practices requires the understanding of sev-
eral factors. First, the feasibility and the modes of implementation of these pmctices
should be considered in Iight of the variability of the environment (soil, climate, and
human pressure on land) and the diversity of farming practices (farmers’ objectives,
socioeconomic means and constraints). Then, atmospheric nitrogen fixation (N2 fixation)
and competition problems with associated crops should be addressed. The importance of
N, fixation is well recognized (Bowen et al., 1990). However, few tests have been con-
ducted to evaluate the N, fixation in natural ecosystems (Hansen & Paste, 1987; Dome-
nach et al., 1989; Yoneyama et al., 199Oa, 1990b; Sanginga et al., 1990, 1991; Mariotti et
al., 1992). Many N,-fixing trees (NFTs) have a potential use in agroforestry systems such
as alley cropping. However, few have been evaluated (Yoneyama, 1987; Kang et al.,
1990). Tree species such as Leucaena spp. and Gliricidia spp., which have been inten-
sively investigated in the tropical zone of Nigeria, have shown high potential for N, fixa-
tion (Kang et al., 1990). Nevertheless, such leguminous species are difficult to establlish in
semiarid and arid zones for several reasons. Among them, besides socioeconomic reasons,
one reason is the establishment and the difficult functioning of N, fixation of these species.
Thus, the evaluation of the percentage of N, fixation (%Ndfa) in NFTs under natural sys-
tems is a primary requirement for the improvement of agroforest.ry techniques. It is indeed
by improving the Ndfa, when it is below the N, fixation potential (NI?), that a sustainable
agroforestry system may be established. The present study is devoted to the evaluation of
the N, fixation by NFTs growing under natural conditions.
Materials and Methods
Detennination of the Percentage of Fixed Nz (%Ndfa) by the Nz-Fixing Trees (NFTs)
Three stations at the Institut Senegalais de Recherches Agricoles (ISRA), namely, Bam-
bey (17”50’ N, 14”75 w) in the northem Sudanian zone, Nioro (16”35’ N, 13”75’ W) in
the southem Sudanian zone, and Djibelor (17’35’ N, 12”50’ W) in the sub-Guinean zone,
representing three ecological zones, were selected. The definition of these zones was based
on isohyets representing the rainfall probability of 5 years out of 10 in the 1970-1990 peri-
od. The isohyets were 400,600, and 1100 mm, respectively, at Bambey, Nioro, and Dji-
belor. Soils were mainly Ultisols and Alfisols (Soi1 Survey Staff, 1992; Singer & Munns,
1996), with a net predominance of kaolinite in the clayey fraction (Charreau & Nicou,
1971). The clay content ranged from 4 to 12% from Bambey to Djibelor. These soils pre-
sented mediocre physical properties, namely, weakly developed structure, very weak
structural stability, and very poor water-holding capacity. The available water reseye
ranged from 3 to 10% in the topsoil, and available P was less than 10 mg kg-‘.
Leaf samples of NETs and non-N,-fixing trees were harvested in May 1994 from t r-
mina1 parts of tree branchds in hedgerows of agroforestry trials. Working in fertilized A a-
cia, Zakra et al. (1991, 1*3) have found that a substantial amount of N in leaves may be
from remobilization from other plant parts. For estimating the N, fixation by Alnus gl t,i-
nosa (L.) Gaertn. Using a ‘% labeling method, Domenach and Kurdali (1989)
l
reco I-
mended sampling for analysis from the most recently formed leaves at the end of the
ing season to avoid the inffluence of nitrogenous reserves of other plant parts. It is
that the translocated N frdm mot and trunk reserves cari represent 10% of total N
mergues, 1997). TO minirfiize the inlfluence of remobilized N from mot and trunk o
newly formed leaves were sampled. Four to 5 trees pcr species or line within species w
N2 Fixation by Tree Légumes in Senegal
247
chosen at random and 4-l 0 branches per tree were sampled to have 4-5 fresh matter sam-
ples weighing 500 g each. This sampling was undertaken at the cutting or pruning period
of the trees. After harvest, each sample was oven-dried at 80°C and then finely ground to
obtain 10 g. Total N content and natural 15N abundance of samples were measured by a
double-introduction mass spectrometer at the Laboratoire Central du Centre National de la
Recherche Scientifique de Lyon, France.
Soi1 samples were taken from topsoil (O-20 cm) at the three sites in NFfs growing
area (three replicates) and outside of this area (three replicates). The total N content and
natural ‘$ abundance of these soi1 samples were determined using the mass spectrometer
described earlier.
The natural 15N abundance (61sN: relative isotopic excess per thousand) of the sam-
ple was expressed as follows:
‘“N/‘~[sample] - l%/lsN[air]
615N = 1000
‘$J/“N[air]
where “?@N[air] = 0.3663, and 615N is expressed in parts per thousand.
The 15N isotopic method for quantifying the N, fixation is based on the S15N varia-
tion. The S’$I values in the N source and reactions by which plants metabolize N coming
from this source define the average 6’?I values in the plant. Plants depending on soi1
exhibit positive 6r5N values, rather close to those of the soi1 (Domenach et al., 1989). Con-
sequently, b15N values of plants cultivated on soi1 where minera1 N is available are lower
if these plants fix N, (Amarger et al., 1979). Thus, based on reference plants (non-N,-
fixing), which depend on soi1 N, the percentage of fixed N, in plants (%Ndfa) cari be cal-
culated using the following formula (Amarger et al., 1979):
&Jref - 6?Nfix
%Ndfa = 100 &ref - 615Np
where 6l$ref, 6t5Nfix, and 6’“Np are, respectively, S’% values of the reference plant, the
N,-fixing plant, and the Nz-fixing plant growing on an inorganic N-free medium. Refer-
ence plants were chosen close to NFfs in such a way that they feed on the same soi1 N
pool-
Determination of Sl”Np
Gliricidia sepium (Jack.) Kunth ex Walp. seeds were scarified in concentrated sulfuric
acld for 15 min and rinsed several times with sterilized distilled water to remove acid.
Seeds were then sown in plastic bags (3 seeds per bag, 4 bags), each filled with 650 g
sterilized Sand. Seedlings were thinned to 1 plant per bag 15 days after emergence and
then inoculated with liquid inoculum of Bradyrhizobium TAL5 isolated on Gliricidia
spp. in Nicaragua (provided by the Centre de Coopération International en Recherche
Agronomique pour le DeveloppementICultures Annuelles, Montpellier, IFrance). The
TAL5 strain was cultured in liquid yeast extract mannitol broth (Vincent, 1970) at 28°C
for 7 days. At thinning, immediately after strain cultivation, 1 ml of suspension contain-
ing approximately 109 cells was added to the soi1 near the seedlings. Bags were watered
daily with an inorganic N-free medium (Vincent, 1970) to keep the moisture content near
field capacity. At 100 days after planting, shoots were tut at soi1 level, mots were washed,
248
M. Ndiaye and F. Ganry
and nodules were collected. Each lant part, oven-dried at 70°C was ground and analyzed
for nitrogen content and natural 15N abundance.
Statistical Analysis
The data were analyzed statistically using a statistical package (STAT-ITCF) for ,the PC
computer (Beaux et al., 1988). Mean and standard deviation from mean were found for
each species or line.
Results
N Content and s”N of Soils
The resulting N contents and 6i5N values of the soils are shown in Table 1. The variations
of i5N abundance of soils from areas with and without growing trees were similar, allow-
ing the use of average values from six replicates. The N content values of soi1 were 0.280,
0.333, and 0.720 g kg-‘, respectively, at the Bambey, Nioro, and Djibelor sites. The 615N
values were +6.58 at Bambey, +6.03 at Nioro, and +6.85 at Djibelor. These 6i5N values
of soils are within the range observed in Brazil and Philippines soils, where a range of +5
to +1 1 was reported (Yoneyama, 1987). A similar range (+3.7 to +9.5) has been reported
for Thailand soils (Yoneyama et al., 1993). The range (+7.7 to +8.2) observed by Snoeck
(1995) for Burundi soils was, however, higher than that reported in the present study.
S’“Np of N,-Fixing Plants Growing on an Inorganic N-Free Medium
The 615N in leaves varied from -1.20 to -2.21 with an average value of -1.70 :+ 0.32
(Table 2). Similar values have been reported by Domenach et al. (1989) in Prosopis
grandulosa TOIT. (-1.70) and by She:arer et al. (1983) in Alnus spp. (-1.80) but a superior
value was found by Mariotti et al. (1992) in Casuarina equisotifolia (L.) (-1 .OO). For this
study, the 6i5Np value of -1.70 f 01.32 has been chosen for estimating the Ndfa of tree
species at the three sites.
S’“N of Reference Plants
Reference plants included the nonllegumes Azadirachta indien JU~S. and Guiera sene-
galensis J.F. Gmel., and non-N,-fixing legumes belonging to the Caesalpinoidose: Cassis
Table 1
Total N contient (g kg-‘) and 615N (parts per thousand) of soils
Soils
N Cg kg-‘)
S15N
Bambey
0.280 t 0.014
+6.58 2 0.10
Nioro
0.333 t 0.052
+6.03 + 0.24
Djibélor
0.720 2 0.034
+6.85 5 0.10
Note. Soi1 sampling horizon = O-20 cm. Each value is the mean of six
replicates. Numbers following ? are standard deviation from mean.
N2 Fixation by Tree Lqumes in Senegal
249
Table 2
Values of 615Np of Gliricidia sepium inoculated with the
Bradyrhizobium strain TALS and grown on an inorganic
N-free medium
Leaf samples
6r5Np
1
-1.20
2
-1.74
3
-1.76
4
-2.21
Mean
-1.70
Standard deviation from mean
50.32
siamea, (Lam.) syn. Sennu siamea (Lam.) H.S. Hirwin & Barneby, Cassia sclerospermu
A. Cunn ex Vog, Bauhinia rufescens (Lam.), and Piliostigma reticuhtum (DC.) Hochst
(Allen & Allen, 1981; Sprent & Sutherland, 1990). The SlsN values of reference plants
(615N ref) were analyzed (Table 3). From the results, the average of the two reference
plants that exhibited the closest values of SlsN to that of the soi1 was used. Thus, the
615Nref values retained were +5.49 at Bambey, +5.57 at Nioro, and +6.89 at. Djibelor. This
6?lref value was higher in the south (humid zone) than in the north (dry zone). Con-
ceming A. indica, the 615Nref varied from +4.95 at Bambey to +8.22 at Djibelor, and took
a intermediary value of +6.16 at Nioro.
Estimation of the Percentage of Fixed N2 (%Ndfa)
The %Ndfa values of leaves harvested from trees growing in different zones of Senegal
are indicated in Table 3. Among NFTs analyzed at Bambey, Gliricidia sepium (Jacq.)
Kunth ex Walp. showed a 615N value (+4.29) close to the 615Nref (+5.49), indicating a low
level of N, fixation (Ndfa = 17%). In contras& the values of Hardwickia binuta Roxb.
(+3X) and Prosopis cineraria (L.) Druce (+3.97) were low, suggesting respective N, fix-
ations (Ndfa) of 22 and 21%. However, G. sepium showed the highest N content among
the three species while having the lowest %Ndfa.
Percentages of fixed N, (%Ndfa) at Nioro showed that a11 species, except Acacia
holosericea Cunn. ex G. Don. and Prosopis cineraria, had SlsN values similar to the
&lref value considering the experimental error (Table 3). These two species showed
respective Ndfa values of 39 and 21%. As found at Bambey, the highest value of N con-
tent (3.96%) was observed in leaves of G. sepium (ILG 50), which fixed only 6% of its
total N.
At Djibelor, the 615N values for G. sepium lines were significantly lower than the aver-
age S’% ref value (+6.89). Their SlsN values varied from -1.30 for ILG61 to +0.34 for
ILG58. Contrarily to Bambey and Nioro, the contribution of N, fixation (Ndfa) to N nutri-
tion of plants was very high at Djibelor, ranging from 76% for ILG58 to 95% for ILGOl.
Improvement Test of the N,jk.tion of G. sepium at Nioro
The low levels of Ndfa observed at Nioro prompted us to define the best way to improve
the N, fixation. Water and minera1 nutrients such as P are the main limiting factors of the
_-
I-
-mm-.‘.u1--
-
250
M. Ndiaye and F. Guru-y
Table 3
Total N content (CG), 615N I:parts per thousand), and Ndfa (96) of plants
Species
N 6)
615N
Ndfa (%a
Bambey site
Nonnodulating legumes
Cassia siamea
22.2 +- 0.12
4.29 + 0.11
N:D
Cassia sclerosperma varl
1.58 t 0.04
4.84 -t 0.09
ND
Cassia sclerosperma var2
1.58 + 0.05
4.73 2 0.10
ND
Cassia sclerosperma var3
1.62 + 0.03
4.66 2 0.18
NID
Bauhinia rufescens
1.81 + 0.08
6.03 + 0.42
NID
Nonlegumes
Azadirachta indica
1.82 t 0.05
4.95 + 0.25
NID
Mean for reference plants
retained”
1.81
5.49
NID
Nodulating legumes
Gliricidia sepium
3.93 t 0.07
4.29 + 0.27
1’7
Hardwickia binata
1.66 2 0.03
3.90 + 0.18
22
Prosopis cineraria
2.09 ‘-e 0.05
3.97 + 0.14
21
Mean for nodulating plants
2.56
4.05
20
Nioro site
Nonnodulating legumes
Piliostigma reticulatum
1.49 f. 0.06
4.73 + 0.16
ND
Cassis siamea
2.60 -I 0.09
4.58 + 0.04
ND
Nonlegumes
Azadirachta indica
2.99 + 0.11
6.16 t 0.59
ND
Guiera senegalensis
1.29 + 0.04
4.99 -t- 0.49
ND
Mean for reference plants
retainedb
2.34
5.57
ND
Nodulating legumes
Acacia holosericea
1.52 2 0.03
2.74 2 0.09
39
Prosopis cineraria
‘2.10 2 0.06
4.01 2 0.10
211
Gliricidia sepium ILGSO
3.% +r 0.06
5.11 t 0.09
6
Gliricidia sepium ILG.5.5
3.65 + 0.12
5.39 + 0.04
2
Gliricidia sepium HYB
3.68 + 0.13
5.48 t 0.16
11
Hardwickia binata
1.46 k 0.05
5.66 2 0.04
0
Leucaena leucocephalq
.3.82 I!I 0.10
5.69 + 0.12
0
Meau for nodulating plan+
2.89
4.87
10
Djibélor site
Cassia siamea
I
:3.22 2 0.12
+5.57 +I 0.10
ND
Azadirachta indica I
:2.45 k 0.09
+8.22 t 0.43
ND
Mean for reference plants!
retainedc
2.83
+6.89
ND
G. sepium ILG 50
4.90 2 0.15
-0.55 2 0.11
87
G. sepium ILG 52
I
5.06 2 0.19
-0.08 i 0.23
81
G. sepium ILG 54
4.69 + 0.13
+0.27 t 0.29
77
N, Fixation by Tree L.egumes in Senegal
251
Table 3 (Continued)
Total N content (%), S15N (parts per thousand), and Ndfa (%) of plants
Species
N @)
FN
Ndfa (%)
G. sepium ILG 55
5.45 t 0.20
-0.55 + 0.11
87
G. sepium ILG 56
4.99 k 0.22
-0.75 -+ 0.22
89
G. sepium ILG 57
4.68 2 0.11
-0.54 + 0.09
86
G. sepium ILG 58
4.61 2 0.12
+0.34 2 0.01
76
G. sepium ILG 59
4.84 2 0.28
-0.34 2 0.12
84
G. sepium ILG 60
4.94 i 0.09
-0.41 rt 0.26
85
G. sepium ILG 61
4.84 -c 0.13
-1.30 + 0.12
95
G. sepium ILG 62
4.34 2 0.08
+0.27 t 0.14
77
G. sepium ILG 63
4.40 k 0.09
-0.07 rt 0.07
81
G. sepium HYB
4.44 + 0.12
+o.oo + 0.09
80
Mean for nodulating plants
4.78
-0.20
83
Note. Each value is the mean of four replicates. Numbers following + are standard deviation from
mean. ND, not determined.
“Mean of G”Nref of B. rufescens and A. indica.
%Jean of 6”Nref of G. senegalensis and A. indica.
‘Mean of 615Nref of C. siamea and A. indica.
N, fixation in this zone. This was why we have cultivated G. sepium in plastic bags under
greenhouse conditions by providing water and mineral nutrients (except N). Two treat-
ments-Tl (G. sepium growing on Nioro soi1 where tested trees have been planted) and
T2 (G. sepium inoculated with a Bradyrhizobium strain TAL5 and growing on Nioro
soil)-were compared in a completely randomized design with four replicates. The results
obtained after three months of cultivation (Table 4) showed a higher 615N value for G.
sepium than those obtained in situ (Table 3). It cari be deduced from these test results that
water and mineral elements (N excluded) were not the limiting factors of N, fixation at
Nioro. The major limiting factor was probably a deficiency of the symbiosis due to
absence of specific native strains of Bradyrhizobium or to other constraints different from
lack of water or minera1 elements. Inoculation with an effective strain of Bradyrhizobium
Table 4
Inoculation effect of the Bradyrhizobium TAL5 strain on dry matter (g plant-‘)
and S”N (parts per thousand) of Gliricidia sepium
Dry weight (g plant-‘)
S15N
Treatment
Stems + leaves
Nodules
Stems + leaves Nodules
Tl: G. sepium
0.53 2 0.08
ni1
9.09 t, 2.59
ni1
+Nioro soi1
T2: G. sepium
4.87 t 0.96
0.61 2 0.24
-1.99 It 0.08
12.69
+ Nioro soi1
+ TAL5 inoculum
Note. E&h value is the mean of four replicates. Numbers following i are standard
deviation from mean.
-
- - - I - - I
252
M. Ndiaye and F. Ganry
such as TALS, which should be tested under field conditions, is a possible improvement
technique of the N, fixation (Ndfa).
Discussion and Conclusion
The validity of the estimation of the N, fixation based on the 15N technique lies in the
choice of the reference plant, which is generally considered the main source of error (Fried
et al., 1983). Its choice is particularly difficult for NFTs for which nonnodulating lines are
not yet available. Cassia siamea (Lam.) and Eucalyptus grandis Hi11 ex Maiden have
already been used as reference plants for estimating the N, fixation of Leucaenu leuco-
cephala (Lam.) de Wit. and Acacia albidu (L.) syn. Fuidherbia albidu (Del.) A. Chev.
(Sanginga et al., 1990). Another diffïculty of the 615N method is the requirement that ref-
erence plants be very close to NFR TO reduce the possible errors in choosing the refer-
ence plants, Domenach et al. (1989) have proposed an average 615N value of several ref-
erence plants (rather than one) growing in the same environment as that of N~S. In the
present study, the average of the two closest 615Nref values to that of the soi1 was used at
each site. Our results showed that the 615N values of the soi1 were high enough (>4 parts
per thousand) and homogeneous enough within a site to allow the utilization of the 615N
method (Peoples et al., 1988).
At Bambey, the 615Nref (particularly that of C. siumeu) was lower than that of the soil,
but 615Nref values found at Nioro and Djibelor were very close to or identical to th’ose of
soils. These 615Nref increased with an ecological “north-south” gradient (Bambcy, +5.49;
Nioro, +5.57; Djibelor, +6.89).
The percentage of fixed N, was very low at Bambey (Ndfa I 22%) and Nioro (Ndfa
5 39%) and very high at Djibclor (76% I Ndfa 195%).
Based on the results of the experiment on Nioro soi1 under controlled conditions, we
advanced the hypothesis of lack of N, fixation due to the absence of Brudyrhizobium
strains or the inability of these strains to express themselves in the Nioro environment. In
contrast, specific strains of Bradyrhîzobium were present and were able to efficiently fix
N, at Djibelor. Therefore, we obviously showed that the two zones (Sudano-Sahelian and
sub-Guinean) are strongly contrasting in respect to N, fixation. In the Sudano-Sahelian
zone, the main constraints are the low clay content of soils (17%), and low rainfall (~$00
mm of precipitation per year) over a duration not exceeding 4 months, and secondarily
high concentration of minera1 N in the soi1 solution induced by marked drying and rewet-
ting cycles (Blondel, 1971; Sprent, 1976). These environmental conditions are not favor-
able to the establishment and expression of Bradyrhizobium strains because the nodulation
of host plant is insufficient and inadlequately repeated on time for developing and main-
taining specific population of native strains in the soil. However, these strains cari be re-
sent deep in the soi1 and fdrm effective nodules (Felker, 1986; Virginia et al., 1986; Du uy
& Dreyfus, 1992). In contrast, under a more humid climate like that at D.jibelor (21 i00
mm of precipitation per year), abundant nodulation is observed (Dupuy & Dreyfus, 112!),
thus favoring the buildup of efficient Bradyrhizobium strains when nodulations are redat-
ed. The findings of Ducousso et al. (1995) are in agreement with our explanation. Wo/rk-
ing in Acacia albidu, they have shown that most Bradyrhizobium strains isolated from khe
Sahelian zone are ineffective. They further demonstmted that the proliferation of Br ad,y-
rhizobium populations is strongly dependent on the durable presence of the host tree. $he
low values of %Ndfa found in the Egambey and Nioro sites showed that NFTs eval
ed
in these zones were more dependent on soi1 N than on the N,
i
fixation, because of m jor
constraints affecting N, fixation. To fix N,, Ieguminous trees need at least 700 mm (of
rainfal 1.
~
N2 Fixation by Tree Legumes in Senegal
253
The #“N values observed in C. siamea (+4.29 at Bambey, +4..58 at Nioro, and +5.57
at Djibelor) were very close to those of the S15N of NFTs reputed to fix N, (Allen & Allen,
1981). Nevertheless, the hypothesis of N, fixation by C. siamea cannot be confirmed due
to a number of reasons, including morphological factors (immobilization of Rhizobium at
the absorbent root hair base, reduced cortex) and biochemical factors (production of
antibacterial compounds, phenolic compounds, tannins and quinones in root cells) that
militate against the nodule formation on C. siamea roots (Allen & Allen, 198 1). Based on
our results, no valuable hypothesis cari be formulated to explain the low 615N values in C.
siamea. This induces interesting fields of investigation in order to find causes and deter-
mining factors of this phenomenon.
References
Allen, E. K., and 0. Allen. 1981. The nodulation profile of the genus Cassis, pp. 113-122, in P. S.
Nutman, ed., Symbiotic nitrogen jxution in plants. International Biology Programme no. 7.
Cambridge University Press, Cambridge.
Amarger, N., A. Mariotti, F. Mariotti, J. C. Durr, C. Bourguignon, and B. Lagacherie. 1979. Esti-
mate of symbiotically fixed nitrogen in field grown soybean using variations in % natural
abundance. Plant and Soi1 523269-280.
Beaux, M. F., H. Gouet, J. P. Gouet, P. Morleghem, G. Philippeau, J. Tranchefort, and M. Vemeau.
1988. Manuel d’utilisation STAT-ITCF Version N”4. Institut Techniques des Ceréales et des
Fourrages (ITCF), Services des Etudes Statistiques, Paris.
Blonde], D. 1971. Contribution à la connaissance de la dynamique de l’azote minéral en sol sableux
(Dior) au Sénégal. Agronomie Tropicale 26~1303-1333.
Bowen, G. D., N. Sanginga, and S.K.A. Danso. 1990. Biological nitrogen fixation in agroforestry-
An overview, pp. 170-175, in Proceedings of 14th International Congress of Soi1 Science, Vol-
ume III. International Institute of Agroenvironmental Sciences, Kyoto, Japan.
Charreau, C., and R. Nicou. 1971. L’amélioration du profil cultural dans les sols sableux et sablo-
argileux de la zone tropicale sèche ouest-africaine et ses incidences agronomiques. Bulletin
Agronomique No. 23. IRAT, Paris.
Domenach, A. M., and F. Kurdali. 1989. Influence des réserves azotées sur la formation des feuilles
d’dlnus glutinosa et ses conséquences dans l’estimation de la fixation d’azote. Canadian Jour-
nul of Botany 67:865-871.
Domenach, A. M., F. Kurdali, and R. Bardin. 1989. Estimation of symbiotic dinitrogen fixation in
alder forest by the method based on natural % abundance. Plant and Soi1 11851-59.
Dommergues, Y. 1997. Les arbresfiateurs d’azote dans l’aménagement des écosystèmes méditer-
ranéens et tropicaux. (in press).
Ducousso, M., B. Sougoufara, N. Dupuy, and B. Dreyfus. 1995. Rôle de deux acacias (Acacia
mungium et Acacia albida) dans le maintien et l’amélioration de la fertilité des sols en Basse
Casamance. Rapport final du projet Agroforesterie III No. 91 L 0685. Ministère de I’Enseigne-
ment Supérieur et de la Recherche, Paris.
Dupuy, N. C., and B. L. Dreyfus. 1992. Brudyrhizobrium populations occur in deep soi1 under legu-
minous tree Acacia albida. Applied ami Environmental Microbiology 582415-2419.
Felker, P. 1986. Treephzntings in semi-arid regions. Elsevier, New York.
Fried, M., S.K.A. Danso, and F. Zapata. 1983. The methodology of measurement of N, fixation by
non-legumes as referred from field experiments with legumes. Canadian Journal of Microbiol-
ogy 29: 1053-1062.
Hansen, A. P., and J. S. Pate. 1987. Evaluation of the % natural abundance method and xylem sap
analysis for assessing N, fixation of understorey legumes in Jarrah (Eucalyptus arginatu (Donn
ex Sm.)) forest in South West. Australian Journal of Experimental Botany 38: 1446-1458.
Kang, B. T., L. Reynolds, and A. N. Atta-Krah. 1990. Alley farming. Advances in Agronomy
43:315-357.
Mariotti, A., B. Sougoufara, and Y. R. Dommergues. 1992. Estimation of nitrogen fixation using the
254
M’. Ndiaye and F. Ganty
natural abundance method in a plantation of Casuarina equisetifolia (Forst). Soi1 Biolo y and
Biochemistry 24647$63.
Peoples, M. B., D. F. Herrbdge, and J. F. Bergersen. 1988. Measurement of nitrogen fixation in trop
and shrub legumes, pp. 223-238, in Green manure in rice farming. Proceedings of a s:nnpo-
sium on sustainable agriculture, 25-29 May 1987. International Rice Research Institute, Los
Banos, the Philippines.
Sanginga, N., S. K. A. Danso, F. Zapata, and G. D. Bowen. 1990. Influence of reference trees on N,
fixation estimates in keucaena leucocephala and Acacia albida using %-labeling techn clues.
Biology and Fertility of Soi1 934, L-346.
Sanginga, N., G. D. Bowen, and S. K. A. Danso. 1991. Intra-specific variation in P accumulat on of
Leucaena leucoceph.ala and Gliricidia sepium as influenced by soi1 phosphate status. Plant ami
Soi1 133:201-208.
Shearer, G., D. H. Khol, R. A. Virginia, B. A. Bryan, J. L. Skeeters, E. T. Nilsen, M. R. Shari i., and
P. W. Rundel. 1983. Estimates of N, fixation from variation in the natural abundance of i5N in
Sonoran Desert ecosystem. Oecologia (Berlin) 56365-373.
Singer, M. J., and D. N. Munns. 1996. Soils. An introduction. Prentice Hall, Upper Saddle River,
New Jersey.
Snoeck, D. 1995. Interactions entre végétaux fixateurs d’azote et non fixateurs en culture mixte: Cas
des Leucaena spp.associéees à Coffea arabica L. au Burundi. Thèse de Doctorat. lJni@sité
Claude Bernard-Lyon 1, Lyon, France.
Soi1 Survey Staff. 1992. Keys to soi1 taxonomy, Monograph No. 19.5th ed. Agency for Inte
tion-
a1 Development, U.S. Department of Agriculture. Pocahontas Press, Blacksburg, Virgini
Sprent, J. 1976. Water deficit and nitrogen fixing root nodules, pp. 291-315, in T. T. Kozlows1i ed.,
Water-dejcit andplant growth. Academic Press, New York.
Sprent, J. J., and Sutherland, J. M. 1!390. Nitrogen fixing woody legumes. Nitrogen Ftiing 1Irees
i -
Research Reports 8: 17-3 1.
Vincent, J. M. 1970. A manualfor thepractical study of the root nodule bacteria. International Biol-
ogy Programme Handbook No. 15. Blackwell Scientific Publications, Oxford, UK.
Virginia, R. A., M. B. Jenkins, and W. M. Jarrell. 1986. Depth of mot symbiont occurrence in soil.
BioZogy and Fertility of Soi1 2: 127-l 30.
Yoneyama, T. 1987. N, fixation and natural abundance of leguminous plants and Azolla. Bulletin of
National Institute of Agrobiologiscal Resources (Japan) 3z59-87.
Yoneyama, T., K. Kouno, and J. Yazaki. 1990a. Variation of natural % abundance of cro ds and
soils in Japan with special reference to the effect of soi1 conditions and fertilizer applioation.
Soi1 Science and Plant Nutrition 36667-675.
Yoneyama, T., T. Mukarami, N. Boonkerd, P. Wadisirisuk, S. Siripin, and K. Kouno. 1990~. Nat-
ural ‘-% abundance in shrub and tree legumes, Casuarina and non fixing plants in Th ‘land.
i
Plant and Soi1 128287-292.
Yoneyama, T., T. Muroka,T. Murakami, and N. Boonkerd. 1993. Natural abundance of “N in trop-
ical plants with emvbasis on tree leaume. Plant and Soi1 153:293-295.
Zakra, N., Ouvrier, M., and G. De Taffin. 1991. Partitioning of 15N labelled mineral nitrogen i4I Aca-
cia and coconut, pp. 79-386, in .Proceedings of an International Symposium on the Use 4
ble Isotopes in Plant Nutrition, Soi1 Fertility and Environmental Studies, jointly organi 21 fZ
the International Ato ic Energy ,4gency (IAEA) and the Food and Agriculture Organizat/ion of
the United Nations ( AO), Vienna, l-5 October 1990. IAEA, Vienna.
Zakra, N. Weaver, R. W., and G. De Taffin. 1993. Distribution of ‘%I from ammonium fertil zer in
field grown Acacia. ~lant amISoi 151:147-150.
Legumes in Threé Ecological Zones from the North
to the South of Senegal
<.
,j
’
M., NDIAYE
Jf ? Pj.w~ *lc;q+
Institut Sénégalais de Recherches Agricoles
,
Centre Nord Bassin Arachidier
JJ7/73-
.,
Bambey, Sénégal
,
. . ..b...1,.
I\\
i-4. ‘ii..,,. , <a “.,
&
.-. ,.” -.. . . . . * ._._,-.
F. GANRY
,_
.Iv.&
.w -.- .-4. d ,” / i-&*.,“,
Centre de Coopération Internationale en Recherche Agronomique pour le
Développement
Département des Cultures Annuelles
Montpellier, France
The utilization
NJïxing trees (NFTs) in agroforestry, in semiarid regions, is O&n
limited by weak N, fixation. This study was conducted to evaluate the N&ing capac-
ity of trees growing under natural conditions for 10 years with neither chemical fertil-
izer nor inoculum application. For this purpose, natural isotopic “N abundance (b5N)
of leaves harvested from NFTs and reference plants (zen growing in three agroeco-
logical zones of Senegal were analyzed. At each site (one site per zone), average g5N
ref was calculated N contents and b5N values increasedfiom the north in the Sudano-
Sahelian zone (Bambey and Nioro) ta the south in the sub-Guinean zone (Djibelor).
The percentage ofjixed N, (%Ndfa) was low at both Bambey (Ndfa 5 22%) and Niera
(Ndfa c; 39%) and very high at Djibelor (76% 5 Ndfa s 95%). TO explain the low lev-
els of Ndfa observed at Bambey and Nioro, the hypothesis of lack of Bradyrhizobium
strains in these soils was made. This hypothesis was confirmed by results obtainedfrom
a tria1 conducted under controlled conditions, which showed absence of nodulation on
Gliricidia sepium growing under optimal conditions of water and nutrient supp[y.
Improvement of NZ fiation of NFTs in the semiarid zone is a prerequisite for imple-
mentation of sustainable agroforestry systems.
Keywords N,-fixing trees, natural “N abundance, percent of fixed N:!, Sudano-
Sahelian and sub-Guinean zones
Received 20 August 1996; accepted 12 December 1996.
We thank P. N. Sall, S. Badiane, S. A. Ndiaye, B. Ndour, and 1. Diaïté (researchers at the Insti-
tut Sénégalais de Recherches Agricoles, ISRA) for allowing us to sample tree legumes, and 1. Dioum
and G. Deme (technicians at ISRA) for their help in sampling trees and soils. We are very grateful
for the advice in tree sampling received from Dr. A. M. Domenach. We also thank R. Oliver for his
assistance in analyzing natural ‘% abundance. Professor Yvon R. Dommergues is acknowledged for
his valuable comments on earlier versions of the manuscript. The comments of anonymous review-
ers are much appreciated.
Address correspondence to Dr. F. Gam-y, Centre de Coopération Internationale en Recherche
Agronomique pour le Développement, Departement des Cultures Annuelles, BP 5035,34032 Mont-
pellier, France.
245
Arid Soi1 Research and Rebabilitation. 11:245-254, 1997
Copyright 0 1997 Taylor L Francis
0890-3049/97 $12.00 + .oo
-.-,-l--^--l-l-l-
---
.- .-
-------
c
246
M. Ndiaye and F. Ganry
The establishment of sustainable agroforestry practices requires the understanding of sev-
eral factors. First, the feasibility and the modes of implementation of these pmctices
should be considered in Iight of the variability of the environment (soil, climate, and
human pressure on land) and the diversity of farming practices (farmers’ objectives,
socioeconomic means and constraints). Then, atmospheric nitrogen fixation (N2 fixation)
and competition problems with associated crops should be addressed. The importance of
N, fixation is well recognized (Bowen et al., 1990). However, few tests have been con-
ducted to evaluate the N, fixation in natural ecosystems (Hansen & Paste, 1987; Dome-
nach et al., 1989; Yoneyama et al., 199Oa, 1990b; Sanginga et al., 1990, 1991; Mariotti et
al., 1992). Many N,-fixing trees (NFTs) have a potential use in agroforestry systems such
as alley cropping. However, few have been evaluated (Yoneyama, 1987; Kang et al.,
1990). Tree species such as Leucaena spp. and Gliricidia spp., which have been inten-
sively investigated in the tropical zone of Nigeria, have shown high potential for N, fixa-
tion (Kang et al., 1990). Nevertheless, such leguminous species are difficult to establlish in
semiarid and arid zones for several reasons. Among them, besides socioeconomic reasons,
one reason is the establishment and the difficult functioning of N, fixation of these species.
Thus, the evaluation of the percentage of N, fixation (%Ndfa) in NFTs under natural sys-
tems is a primary requirement for the improvement of agroforest.ry techniques. It is indeed
by improving the Ndfa, when it is below the N, fixation potential (NI?), that a sustainable
agroforestry system may be established. The present study is devoted to the evaluation of
the N, fixation by NFTs growing under natural conditions.
Materials and Methods
Detennination of the Percentage of Fixed Nz (%Ndfa) by the Nz-Fixing Trees (NFTs)
Three stations at the Institut Senegalais de Recherches Agricoles (ISRA), namely, Bam-
bey (17”50’ N, 14”75 w) in the northem Sudanian zone, Nioro (16”35’ N, 13”75’ W) in
the southem Sudanian zone, and Djibelor (17’35’ N, 12”50’ W) in the sub-Guinean zone,
representing three ecological zones, were selected. The definition of these zones was based
on isohyets representing the rainfall probability of 5 years out of 10 in the 1970-1990 peri-
od. The isohyets were 400,600, and 1100 mm, respectively, at Bambey, Nioro, and Dji-
belor. Soils were mainly Ultisols and Alfisols (Soi1 Survey Staff, 1992; Singer & Munns,
1996), with a net predominance of kaolinite in the clayey fraction (Charreau & Nicou,
1971). The clay content ranged from 4 to 12% from Bambey to Djibelor. These soils pre-
sented mediocre physical properties, namely, weakly developed structure, very weak
structural stability, and very poor water-holding capacity. The available water reseye
ranged from 3 to 10% in the topsoil, and available P was less than 10 mg kg-‘.
Leaf samples of NETs and non-N,-fixing trees were harvested in May 1994 from t r-
mina1 parts of tree branchds in hedgerows of agroforestry trials. Working in fertilized A a-
cia, Zakra et al. (1991, 1*3) have found that a substantial amount of N in leaves may be
from remobilization from other plant parts. For estimating the N, fixation by Alnus gl t,i-
nosa (L.) Gaertn. Using a ‘% labeling method, Domenach and Kurdali (1989)
l
reco I-
mended sampling for analysis from the most recently formed leaves at the end of the
ing season to avoid the inffluence of nitrogenous reserves of other plant parts. It is
that the translocated N frdm mot and trunk reserves cari represent 10% of total N
mergues, 1997). TO minirfiize the inlfluence of remobilized N from mot and trunk o
newly formed leaves were sampled. Four to 5 trees pcr species or line within species w
N2 Fixation by Tree Légumes in Senegal
247
chosen at random and 4-l 0 branches per tree were sampled to have 4-5 fresh matter sam-
ples weighing 500 g each. This sampling was undertaken at the cutting or pruning period
of the trees. After harvest, each sample was oven-dried at 80°C and then finely ground to
obtain 10 g. Total N content and natural 15N abundance of samples were measured by a
double-introduction mass spectrometer at the Laboratoire Central du Centre National de la
Recherche Scientifique de Lyon, France.
Soi1 samples were taken from topsoil (O-20 cm) at the three sites in NFfs growing
area (three replicates) and outside of this area (three replicates). The total N content and
natural ‘$ abundance of these soi1 samples were determined using the mass spectrometer
described earlier.
The natural 15N abundance (61sN: relative isotopic excess per thousand) of the sam-
ple was expressed as follows:
‘“N/‘~[sample] - l%/lsN[air]
615N = 1000
‘$J/“N[air]
where “?@N[air] = 0.3663, and 615N is expressed in parts per thousand.
The 15N isotopic method for quantifying the N, fixation is based on the S15N varia-
tion. The S’$I values in the N source and reactions by which plants metabolize N coming
from this source define the average 6’?I values in the plant. Plants depending on soi1
exhibit positive 6r5N values, rather close to those of the soi1 (Domenach et al., 1989). Con-
sequently, b15N values of plants cultivated on soi1 where minera1 N is available are lower
if these plants fix N, (Amarger et al., 1979). Thus, based on reference plants (non-N,-
fixing), which depend on soi1 N, the percentage of fixed N, in plants (%Ndfa) cari be cal-
culated using the following formula (Amarger et al., 1979):
&Jref - 6?Nfix
%Ndfa = 100 &ref - 615Np
where 6l$ref, 6t5Nfix, and 6’“Np are, respectively, S’% values of the reference plant, the
N,-fixing plant, and the Nz-fixing plant growing on an inorganic N-free medium. Refer-
ence plants were chosen close to NFfs in such a way that they feed on the same soi1 N
pool-
Determination of Sl”Np
Gliricidia sepium (Jack.) Kunth ex Walp. seeds were scarified in concentrated sulfuric
acld for 15 min and rinsed several times with sterilized distilled water to remove acid.
Seeds were then sown in plastic bags (3 seeds per bag, 4 bags), each filled with 650 g
sterilized Sand. Seedlings were thinned to 1 plant per bag 15 days after emergence and
then inoculated with liquid inoculum of Bradyrhizobium TAL5 isolated on Gliricidia
spp. in Nicaragua (provided by the Centre de Coopération International en Recherche
Agronomique pour le DeveloppementICultures Annuelles, Montpellier, IFrance). The
TAL5 strain was cultured in liquid yeast extract mannitol broth (Vincent, 1970) at 28°C
for 7 days. At thinning, immediately after strain cultivation, 1 ml of suspension contain-
ing approximately 109 cells was added to the soi1 near the seedlings. Bags were watered
daily with an inorganic N-free medium (Vincent, 1970) to keep the moisture content near
field capacity. At 100 days after planting, shoots were tut at soi1 level, mots were washed,
248
M. Ndiaye and F. Ganry
and nodules were collected. Each lant part, oven-dried at 70°C was ground and analyzed
for nitrogen content and natural 15N abundance.
Statistical Analysis
The data were analyzed statistically using a statistical package (STAT-ITCF) for ,the PC
computer (Beaux et al., 1988). Mean and standard deviation from mean were found for
each species or line.
Results
N Content and s”N of Soils
The resulting N contents and 6i5N values of the soils are shown in Table 1. The variations
of i5N abundance of soils from areas with and without growing trees were similar, allow-
ing the use of average values from six replicates. The N content values of soi1 were 0.280,
0.333, and 0.720 g kg-‘, respectively, at the Bambey, Nioro, and Djibelor sites. The 615N
values were +6.58 at Bambey, +6.03 at Nioro, and +6.85 at Djibelor. These 6i5N values
of soils are within the range observed in Brazil and Philippines soils, where a range of +5
to +1 1 was reported (Yoneyama, 1987). A similar range (+3.7 to +9.5) has been reported
for Thailand soils (Yoneyama et al., 1993). The range (+7.7 to +8.2) observed by Snoeck
(1995) for Burundi soils was, however, higher than that reported in the present study.
S’“Np of N,-Fixing Plants Growing on an Inorganic N-Free Medium
The 615N in leaves varied from -1.20 to -2.21 with an average value of -1.70 :+ 0.32
(Table 2). Similar values have been reported by Domenach et al. (1989) in Prosopis
grandulosa TOIT. (-1.70) and by She:arer et al. (1983) in Alnus spp. (-1.80) but a superior
value was found by Mariotti et al. (1992) in Casuarina equisotifolia (L.) (-1 .OO). For this
study, the 6i5Np value of -1.70 f 01.32 has been chosen for estimating the Ndfa of tree
species at the three sites.
S’“N of Reference Plants
Reference plants included the nonllegumes Azadirachta indien JU~S. and Guiera sene-
galensis J.F. Gmel., and non-N,-fixing legumes belonging to the Caesalpinoidose: Cassis
Table 1
Total N contient (g kg-‘) and 615N (parts per thousand) of soils
Soils
N Cg kg-‘)
S15N
Bambey
0.280 t 0.014
+6.58 2 0.10
Nioro
0.333 t 0.052
+6.03 + 0.24
Djibélor
0.720 2 0.034
+6.85 5 0.10
Note. Soi1 sampling horizon = O-20 cm. Each value is the mean of six
replicates. Numbers following ? are standard deviation from mean.
N2 Fixation by Tree Lqumes in Senegal
249
Table 2
Values of 615Np of Gliricidia sepium inoculated with the
Bradyrhizobium strain TALS and grown on an inorganic
N-free medium
Leaf samples
6r5Np
1
-1.20
2
-1.74
3
-1.76
4
-2.21
Mean
-1.70
Standard deviation from mean
50.32
siamea, (Lam.) syn. Sennu siamea (Lam.) H.S. Hirwin & Barneby, Cassia sclerospermu
A. Cunn ex Vog, Bauhinia rufescens (Lam.), and Piliostigma reticuhtum (DC.) Hochst
(Allen & Allen, 1981; Sprent & Sutherland, 1990). The SlsN values of reference plants
(615N ref) were analyzed (Table 3). From the results, the average of the two reference
plants that exhibited the closest values of SlsN to that of the soi1 was used. Thus, the
615Nref values retained were +5.49 at Bambey, +5.57 at Nioro, and +6.89 at. Djibelor. This
6?lref value was higher in the south (humid zone) than in the north (dry zone). Con-
ceming A. indica, the 615Nref varied from +4.95 at Bambey to +8.22 at Djibelor, and took
a intermediary value of +6.16 at Nioro.
Estimation of the Percentage of Fixed N2 (%Ndfa)
The %Ndfa values of leaves harvested from trees growing in different zones of Senegal
are indicated in Table 3. Among NFTs analyzed at Bambey, Gliricidia sepium (Jacq.)
Kunth ex Walp. showed a 615N value (+4.29) close to the 615Nref (+5.49), indicating a low
level of N, fixation (Ndfa = 17%). In contras& the values of Hardwickia binuta Roxb.
(+3X) and Prosopis cineraria (L.) Druce (+3.97) were low, suggesting respective N, fix-
ations (Ndfa) of 22 and 21%. However, G. sepium showed the highest N content among
the three species while having the lowest %Ndfa.
Percentages of fixed N, (%Ndfa) at Nioro showed that a11 species, except Acacia
holosericea Cunn. ex G. Don. and Prosopis cineraria, had SlsN values similar to the
&lref value considering the experimental error (Table 3). These two species showed
respective Ndfa values of 39 and 21%. As found at Bambey, the highest value of N con-
tent (3.96%) was observed in leaves of G. sepium (ILG 50), which fixed only 6% of its
total N.
At Djibelor, the 615N values for G. sepium lines were significantly lower than the aver-
age S’% ref value (+6.89). Their SlsN values varied from -1.30 for ILG61 to +0.34 for
ILG58. Contrarily to Bambey and Nioro, the contribution of N, fixation (Ndfa) to N nutri-
tion of plants was very high at Djibelor, ranging from 76% for ILG58 to 95% for ILGOl.
Improvement Test of the N,jk.tion of G. sepium at Nioro
The low levels of Ndfa observed at Nioro prompted us to define the best way to improve
the N, fixation. Water and minera1 nutrients such as P are the main limiting factors of the
_-
I-
-mm-.‘.u1--
-
250
M. Ndiaye and F. Guru-y
Table 3
Total N content (CG), 615N I:parts per thousand), and Ndfa (96) of plants
Species
N 6)
615N
Ndfa (%a
Bambey site
Nonnodulating legumes
Cassia siamea
22.2 +- 0.12
4.29 + 0.11
N:D
Cassia sclerosperma varl
1.58 t 0.04
4.84 -t 0.09
ND
Cassia sclerosperma var2
1.58 + 0.05
4.73 2 0.10
ND
Cassia sclerosperma var3
1.62 + 0.03
4.66 2 0.18
NID
Bauhinia rufescens
1.81 + 0.08
6.03 + 0.42
NID
Nonlegumes
Azadirachta indica
1.82 t 0.05
4.95 + 0.25
NID
Mean for reference plants
retained”
1.81
5.49
NID
Nodulating legumes
Gliricidia sepium
3.93 t 0.07
4.29 + 0.27
1’7
Hardwickia binata
1.66 2 0.03
3.90 + 0.18
22
Prosopis cineraria
2.09 ‘-e 0.05
3.97 + 0.14
21
Mean for nodulating plants
2.56
4.05
20
Nioro site
Nonnodulating legumes
Piliostigma reticulatum
1.49 f. 0.06
4.73 + 0.16
ND
Cassis siamea
2.60 -I 0.09
4.58 + 0.04
ND
Nonlegumes
Azadirachta indica
2.99 + 0.11
6.16 t 0.59
ND
Guiera senegalensis
1.29 + 0.04
4.99 -t- 0.49
ND
Mean for reference plants
retainedb
2.34
5.57
ND
Nodulating legumes
Acacia holosericea
1.52 2 0.03
2.74 2 0.09
39
Prosopis cineraria
‘2.10 2 0.06
4.01 2 0.10
211
Gliricidia sepium ILGSO
3.% +r 0.06
5.11 t 0.09
6
Gliricidia sepium ILG.5.5
3.65 + 0.12
5.39 + 0.04
2
Gliricidia sepium HYB
3.68 + 0.13
5.48 t 0.16
11
Hardwickia binata
1.46 k 0.05
5.66 2 0.04
0
Leucaena leucocephalq
.3.82 I!I 0.10
5.69 + 0.12
0
Meau for nodulating plan+
2.89
4.87
10
Djibélor site
Cassia siamea
I
:3.22 2 0.12
+5.57 +I 0.10
ND
Azadirachta indica I
:2.45 k 0.09
+8.22 t 0.43
ND
Mean for reference plants!
retainedc
2.83
+6.89
ND
G. sepium ILG 50
4.90 2 0.15
-0.55 2 0.11
87
G. sepium ILG 52
I
5.06 2 0.19
-0.08 i 0.23
81
G. sepium ILG 54
4.69 + 0.13
+0.27 t 0.29
77
N, Fixation by Tree L.egumes in Senegal
251
Table 3 (Continued)
Total N content (%), S15N (parts per thousand), and Ndfa (%) of plants
Species
N @)
FN
Ndfa (%)
G. sepium ILG 55
5.45 t 0.20
-0.55 + 0.11
87
G. sepium ILG 56
4.99 k 0.22
-0.75 -+ 0.22
89
G. sepium ILG 57
4.68 2 0.11
-0.54 + 0.09
86
G. sepium ILG 58
4.61 2 0.12
+0.34 2 0.01
76
G. sepium ILG 59
4.84 2 0.28
-0.34 2 0.12
84
G. sepium ILG 60
4.94 i 0.09
-0.41 rt 0.26
85
G. sepium ILG 61
4.84 -c 0.13
-1.30 + 0.12
95
G. sepium ILG 62
4.34 2 0.08
+0.27 t 0.14
77
G. sepium ILG 63
4.40 k 0.09
-0.07 rt 0.07
81
G. sepium HYB
4.44 + 0.12
+o.oo + 0.09
80
Mean for nodulating plants
4.78
-0.20
83
Note. Each value is the mean of four replicates. Numbers following + are standard deviation from
mean. ND, not determined.
“Mean of G”Nref of B. rufescens and A. indica.
%Jean of 6”Nref of G. senegalensis and A. indica.
‘Mean of 615Nref of C. siamea and A. indica.
N, fixation in this zone. This was why we have cultivated G. sepium in plastic bags under
greenhouse conditions by providing water and mineral nutrients (except N). Two treat-
ments-Tl (G. sepium growing on Nioro soi1 where tested trees have been planted) and
T2 (G. sepium inoculated with a Bradyrhizobium strain TAL5 and growing on Nioro
soil)-were compared in a completely randomized design with four replicates. The results
obtained after three months of cultivation (Table 4) showed a higher 615N value for G.
sepium than those obtained in situ (Table 3). It cari be deduced from these test results that
water and mineral elements (N excluded) were not the limiting factors of N, fixation at
Nioro. The major limiting factor was probably a deficiency of the symbiosis due to
absence of specific native strains of Bradyrhizobium or to other constraints different from
lack of water or minera1 elements. Inoculation with an effective strain of Bradyrhizobium
Table 4
Inoculation effect of the Bradyrhizobium TAL5 strain on dry matter (g plant-‘)
and S”N (parts per thousand) of Gliricidia sepium
Dry weight (g plant-‘)
S15N
Treatment
Stems + leaves
Nodules
Stems + leaves Nodules
Tl: G. sepium
0.53 2 0.08
ni1
9.09 t, 2.59
ni1
+Nioro soi1
T2: G. sepium
4.87 t 0.96
0.61 2 0.24
-1.99 It 0.08
12.69
+ Nioro soi1
+ TAL5 inoculum
Note. E&h value is the mean of four replicates. Numbers following i are standard
deviation from mean.
-
- - - I - - I
252
M. Ndiaye and F. Ganry
such as TALS, which should be tested under field conditions, is a possible improvement
technique of the N, fixation (Ndfa).
Discussion and Conclusion
The validity of the estimation of the N, fixation based on the 15N technique lies in the
choice of the reference plant, which is generally considered the main source of error (Fried
et al., 1983). Its choice is particularly difficult for NFTs for which nonnodulating lines are
not yet available. Cassia siamea (Lam.) and Eucalyptus grandis Hi11 ex Maiden have
already been used as reference plants for estimating the N, fixation of Leucaenu leuco-
cephala (Lam.) de Wit. and Acacia albidu (L.) syn. Fuidherbia albidu (Del.) A. Chev.
(Sanginga et al., 1990). Another diffïculty of the 615N method is the requirement that ref-
erence plants be very close to NFR TO reduce the possible errors in choosing the refer-
ence plants, Domenach et al. (1989) have proposed an average 615N value of several ref-
erence plants (rather than one) growing in the same environment as that of N~S. In the
present study, the average of the two closest 615Nref values to that of the soi1 was used at
each site. Our results showed that the 615N values of the soi1 were high enough (>4 parts
per thousand) and homogeneous enough within a site to allow the utilization of the 615N
method (Peoples et al., 1988).
At Bambey, the 615Nref (particularly that of C. siumeu) was lower than that of the soil,
but 615Nref values found at Nioro and Djibelor were very close to or identical to th’ose of
soils. These 615Nref increased with an ecological “north-south” gradient (Bambcy, +5.49;
Nioro, +5.57; Djibelor, +6.89).
The percentage of fixed N, was very low at Bambey (Ndfa I 22%) and Nioro (Ndfa
5 39%) and very high at Djibclor (76% I Ndfa 195%).
Based on the results of the experiment on Nioro soi1 under controlled conditions, we
advanced the hypothesis of lack of N, fixation due to the absence of Brudyrhizobium
strains or the inability of these strains to express themselves in the Nioro environment. In
contrast, specific strains of Bradyrhîzobium were present and were able to efficiently fix
N, at Djibelor. Therefore, we obviously showed that the two zones (Sudano-Sahelian and
sub-Guinean) are strongly contrasting in respect to N, fixation. In the Sudano-Sahelian
zone, the main constraints are the low clay content of soils (17%), and low rainfall (~$00
mm of precipitation per year) over a duration not exceeding 4 months, and secondarily
high concentration of minera1 N in the soi1 solution induced by marked drying and rewet-
ting cycles (Blondel, 1971; Sprent, 1976). These environmental conditions are not favor-
able to the establishment and expression of Bradyrhizobium strains because the nodulation
of host plant is insufficient and inadlequately repeated on time for developing and main-
taining specific population of native strains in the soil. However, these strains cari be re-
sent deep in the soi1 and fdrm effective nodules (Felker, 1986; Virginia et al., 1986; Du uy
& Dreyfus, 1992). In contrast, under a more humid climate like that at D.jibelor (21 i00
mm of precipitation per year), abundant nodulation is observed (Dupuy & Dreyfus, 112!),
thus favoring the buildup of efficient Bradyrhizobium strains when nodulations are redat-
ed. The findings of Ducousso et al. (1995) are in agreement with our explanation. Wo/rk-
ing in Acacia albidu, they have shown that most Bradyrhizobium strains isolated from khe
Sahelian zone are ineffective. They further demonstmted that the proliferation of Br ad,y-
rhizobium populations is strongly dependent on the durable presence of the host tree. $he
low values of %Ndfa found in the Egambey and Nioro sites showed that NFTs eval
ed
in these zones were more dependent on soi1 N than on the N,
i
fixation, because of m jor
constraints affecting N, fixation. To fix N,, Ieguminous trees need at least 700 mm (of
rainfal 1.
~
N2 Fixation by Tree Legumes in Senegal
253
The #“N values observed in C. siamea (+4.29 at Bambey, +4..58 at Nioro, and +5.57
at Djibelor) were very close to those of the S15N of NFTs reputed to fix N, (Allen & Allen,
1981). Nevertheless, the hypothesis of N, fixation by C. siamea cannot be confirmed due
to a number of reasons, including morphological factors (immobilization of Rhizobium at
the absorbent root hair base, reduced cortex) and biochemical factors (production of
antibacterial compounds, phenolic compounds, tannins and quinones in root cells) that
militate against the nodule formation on C. siamea roots (Allen & Allen, 198 1). Based on
our results, no valuable hypothesis cari be formulated to explain the low 615N values in C.
siamea. This induces interesting fields of investigation in order to find causes and deter-
mining factors of this phenomenon.
References
Allen, E. K., and 0. Allen. 1981. The nodulation profile of the genus Cassis, pp. 113-122, in P. S.
Nutman, ed., Symbiotic nitrogen jxution in plants. International Biology Programme no. 7.
Cambridge University Press, Cambridge.
Amarger, N., A. Mariotti, F. Mariotti, J. C. Durr, C. Bourguignon, and B. Lagacherie. 1979. Esti-
mate of symbiotically fixed nitrogen in field grown soybean using variations in % natural
abundance. Plant and Soi1 523269-280.
Beaux, M. F., H. Gouet, J. P. Gouet, P. Morleghem, G. Philippeau, J. Tranchefort, and M. Vemeau.
1988. Manuel d’utilisation STAT-ITCF Version N”4. Institut Techniques des Ceréales et des
Fourrages (ITCF), Services des Etudes Statistiques, Paris.
Blonde], D. 1971. Contribution à la connaissance de la dynamique de l’azote minéral en sol sableux
(Dior) au Sénégal. Agronomie Tropicale 26~1303-1333.
Bowen, G. D., N. Sanginga, and S.K.A. Danso. 1990. Biological nitrogen fixation in agroforestry-
An overview, pp. 170-175, in Proceedings of 14th International Congress of Soi1 Science, Vol-
ume III. International Institute of Agroenvironmental Sciences, Kyoto, Japan.
Charreau, C., and R. Nicou. 1971. L’amélioration du profil cultural dans les sols sableux et sablo-
argileux de la zone tropicale sèche ouest-africaine et ses incidences agronomiques. Bulletin
Agronomique No. 23. IRAT, Paris.
Domenach, A. M., and F. Kurdali. 1989. Influence des réserves azotées sur la formation des feuilles
d’dlnus glutinosa et ses conséquences dans l’estimation de la fixation d’azote. Canadian Jour-
nul of Botany 67:865-871.
Domenach, A. M., F. Kurdali, and R. Bardin. 1989. Estimation of symbiotic dinitrogen fixation in
alder forest by the method based on natural % abundance. Plant and Soi1 11851-59.
Dommergues, Y. 1997. Les arbresfiateurs d’azote dans l’aménagement des écosystèmes méditer-
ranéens et tropicaux. (in press).
Ducousso, M., B. Sougoufara, N. Dupuy, and B. Dreyfus. 1995. Rôle de deux acacias (Acacia
mungium et Acacia albida) dans le maintien et l’amélioration de la fertilité des sols en Basse
Casamance. Rapport final du projet Agroforesterie III No. 91 L 0685. Ministère de I’Enseigne-
ment Supérieur et de la Recherche, Paris.
Dupuy, N. C., and B. L. Dreyfus. 1992. Brudyrhizobrium populations occur in deep soi1 under legu-
minous tree Acacia albida. Applied ami Environmental Microbiology 582415-2419.
Felker, P. 1986. Treephzntings in semi-arid regions. Elsevier, New York.
Fried, M., S.K.A. Danso, and F. Zapata. 1983. The methodology of measurement of N, fixation by
non-legumes as referred from field experiments with legumes. Canadian Journal of Microbiol-
ogy 29: 1053-1062.
Hansen, A. P., and J. S. Pate. 1987. Evaluation of the % natural abundance method and xylem sap
analysis for assessing N, fixation of understorey legumes in Jarrah (Eucalyptus arginatu (Donn
ex Sm.)) forest in South West. Australian Journal of Experimental Botany 38: 1446-1458.
Kang, B. T., L. Reynolds, and A. N. Atta-Krah. 1990. Alley farming. Advances in Agronomy
43:315-357.
Mariotti, A., B. Sougoufara, and Y. R. Dommergues. 1992. Estimation of nitrogen fixation using the
254
M’. Ndiaye and F. Ganty
natural abundance method in a plantation of Casuarina equisetifolia (Forst). Soi1 Biolo y and
Biochemistry 24647$63.
Peoples, M. B., D. F. Herrbdge, and J. F. Bergersen. 1988. Measurement of nitrogen fixation in trop
and shrub legumes, pp. 223-238, in Green manure in rice farming. Proceedings of a s:nnpo-
sium on sustainable agriculture, 25-29 May 1987. International Rice Research Institute, Los
Banos, the Philippines.
Sanginga, N., S. K. A. Danso, F. Zapata, and G. D. Bowen. 1990. Influence of reference trees on N,
fixation estimates in keucaena leucocephala and Acacia albida using %-labeling techn clues.
Biology and Fertility of Soi1 934, L-346.
Sanginga, N., G. D. Bowen, and S. K. A. Danso. 1991. Intra-specific variation in P accumulat on of
Leucaena leucoceph.ala and Gliricidia sepium as influenced by soi1 phosphate status. Plant ami
Soi1 133:201-208.
Shearer, G., D. H. Khol, R. A. Virginia, B. A. Bryan, J. L. Skeeters, E. T. Nilsen, M. R. Shari i., and
P. W. Rundel. 1983. Estimates of N, fixation from variation in the natural abundance of i5N in
Sonoran Desert ecosystem. Oecologia (Berlin) 56365-373.
Singer, M. J., and D. N. Munns. 1996. Soils. An introduction. Prentice Hall, Upper Saddle River,
New Jersey.
Snoeck, D. 1995. Interactions entre végétaux fixateurs d’azote et non fixateurs en culture mixte: Cas
des Leucaena spp.associéees à Coffea arabica L. au Burundi. Thèse de Doctorat. lJni@sité
Claude Bernard-Lyon 1, Lyon, France.
Soi1 Survey Staff. 1992. Keys to soi1 taxonomy, Monograph No. 19.5th ed. Agency for Inte
tion-
a1 Development, U.S. Department of Agriculture. Pocahontas Press, Blacksburg, Virgini
Sprent, J. 1976. Water deficit and nitrogen fixing root nodules, pp. 291-315, in T. T. Kozlows1i ed.,
Water-dejcit andplant growth. Academic Press, New York.
Sprent, J. J., and Sutherland, J. M. 1!390. Nitrogen fixing woody legumes. Nitrogen Ftiing 1Irees
i -
Research Reports 8: 17-3 1.
Vincent, J. M. 1970. A manualfor thepractical study of the root nodule bacteria. International Biol-
ogy Programme Handbook No. 15. Blackwell Scientific Publications, Oxford, UK.
Virginia, R. A., M. B. Jenkins, and W. M. Jarrell. 1986. Depth of mot symbiont occurrence in soil.
BioZogy and Fertility of Soi1 2: 127-l 30.
Yoneyama, T. 1987. N, fixation and natural abundance of leguminous plants and Azolla. Bulletin of
National Institute of Agrobiologiscal Resources (Japan) 3z59-87.
Yoneyama, T., K. Kouno, and J. Yazaki. 1990a. Variation of natural % abundance of cro ds and
soils in Japan with special reference to the effect of soi1 conditions and fertilizer applioation.
Soi1 Science and Plant Nutrition 36667-675.
Yoneyama, T., T. Mukarami, N. Boonkerd, P. Wadisirisuk, S. Siripin, and K. Kouno. 1990~. Nat-
ural ‘-% abundance in shrub and tree legumes, Casuarina and non fixing plants in Th ‘land.
i
Plant and Soi1 128287-292.
Yoneyama, T., T. Muroka,T. Murakami, and N. Boonkerd. 1993. Natural abundance of “N in trop-
ical plants with emvbasis on tree leaume. Plant and Soi1 153:293-295.
Zakra, N., Ouvrier, M., and G. De Taffin. 1991. Partitioning of 15N labelled mineral nitrogen i4I Aca-
cia and coconut, pp. 79-386, in .Proceedings of an International Symposium on the Use 4
ble Isotopes in Plant Nutrition, Soi1 Fertility and Environmental Studies, jointly organi 21 fZ
the International Ato ic Energy ,4gency (IAEA) and the Food and Agriculture Organizat/ion of
the United Nations ( AO), Vienna, l-5 October 1990. IAEA, Vienna.
Zakra, N. Weaver, R. W., and G. De Taffin. 1993. Distribution of ‘%I from ammonium fertil zer in
field grown Acacia. ~lant amISoi 151:147-150.