Biol Fert Soils (1985) 1: 15-23 V/ Springer-Verlag ...
Biol Fert Soils (1985) 1: 15-23
V/ Springer-Verlag 1985
Inoculation with Glomrrsmosseae improves
N2 fixation by field-grown soybeans
F. Ganry’, H.G. Diem’, J. Wey’, and Y.R. Dommergues’
1 1RAT;ISRA Centre national de recherches agricoles. Bambey. Sénégal
2 CNRSORSTOM: B.P. 1386. Dakar. Sénégal
Summary. A field study carried out in a sandy,
cular mycorrhizae (VAMji on N, fixation by legumes
relatively acid Senegalese soi1 with a low soluble P
especially in P-deficient soils. Because of the general-
content (7 ppm) and low vesicular-arbuscular my-
ly low availability of P in tropical soils, the potential
corrhizal (VAM) populations showed that soybean
for the exploitation of VAM in the culture of legumes
responded to Glomus YtzosscIle inoculation when the
seems to be greater than in temperate soils. Howev-
soluble P level in the soi1 had been raised by the
er, this view should be tempered by the fact that
addition of 22 kg P ha-‘. In P-fertilized plots,
limitations exist in the field that often obliterate the
N, -fixation of soybean, assessed by the A value me-
stimulation of legume N:-fixing activity by VAM.
thod. was 109 kg NI fixed ha-’ when plants were
Thus field experiments are needed to find out wheth-
inoculated with Rhizobium alone and it reached 139
er inoculation with VAM cari improve NZ fixation by
kg NZ fixed ha-’ when plants vvere dually inoculated
legumes. Up to now only a few field experiments
wi th Rhizobium and Glomus mosseue using an algina-
dealing with legumes have been set up in the tropics
te bead inoculum. In addition to this N, fixation
(e.g.. Islam et al. 1980; Islam and Ayanaba 1981;
increase (t-28%), Glomcis mosseae inoculation signi-
Bagyaraj et al. 1979). In some cases the results were
t’icantly improved grain yield (+ 13%) and total N
inconclusive and no reliable method was used to
content of grains (i-16%). This success was attribu-
assess the effect of VAM inoculation on N, fixation.
ted mainly to the low infection potential of the native
A previous field study in Senegal, Ganry et al.
V.AM populations in the experimental site. In treat-
(1982), indicated that inoculation of soybean with
ments without soluble P or with rock phosphate, no
Glomus mosseae increased the harvest index and
effect of VAM inoculation was observed.
N? fixation assessed using the A value method (Fried
and Broeshart 1975). The rainfall was irregular du-
Key words: Glomus mosseae - Soybean - Inoculation
ring the growth cycle and drought spell occurred
- A value method - Senegal
during pod filling. These unfavorable climatic condi-
tions probably affected the NI-fixing activity of the
plants since even under the best conditions (i.e., in
P-fertilized plots inoculated with Gfomus mosseae)
Laboratory and greenhouse experiments have clearly
the total amount of fixed NZ was relatively low
demonstrated the beneficial effect of vesicular-arbus-
(60 kg ha-‘) and only 41% of trop N was derived from
Nz fixation. SO it appeared necessary to repeat this
Ofj+rint reyursts /o: Y. R.Dommergues
field experiment, hopefully under more favorable
climatic conditions. In addition, we carefully chose
prelimirzary pot experirnent Jor choosing
an experimental site with a low VAM inoc&m
the experimental size
poten&], SO that we could expect a satisfa :t::)ry
response of the trop to VAM inoculation.
The experimental site, which was selected from three
sites (A, B, C) for its lower VAM infection potentiai.
necessitated a preliminary pot experiment. We com-
Material and methods
pared the endomycorrhizal infection percentage of
soybean (cv. 44Ai73) grown in pots filled with 3 kg
The field experiment was carried out at the JSFA
ofsoils A, B, C (nonsterile soils). For each soi1 there
(Institut Sénégalais de Recherches Agricole:) re-
were two treatments: treatment 0: no inoculation:
search station of Sefa, South Senegal, in 1982 The
treatment G: inoculation with Gfomus mosseae. In-
soi1 was a leached ferruginous tropical soi1 (alficc nitru-
oculation ws achieved hy introducing 20 beads of
stox) in which soybeans had never been grown ;and
alginate-entrapped Glomus mosseae in the rhizo-
which had lain fallow for 5 years hefore the e::peri-
sphere of the soybean seedlings when they were at
ment.
their first-leaf stage. Pots were placed in a green-
house under the climatic conditions prevailing at
Dakar in February and March.
Materials
The soybean cultivar used was cv. ISRA-IRAT 26/72
Prcliminar experinzent,for choosing the standard trop
obtained from the CNRA (Centre National d: Re-
cherches Agronomiques) station, Bambey, Seriegal.
TO choose the standard trop (i.e., the non-N,-fixing
The Rhizobium peat base inoculum, which contained
trop to be compared with the NI-fixing one) we
3 x 10’ living cells g-’ (fresh weight), was applied by
planted soybean cv. ISRA-IRAT 26172 (the same cv.
hand to the seedling bed at the rate of 25 kg ha-‘, <Ywo
as thc one used in the field experiment) with two
strains of Rhizobiurn were used: an effective strain,
treatments: (1) no inoculation (to avoid contamina-
USDA 138, and an ineffective one, strain Gl (Laga-
tion, plots were separated by sheets of corrugated
cherie et al. 1977). Gl was used in the standard trop.
iron) and (2) inoculation with the ineffective strain
The VAM inoculant was prepared as wet be:.d!s of
G 1. Uninoculated soybeans were found to be slightly
Gfomus mosseae entrapped in alginate accord ng to
nodulated either by native strains compatible with
the method proposed by Diem et al. (I%l). E,ach
soybean as already observed by P. Jara (persona1
bead contained ca. 12 mg (fresh weight) of infeeted
communication) in many Senegalese soils. The atom
roots, spores, and hyphae. Inoculation was per Form-
percentage “‘N excess in soybeans inoculated with
ed by introducing 10-15 beads beneath the seed 3-4
the ineffective strain was higher than that of inocula-
cm deep into the SO~I. It is interesting to note t rat to
ted soybeans, indicating that even if the former soy-
obtain 1 1 of this Glomus mosseae inoculant it is
beans fixed a small amount of NI. their fixation was
necessary to grow an area of 0.3 m’of Vipa unjpicu-
lower than that of uninoculated soybeans. Thus soy-
lata (uninoculated with Rhizobium), thc pla rt we
beans inoculated with the ineffective strain were
routinely use to multiply Glomus mosseae il the
chosen as the standard trop.
greenhouse.
Experimental desip
Rainfall
A split-plot experimental design was used with eight
The total rainfall before sowing (May 1 to Ju .y 17)
replicates. The main treatments were:
was 170 mm and during the growth cycle (July 1,7 to
October 10, harvest time) it was 692 mm. The r;.infall
1. Inoculation with the ineffective Rhizobi~4rn
distribution was fairly even, without any rnarkc,d dry
strain; N fertilizer: 90 kg N ha” (I-90 N)
spell.
2. Inoculation with the effective Rhizohium strain;
starter N fertilizer: 17 kg N ha-’ (R-17 N)
F. Ganry et al.: Glomus mosseae improves NJ fixation by soybeans
17
3. Inoculation with the effective Rhizobium strain
Results
and Glomus mosseae; starter N fertilizer:
17 kg N ha-’ (RG-17 N)
Pot experiment
The subretitments were:
Table 1 shows that soi1 C had a lower VAM infection
1. No P addition (OP)
potential than soils A and B; thus soi1 C was chosen
2. P added as supertriple, 22 kg P ha-’ (super)
for the field experiment. The main characteristics of
3. P added as Taiba rock phosphate, 22 kg P ha-’
the soi1 C (O-20 cm horizon) were as follows: sand
(rock P)
(20-2000 pm), 83% pH H?O (1:2.5), 6.2; organic C,
0.40%; organic N, 0.038%; total exchangeable ca-
Each main plot (40.25 m’) was divided into three
tions, 1.29 mEq 10.‘g; total exchange capacity,
subplots (12.25 m’). In each subplot an area of
1.66 mEq l@‘g; total P. 197 ppm; available P
6.10 m’ was used for yield estimation and an area of
(Truog), 7 ppm. One should note that soi1 C differed
1.65 m’ was used for the 15N labeling. Al1 the experi-
only from soils A and B in the fact that the root
mental plots were fertilized with KCl at the rate of infection by native VAR4 fungi was significantly lo-
90 kg h-l. Labeled N fertilizer was applied as
wer at the 20th and 38th days, but was the same when
(‘SNH4)rSOJ with 1.01 atom percent ‘“N excess for
the plants wer 60 days old. In a11 soils. plants respon-
the 90 kg N ha-’ application and 4.73 atom percent
ded similarly to the inoculation with Glomus
‘“N excess for the 17 kg N ha-’ application.
mosseae, indicating that no limiting factor occurred
Analysis of the plants
in preventing the establishment of G. mosseae.
Plants were carefully harvested avoiding contami-
Table 1. VAM infection frequency of soybeans grown in
nation with soi1 N; leaves, stems, husks, and grains
nonsterile soils A, B, and C from the Sefa experimental
were sampled and analyzed separately. The samples
station (preliminary pot experiment)
were dried at 6S”-70°C for 24 h, weighed, ground into
a 40-mesh powder, and anaiyzed for total N content
VAM infection frequency (%) at
according to the Kjeldahl method. ‘“N analyses were
Soils
Treat-
20th day
38th day
40th day
carried out at the Seibersdorf Laboratory (IAEA)
ments
using Dumas’ method (the combustion performed in
A
0
9
17
5 5
this technique converts total N directly to NJ and
G
14
19
57
emission spectrometry. For the sake of simplification
B
0
19
20
55
figures related to leaves. stems, and husks were
G
18
25
54
pooled under the term shoot, but data related to
C
0
4
12
56
G
18
24
61
grains were presented separately.
The amount of N? fixed was evaluated according to
0, no inocultion; G, inoculation with G. mosseae
the A value method (Fried and Broeshart 1975).
Each value is the mean of the results from tïve replicates
Root samples from each treatment were stained
with trypan blue in lactophenol using the method of
Phillips and Hayman (1970).
Field experiment
Frequency and intensity of VAM infection were
then assessed according to Ollivier et al. (1983). The
Infection by VAM. Table 2 shows that there was no
interpretation of the whole set of data was performed
significant interaction between the main treatments
according to the test proposed by Quidet and Mas-
and the subtreatments. The only significant main
mejean (1982), which indicated the level of sigifi-
effect on infection frequency was that of inoculation
cnnt (P = 0.05) of the main treatments and their
with Glomus mosseae within subtreatment Super
possible interaction. Within each treatment, sub-
(application of superphosphate): without G. mosseae
‘ireatments were compared and within each subtreat-
inoculation, the infection frequency was 65% and
ment the main treatments were also compared. Re-
with inoculation it was 87% at the 26th day. This
lated LSDs are indicated at the bottom of each table
effect disappeared when plants were older (40th
of data.
day). There were significant effects of G. mosseae
-..
-,
18
F. Ganry et al.: G1omu.s rnosseae improves NJ fixation by soybeans
Table 2. Nodule weight, frequency, and intensity of VAM ir faction of soybean roots 26 and 40 days after in-
oculation witb Rbizabium japonicnm (USDA 138) alone (RI or with Rhizobium japonicum (USDA 138) plus
Glomus mosseae (RG)
--~
-
Treatments
26 day.
40 days
1
~--
Main treatments’
Subtreat-
Nodule
VAM infection
Nodule
VAM infection
ments
-.
-
Inocu-
N ferti-
(P ferti-
dry wtb
Frequency
Intensity dry wtb Frequency Intensity
lation
lizer
lizer)
(%)
w
(96)
(%)
-
-
-
-
R
17N
OP
14
70
25
74
98
42
R
17N
Super
27
65
25
103
96
42
R
17N
Rock P
2 1
68
27
98
96
45
RG
17N
OP
22
73
34
76
91
39
RG
17N
Super
61
87
35
98
92
40
R G
17N
Rock P
45
75
28
87
93
42
LSD between the main treatments
10
6
5
N S
within the same subtreatment
LSD between the subtreatments
1 2
5
4
N S
within the same main treatment
~-
a 17N: N ferrilizer added at the rate of 17 kg N ha-’
b Geometrical mean for dry wt. at the 26th day
inoculation on infection intensity: one wi
RG-17N) and form of P fertilizer (Super versus rock
subtreatment Super (25% versus 35%) and t
phosphate) were observed for (1) grain yield express-
within the subtreatment OP (25% versus 34%).
ed as kilograms dry weight per hectare, (2) grain yield
The coefficient of variation of infection f
expressed as kilograms N per hectare, (3) grain N
was 4.5% for the whole set of plots inocul
concentration, and (4) grain and shoot total N ex-
G. mosseae, which is a much lower figure
pressed as kilograms N per hectare.
for plots not inoculated with G. mosseae (
Harvest index (grainlshoot ratio): Comparing soy-
bean inoculated with Rhizohium USDA 138 and
NJ Fixation. Nodulation (Table 2): The
G. mosseae (treatment RG), we note (Table 5) that
significant interaction between the main treat
in the absence of P fertilizer, dual inoculation (treat-
and the subtreatments. There were two
ment RG) increased only slightly the harvest index
main effects:
expressed as dry weight (+2%), total N (6%). and
1. Inoculation with G. mosseae incr
total P (+9%): in the presence of P fertilizer (Super)
weight of nodules per plant in the ear
the effect of dual inoculation (treatment RG) on
stage (from 21 to 45 mg). This effect
harvest index was much more marked: the increase of
ed when plants were older (40th day)
harvest index expressed as dry weight, total N, and
2. Effect of P addition, whatever form
total P was t>%, 13%, and 22% respectively.
Table 5 shows that inoculation of soybean with
Percentage of N derived from NJ fixatio
Rhizobium USDA 138 (R) instead of the ineffective
There was no significant interaction
strain of Rhizohium (1) improved the harvest index
main treatments and the subtreatments
when expressed on dry weight and total N basis.
two significant main effects:
1. Effect of G. mosseae inoculation on
Concluding remarks: When the level of soluble P in
N derived from NZ fixation within t
the soi1 was raised by adding 22 kg P ha-‘, inoculation
of soybean with
ment Super (R, 69.8%; RG, 75.9%)
Glomus mosseae increased
2. Effect of P addition, whatever form
NZ fixation (+28%). the grain yield (f 13%). and the
harvest index based on dry weight (+12%). The
Grain yield (Table 4): Significant interacti
increase in grain protein resulting from G. mosseae
tween inoculation with G. mosseae (R-17N
inoculation was relatively modest (+16%) but the
1 9
Table 3. Sources of N (expressed as a percentage of total plant N or in kg N ha-‘) in soybean
inoculated with Rhizobium japonicum (USDA 138) alone (R), or with Rhizobium japonicum
(USDA 138) plus Glomus mosseae (KG)
--_.l-l_l~
Treÿtments
Main treatments
Subtreat-
Fertilizer N
Soi1 Nb
Fixed N2
-
ments
Inocu-
N ferti- (P ferti-
%
kgN ha-l
%
kg N ha-l
%
kg N ha-l
lation
lizera
lizer)
-
R
17N
O P
2.6
2.8
30.1
32.8
67.3
73.1
R
17N
Super
2.0
3 . 1
28.2
44.0
69.8
109.0
R
17N
Rock P
2.0
2.9
26.2
39.2
73.8
110.3
RG
17N
OP
2.6
3 . 1
30.6
36.7
66.9
80.2
RG
17N
Super
1.6
3.0
22.5
41.3
75.9
139.3
RG
17N
Rock P
2 . 1
3.2
26.3
40.0
71.6
108.6
LSD between the main treat-
0.40
NS
5 . 3
14.5
ments within the same sub-
treatment
LSD between the subtreat-
0.44
NS
5.7
13.4
ments within the same
main treatment
a Applied at the rate of 17 kg N ha-l
b Calculated from fixation and labeled fertilizer data
Table 4. Grain yield (expressed as k dry wt ha-l or kg total N ha-l), grain N content (%), grain and shoot total
-B
N and total P (expressed as kg ha ) of soybean inoculated with an ineffective strain of Rbizobium japonicum
(I), with Rhizobium japonicum (USDA 138) alone (R), or with Rhizobium juponicum (USDA 138) plus
C~O~US mosseae (KG)
Main treatments
Sub-
Grain yield
Grain N
Grain and
Grain and
ments
content
shoot
shoot
total N
total P
Inocu-
N ferti-
(P ferti-
kg dry wt ha-la
kg N ha-’
W)
kg N ha-l
kg P Ila-’
lation
lizer
lizer)
~-
1
90N
OP
1093 (16)
65.7
6.01
84.2
5.6
1
90N
Super
1725 (15)
101.6
5.89
127.2
11.1
1
90N
Rock P
1482 (11)
86.2
5.81
109.1
8.2
R
17N
OP
1423 (21)
90.3
6.47
112.0
6.0
R
17N
Super
2017 (16)
133.8
6.58
161.0
10.0
R
17N
Rock P
1888( 9 )
124.6
6.62
150.2
8.2
RG
17N
OP
1431(21}
98.2
6.60
120.0
6.9
RG
17N
Super
2290 (11)
154.7
6.76
183.7
11.8
RG
17N
Rock P
1892 ( 7)
126.0
6.69
152.0
8.3
LSD between the main treatments
192
12.8
0.14
16.2
0.79
within the same subtreatment
L S D between the subtreatments
197
13.5
0.14
17.0
0.81
within the same main treatment
---__--
a In parentheses, coefficient of variation (%)
--
--
---
2 0
Table 5. Harvest index of soybean inoculated with an i n : H:ective
have been used successfully on a small scale, but the
strain of Rhizobium japonicum (l), with Rhizobium
0Jzicum
ia. 7’
method would require nearly 2 t ha-’ of soi1 inoculum
(USDA 138) alone (R’j, or with Rhizobium j~p()ZkZ rw (USDA
and the technology of pellets production needs to be
138) plus Glomus mosseae (RG)
-~~-
- --.-
improved (Hayman et al. lY81). Witty and Haymnan
Main treatments
Subtreat-
Harvest index expie ssed
(1978) used the fluid drilling of mycorrhizal soi1
ments
on the bases of - -_- wet-sicvings suspended in methylcellulose. Unfortu-
Inocu-
N ferti-
(P ferti-
Dry
Total
Total
nately such a technique is not suitable when drought
lation
lizera
lizer)
w e i g h t N
,P
conditions coincide with sowing (Hayman et al.
-
.-
-
-
1981).
1
90N
O P
0.33
3.55
3.8
R
17N
O P
0.43
4.18
4.3
In the field experiment reported in this paper we
R G
17N
OP
0.44
4.42
4.7
used the same type of alginate bead inoculum as the
1
90N
Super
0.36
3.98
4.4
one used earlier (Diem et al. 1978; Ganry et al. 1982).
R
17N
Super
0.48
4.71
4.5
The results presented here clearly confirm the conclu-
R G
17N
Super
0.51
5.33
5 . 5
__--.---
. ..-
.-
sion of this previous tria1 that VAM inoculum made
Harvest index: grainkhoot ratio
of spores, mycelium. and homogenized mycorrhizal
a 9ON, 17N: N fertilizer applied at the rate of 90 or 1’ 7 ki;; N ha‘l
roots entrapped in alginate beads cari successfully be
used for large-scale field inoculation. The alginate
absolute value was + 132 kg ha-‘, which is SUI ) it. :mtial
bead inoculum presents the following advantages:
gain if we consider that this figure is equivi 16:nt to
1, Simple preparation of a large amount of inocu-
1300 kg Pearl millet grain (assuming that the1 p1.otein
lum
content of this cereal is 10%). Inter( :st!ingly ,
2. Easy storage and transportation of the inoculum
G. mosseae inoculation reduced the coeffi(xi6 tnt of
to the field
variation of the grain yield expressed in Weight,
3. Easy incorporation into the soi1 at planting time
which confirms previous observations (Gan‘Y et al.
1982). In the 1980 experiment (Ganry et 211. 1982)
when no Super was added, total P content >f grain
Forms of’ phosphate to he used if2 combination with
and shoots expressed as kg P ha-’ was not affl:Cfed by
VA M inoculatiorz
G. mosseue inoculation since it was 6.9 in th 1 ,1.’ mino-
culated plots and 63 in the inoculated ones, t U’! when
A number of greenhouse experiments have shown
Super was added total P content was raised fi,011 n 10.6
that in sterile P-deficient soils addition of soluble P UP
in the uninoculated plots to Il .2 in the in1 )Cl Ilated
to an appropriate level is required to obtain a signi-
ones. Similar effects were observed in the 198 2,E :xper-
ficant response to VAM inoculation (Gianinazzi-
iment (present report), the related figure.S heing
Pearson and Diem lY82). In other words, in these
respectively 6.9 (uninoculated plots) and 6.0 ( iqlc Jcula-
soils we cari expect a positive interaction between
ted plots) in the absence of Super, and 10.0 (, Ir,.i nocu-
VAM inoculation and fertilization with soluble P. but
lated plots) and 11 .X kg P ha-’ (inoculated ~P’C )ts) in
such an interaction does not occur if soluble P is
the presence of S’uper.
replaced by insoluble P except when insoluble P
/
(rock phosphate) is solubilized subsequently tu its
application. These conclusions were confirmed in the
Discussion
field by a previous study (Ganry et al. 1982) and
reconfirmed here. thus it cari be claimed that, in some
Field inocrrlum preparation
tropical P-deficient soils. an appropriate addition of
soluble P may be required to optimize the response to
Although procedures usually propos4 for rflycorr-
VAM inoculation. Besides this concept our experi-
hizal inoculum preparation and field inl)culation
ment indicated that the form of added rock phospha-
have been reviewed by Hayman et al. (1981) and
te should be carfullyconsidered. Taiba rock phospha-
Menge and Timmer (19X2), only a few rep rts have
te that was used in our experiment was net taken up
dealt with field trials performed with the new forms
by mycorrhizal plants growing in the fie]d aithou&
of inoculum that have been proposed in the past
the soil wx sli&tly acidic (pH 6.2). T&s resu/tis
Years. soi] Pekts tht were devised by HiII1 (197~))
apparentlj nOt consistent with those ohtained by
Mosse et al. (1976), who found that inoculation with
Based on several papers on the effect of mycorr-
VAM fungi greatly improved the utilization of Gafsa
hizal inoculation (Powell et al. 1980: Abbott and
rock phosphate in some acidic soils (pH, 5-6.4). The
Robson 1978; Owusu-Bennoah and Mosse 1979), it is
reason for the discrepancy is probably that the rock
suggested that inoculated plants benefit especially
phosphate w’e used was very resistant to solubiliza-
from an early mycorrhizal infection compared with
tion or that some limiting factors occurred hindering
uninoculated ones. Early mycorrhizal infection al-
activity of rhizospheric phosphate-dissolving bacteria
lows the host plants, especially legumes, to increase P
(e.g., unfavorable water regime; lack of energetic
uptake in the first growth stage when P requirements
substrates).
are high. Carling et al. (1979) wrote that soybean
derives maximum benefit from a VAM fungus only if
it receives maximum exposure at the early seedling
Kesponse of field-gros plants to VA M inoculation
stage.
The effect of earlier VAM infection resulting from
In contrast with pot experiments, large-scale field
inoculation upon nodule weight of soybean is obvious
experiments on VAM inoculation have seldom been
when comparing the data of Table 2, subtreatmem
successful. Thus the satisfactory response to VAM
Super (the most favorable to the expression of the
inoculation we obtained here for the second time in
beneficial effect of VAM). In the early growth stage
Senegalese soils should be discussed. Probably the
(26th day) Glomus infection (frequency and intensi-
main cause of this response is that the experimental
ty) and nodule weight in plants inoculated with Glo-
design was set up in a soi1 chosen for its low VAM
mus mosseae (treatment RG-17N, Super) were 87%.
infection potential, but one cannot totally exclude
35%, and 61 mg respectively. whereas the corre-
the hypothesis that the introduced strain of Glomus
sponding figures for plants uninoculated with Rhizo-
mosseae was more effective than the native VAM
bium alone (treatment R-17N, Super) were only
strains and that it was competitive enough to outclass
65%, 25%‘. and 27 mg respectively. Later (40th day)
them.
there were no differences between the treatments.
suggesting that assessment of infection at a late stage
Low irzfkction potential oJ‘ the native VAM popula-
in plant growth is inadequate to ascertain the effect of
tion. Actually this expression refers to the inability of
inoculation. This is in agreement with the suggestion
the native VAM population to infect the plant in its
of Abbott and Robson (1981) that only an early
early growth stage. the late infection being possibly
assessment of infection is valuable in studies of plant
attributed either to the low number of native VAM
growth response to mycorrhizal inoculation.
propagules or to some intrinsic characteristics (such
Whereas the “rhizobiologist” is familiar with the
as low infectivity ability) of these propagules.
idea of evaluating the best locations for successful
Late infection by native VAM populations proba-
inoculation trials, the “mycorrhizologist” very sel-
bly explains why Howeler et al. (1982) found that
dom cares about that problem. which explains most
cassava -- a mycorrhiza-dependent plant which is
failures of VAM inoculation that are a11 the more
most easily infected by VAM fungi - responded
frequent as VAM fungi are ubiquitous organisms.
positively to inoculation only when VAM popula-
From this point of view preliminary inoculation exper-
tions were low. Similarly we obtained a positive
iments in the laboratory of greenhouse for selecting
response of soybean to VAM inoculation in soi1 C
sites of inoculation in the field are recommended and
chosen for its low VAM infection potential (Table 1).
methods for assessing indigenous VAM popuIations
Lt is also necessary to indicate that in our experiment
in soi1 (Porter 1979: Wilson and Trinick 1983) should
inoculation was performed by placing the Glomw
be developed.
rnosseae propagules (entrapped in the alginate be-
ads) in the seedbed (that is close to the germinating
Effectivity and competitive abiiity of the introduced
soybeans), which is a strategic position conferring a
strain of Glomus mosseae. The double concept of
marked advantage to Glomzu mosseae upon the na-
VAM effectivity (or capability to enhance absorption
tive VAM fungi (probably irregularly and widely
of nutrient by the host plant) and competitiveness is
scattered throughout the soi]) and also making possi-
still not very well established. However, a number of
ble an early infection of the soybean roots.
experiments support this concept. Thus Barea et al.
“~I”~‘,,,l~,ir---
_~.-
-.--
-.-
22
F. Ganry et al.: Glomus moueue improves NI fixation by soyheans
(1980) and Kucey and Paul (1982) have shown tha
Influence of the distribution qf prec-ipitation
Glomus mosseae improved P uptake more than indig
enous VAM fungi.
When comparing the data obtained in 1980 (Ganry et
Similary Powell et al. (1980) found that indigeno
al. 1982) and 1982 (present note) in P-fertilized plots
VAM .fungi were ineffective in many soils and t
inoculated with Clornus mosseae, we see that the
inoculation by more effective VAM fungi wo
amoung of NZ fixed was more than twice as high in
result in positive responses even in nonsterile so
1982 (139.3 kg ha-‘) than in 1980 (63 kg ha”). Similar-
containing a high indigenous VAM population,
ly the percentage of plant N derived from NI fixation
Kucey and Paul (1982) related the effectiveness of
was much higher in 1982 (75.9%) than in 1980
the introduced VAM fungus in relation to its CO
(41.4%). These differences cari be attributed to the
tibility with the host. They have shown that
harmful influence of drought periods on N- fixation
beans growing in a field previously planted wi
that occurred in 1980, a year when the distribution of
wheat would benefit from inoculation wi
precipitation was very irregular.
G’. mosseae because indigenous VAM species spo
By contras& in the P-fertilized plots the increase in
taneously selected by wheat may not be efficient f
N2 fixation resulting from Glomus mosseae inocula-
promoting growth of Faba beans. Furthermore,
tion was identical in the 1980 (+29 kg ha“) and in the
ling and Brown (1980) and Schenck and Smith (1
1982 (+30 kg ha-‘) experiments. Similarly the yield
have found that G. mosseae was more stimulatory tc
increase due to Glomus mosseae inoculation was
soybean than many other species of VAM fungi an
roughly the same in 1980 (-t-302 kg dry weight ha“)
response of soybean to G. mosseae inoculation wa
and 1982 (t273 kg dry weight ha-‘).
particularly marked at a high temperature (36°C).
In 1982, the beneficial effect of VAM inoculation
Since our experimental design was set up in a SO
was attrihuted to the fact that it had induced early
which had lain fallow for many years, we cari assum
infection in the roots of soybean grown in a soi1 with a
that the introduced strain of C/ornus mosseue w
low VAM infection potential whereas in 1980 the
more adapted to soybean than the native VA
inoculation effect was mostly attributed to the fact
microflora. However, this speculation should
that during drought periods, which had occurred
ther explored, using reliable experimental tools
throughout the growth cycle of soybean, VAM pro-
as labeled strains of VAM and “‘P-labeled fertih
tected the host plants against the harmful effects of
In any case, data from the literature and this p
water stress (Sprent 1979; Gianinazzi-Pearson and
suggest that G. mosseae could be one of the
Diem 1982).
VAM fungi promoting maximum growth of soybe
in different soi1 types and climatic conditions.
Acknowledgments. We thank Youssouph Ndiaye and Moussa
Niang for technical assistance. The investigation was funded in
inoculation versus N fertilization
part by IAEA (joint FAOIIAEA coordinated Research Pro-
gramme. Contrat n”DI-SEN 2375).
Inoculation of soybean with effective strain Rhi;
bium USDA 138 increased the yield and grai
References
ratio more than fertilizer application. The a
N? fixed due to Rhizobium inoculation alone
Abbott LK, Robson AD (1978) Growth of subterranean clover in
estimated to be ca 109.0 kg N ha-‘; when soy
relation to the formation of endomycorrhizas by introduced
inoculated simultaneously with Rhizobium
and indigenous fungi in a field soil New Phytol 81575-585
mus mosseae, N? fixation was 139.3 kg N ha.‘,
Abbott LK. Robson AD (1981) Infectivity and effectivcness of
is the highest figure reported in a field expe
vesicular-arhuscular mycorrhizal fungi: effcct of inoculum
Senegal to date. From Tables 3 and 4 we
type. Aust J Agric Res 32:h31-639
Bagyaraj DJ, ManJunath A, Patil RB ( 1979) Interaction hetween a
that increasing N? fixation made possible i
vesicular-arbuscular mycorrhiza and Rhizohiurn and their ef-
yield without depleting soi1 N content.
fects on soybean in the field. New Phytol X2:141-136
F. Ganry ct al.: Glotnus mo.~sscuc improves N2 fixation by soybeans
23
Barea JM. Escudero JL, Azcon-G de Aguilar C (1980) Effects of
Menge JA, Timmer LW (1980) Procedures for inoculation of
introduced and indigenous VA mycorrhiza fungi on nodulati-
plants with vesicular-arbuscular mycorrhizae in the laboratory,
on. growth and nutrition of Medicago sativa in phosphate-fixing
greenhouse, and field. In: Schenck NC (ed) Methods and
soils as affected by P fertilizers. Plant Soi1 5432X3-296
principles of mycorrhizal research. The American Phytopatho-
Carling DE, Brown MF (19X(I) Relative effect of vesicular-arbus-
logical Society, St. Paul, Minnesota, pp 59-68
cular mycorrhizal fungi on the growth and yield of soybeans
Masse B, Powell CL. Hayman DS (1976) Plant growth responses
(G‘&5w max). Soil Sci Soc Amer J 44:528--532
to vesicular-arbuscular mycorrhiza. IX. - Interactions between
Carling DE, Brown MF. Brown FA (1979) Colonization rates and
VA Mycorrhiza, rock phosphate and symbiotic nitrogen fixa-
growth responses of soybean plants infected by vesicular-arbus-
tion. New Phytol76:331-342
cular mycorrhizal fungi. Can J Bot 57: 1769-1772
Ollivier B. Bertheau Y. Diem HG. Gianinazzi-Pearson V. (1983)
Diem HG, Jung G, Mugnier J, Ganry F. Dommergues Y.R.
Influence de la variété de V@a unguiculutu C. Walp dans
(1981) Alginate-entrapped Glomus mosseae for trop inocula-
l’expression de trois association endomycorrhizienne a vesicu-
tion. In: Proc 5th Nacom. University Laval, Quebec, p 23
les et arbuscules. Can J Bot 61:354-358
Fried M, Broeshart H (1975) An independent measuremcnt ofthe
Owusu-Bennoah E, Mosse B (1979) Plant growth responses to
amount of nitrogen fixed by a legume trop. Plant Soi1 43:
vesicular-arbuscular
mycorrhiza. XI. Field inoculation respon-
707-7 I 1
ses in barley, lucerne and onion. New Phytol 83:671-680
Ganry F. Diem HC, Dommergues YR (1982) Effect of inoculation
Phillips JM. Hayman DS (1970) Improved procedure for clearing
with Giomus mosJeue on nitrogen fixation by field grown
roats and staining parasitic and vesicular-arbuscular
mycorr-
soybeans. Plant Soi1 68:321-329
hizal fungi for rapid assessment of infection. Trans Brit Mycol
Gianinazzi-Pearson V, Diem HG (19X2) Endomycorrhizae in the
Soc 551%-161
tropics. In: Dommergues YR. Diem HG (eds) Microbiology of
Porter WM (lY79) The ‘most probable number’ method for enu-
tropical soils and plant productivity Nijhoff/Junk, The Hague,
merating infective propagules of vesicular-arbuscular mycorr-
pp 2flY-2s 1
hizal fungi in soil Auçt J Soi1 Res 17:5lS-519
Hall IR (1979) Soi1 pellets to introduce vesicular-arbuscular my-
Powell CL, Groters M, Metcalfe D (19X0) Mycorrhizal inoculation
corrhizal fungi into soil. Soi1 Biol Biochem 1 l:R5<-86
of a barley trop in the field. NZJ Agr Res 23: 107-109
Hayman DS, Morris El, Page RJ (1981) Methods for inoculating
Quidet P, Masmejean T ( 1962) Organisation. conduite et interpré-
field crops with mycorrhizal fungi. Ann Appl Biol 99:247-253
tation des essais fumure. Services Agronomiques de la Société
Howeler RH, Cadavid LF. Burckhardt E (1982) Response of
Commerciale des Potasse d’alsace. SEDA, Paris, p 108
Cassava to VA mycorrhizal inoculation and phosphorus appli-
Schenck NC, Smith GS (1982) Responses of six species of vesicu-
cation in greenhouse and field experiments. Plant Soi1 69:327-
lar-arbuscular mycorrhizal fungi and their effects on soybean at
339
four soi1 temperatures. New Phytol92:19-3-201
Islam R, Ayanaba A ( 1981) Growth and yield responses of cowpea
Sprent JI (1975) Nitrogen fixation by legumes subjected to water
and maize to inoculation with C/ornus mosseae in sterilized soi1
and light stresses. In: Nutman PS (ed) Symbiotic nitrogen
under field conditions. Plant Soi1 63:505-509
fixation in plants. Cambridge University Press, London. New
Islam R, Ayanaba A. Sanders FE (1980) Response of cowpea
York, pp 405-420
(I/ign~ unguiculata) to inoculation with VA-mycorrhizal fungi
Wilson JM. Trinick MJ (1983) Infection development and interac-
and to rock phosphate fertilization in some unsterilized Niger-
tions between vesicular-arbuscular mycorrhizal fungi. New
ian soils. Plant Soi1 54: 107-l 17
Phytol 93:543-553
Kucey RMN. Paul EA (1982) Vcsicular-arbuscular mycorrhizal
Witty JF, Hayman DS (1978) Slurry inoculation of VA mycorrhi-
populations in various Saskatchewan soils and the effect of
za. Rothamsted Experimental Station Report for 1977. Part
inoculation with Glomus mosseae on Faba hean growth in
1:239-240
greenhouse and field trials. Can J Soi1 Sci 63:87-95
Lagacherie B. Hugot R, Amarger N (1977) SekCtiOn of Rhizo-
hium japonictim strains on their competitiveness for infection.
Ann Agron 2X:379-390
Received September 20, 1984
-.___--
--c.-
__C_.---m-
-.“ml-n---
V/ Springer-Verlag 1985
Inoculation with Glomrrsmosseae improves
N2 fixation by field-grown soybeans
F. Ganry’, H.G. Diem’, J. Wey’, and Y.R. Dommergues’
1 1RAT;ISRA Centre national de recherches agricoles. Bambey. Sénégal
2 CNRSORSTOM: B.P. 1386. Dakar. Sénégal
Summary. A field study carried out in a sandy,
cular mycorrhizae (VAMji on N, fixation by legumes
relatively acid Senegalese soi1 with a low soluble P
especially in P-deficient soils. Because of the general-
content (7 ppm) and low vesicular-arbuscular my-
ly low availability of P in tropical soils, the potential
corrhizal (VAM) populations showed that soybean
for the exploitation of VAM in the culture of legumes
responded to Glomus YtzosscIle inoculation when the
seems to be greater than in temperate soils. Howev-
soluble P level in the soi1 had been raised by the
er, this view should be tempered by the fact that
addition of 22 kg P ha-‘. In P-fertilized plots,
limitations exist in the field that often obliterate the
N, -fixation of soybean, assessed by the A value me-
stimulation of legume N:-fixing activity by VAM.
thod. was 109 kg NI fixed ha-’ when plants were
Thus field experiments are needed to find out wheth-
inoculated with Rhizobium alone and it reached 139
er inoculation with VAM cari improve NZ fixation by
kg NZ fixed ha-’ when plants vvere dually inoculated
legumes. Up to now only a few field experiments
wi th Rhizobium and Glomus mosseue using an algina-
dealing with legumes have been set up in the tropics
te bead inoculum. In addition to this N, fixation
(e.g.. Islam et al. 1980; Islam and Ayanaba 1981;
increase (t-28%), Glomcis mosseae inoculation signi-
Bagyaraj et al. 1979). In some cases the results were
t’icantly improved grain yield (+ 13%) and total N
inconclusive and no reliable method was used to
content of grains (i-16%). This success was attribu-
assess the effect of VAM inoculation on N, fixation.
ted mainly to the low infection potential of the native
A previous field study in Senegal, Ganry et al.
V.AM populations in the experimental site. In treat-
(1982), indicated that inoculation of soybean with
ments without soluble P or with rock phosphate, no
Glomus mosseae increased the harvest index and
effect of VAM inoculation was observed.
N? fixation assessed using the A value method (Fried
and Broeshart 1975). The rainfall was irregular du-
Key words: Glomus mosseae - Soybean - Inoculation
ring the growth cycle and drought spell occurred
- A value method - Senegal
during pod filling. These unfavorable climatic condi-
tions probably affected the NI-fixing activity of the
plants since even under the best conditions (i.e., in
P-fertilized plots inoculated with Gfomus mosseae)
Laboratory and greenhouse experiments have clearly
the total amount of fixed NZ was relatively low
demonstrated the beneficial effect of vesicular-arbus-
(60 kg ha-‘) and only 41% of trop N was derived from
Nz fixation. SO it appeared necessary to repeat this
Ofj+rint reyursts /o: Y. R.Dommergues
field experiment, hopefully under more favorable
climatic conditions. In addition, we carefully chose
prelimirzary pot experirnent Jor choosing
an experimental site with a low VAM inoc&m
the experimental size
poten&], SO that we could expect a satisfa :t::)ry
response of the trop to VAM inoculation.
The experimental site, which was selected from three
sites (A, B, C) for its lower VAM infection potentiai.
necessitated a preliminary pot experiment. We com-
Material and methods
pared the endomycorrhizal infection percentage of
soybean (cv. 44Ai73) grown in pots filled with 3 kg
The field experiment was carried out at the JSFA
ofsoils A, B, C (nonsterile soils). For each soi1 there
(Institut Sénégalais de Recherches Agricole:) re-
were two treatments: treatment 0: no inoculation:
search station of Sefa, South Senegal, in 1982 The
treatment G: inoculation with Gfomus mosseae. In-
soi1 was a leached ferruginous tropical soi1 (alficc nitru-
oculation ws achieved hy introducing 20 beads of
stox) in which soybeans had never been grown ;and
alginate-entrapped Glomus mosseae in the rhizo-
which had lain fallow for 5 years hefore the e::peri-
sphere of the soybean seedlings when they were at
ment.
their first-leaf stage. Pots were placed in a green-
house under the climatic conditions prevailing at
Dakar in February and March.
Materials
The soybean cultivar used was cv. ISRA-IRAT 26/72
Prcliminar experinzent,for choosing the standard trop
obtained from the CNRA (Centre National d: Re-
cherches Agronomiques) station, Bambey, Seriegal.
TO choose the standard trop (i.e., the non-N,-fixing
The Rhizobium peat base inoculum, which contained
trop to be compared with the NI-fixing one) we
3 x 10’ living cells g-’ (fresh weight), was applied by
planted soybean cv. ISRA-IRAT 26172 (the same cv.
hand to the seedling bed at the rate of 25 kg ha-‘, <Ywo
as thc one used in the field experiment) with two
strains of Rhizobiurn were used: an effective strain,
treatments: (1) no inoculation (to avoid contamina-
USDA 138, and an ineffective one, strain Gl (Laga-
tion, plots were separated by sheets of corrugated
cherie et al. 1977). Gl was used in the standard trop.
iron) and (2) inoculation with the ineffective strain
The VAM inoculant was prepared as wet be:.d!s of
G 1. Uninoculated soybeans were found to be slightly
Gfomus mosseae entrapped in alginate accord ng to
nodulated either by native strains compatible with
the method proposed by Diem et al. (I%l). E,ach
soybean as already observed by P. Jara (persona1
bead contained ca. 12 mg (fresh weight) of infeeted
communication) in many Senegalese soils. The atom
roots, spores, and hyphae. Inoculation was per Form-
percentage “‘N excess in soybeans inoculated with
ed by introducing 10-15 beads beneath the seed 3-4
the ineffective strain was higher than that of inocula-
cm deep into the SO~I. It is interesting to note t rat to
ted soybeans, indicating that even if the former soy-
obtain 1 1 of this Glomus mosseae inoculant it is
beans fixed a small amount of NI. their fixation was
necessary to grow an area of 0.3 m’of Vipa unjpicu-
lower than that of uninoculated soybeans. Thus soy-
lata (uninoculated with Rhizobium), thc pla rt we
beans inoculated with the ineffective strain were
routinely use to multiply Glomus mosseae il the
chosen as the standard trop.
greenhouse.
Experimental desip
Rainfall
A split-plot experimental design was used with eight
The total rainfall before sowing (May 1 to Ju .y 17)
replicates. The main treatments were:
was 170 mm and during the growth cycle (July 1,7 to
October 10, harvest time) it was 692 mm. The r;.infall
1. Inoculation with the ineffective Rhizobi~4rn
distribution was fairly even, without any rnarkc,d dry
strain; N fertilizer: 90 kg N ha” (I-90 N)
spell.
2. Inoculation with the effective Rhizohium strain;
starter N fertilizer: 17 kg N ha-’ (R-17 N)
F. Ganry et al.: Glomus mosseae improves NJ fixation by soybeans
17
3. Inoculation with the effective Rhizobium strain
Results
and Glomus mosseae; starter N fertilizer:
17 kg N ha-’ (RG-17 N)
Pot experiment
The subretitments were:
Table 1 shows that soi1 C had a lower VAM infection
1. No P addition (OP)
potential than soils A and B; thus soi1 C was chosen
2. P added as supertriple, 22 kg P ha-’ (super)
for the field experiment. The main characteristics of
3. P added as Taiba rock phosphate, 22 kg P ha-’
the soi1 C (O-20 cm horizon) were as follows: sand
(rock P)
(20-2000 pm), 83% pH H?O (1:2.5), 6.2; organic C,
0.40%; organic N, 0.038%; total exchangeable ca-
Each main plot (40.25 m’) was divided into three
tions, 1.29 mEq 10.‘g; total exchange capacity,
subplots (12.25 m’). In each subplot an area of
1.66 mEq l@‘g; total P. 197 ppm; available P
6.10 m’ was used for yield estimation and an area of
(Truog), 7 ppm. One should note that soi1 C differed
1.65 m’ was used for the 15N labeling. Al1 the experi-
only from soils A and B in the fact that the root
mental plots were fertilized with KCl at the rate of infection by native VAR4 fungi was significantly lo-
90 kg h-l. Labeled N fertilizer was applied as
wer at the 20th and 38th days, but was the same when
(‘SNH4)rSOJ with 1.01 atom percent ‘“N excess for
the plants wer 60 days old. In a11 soils. plants respon-
the 90 kg N ha-’ application and 4.73 atom percent
ded similarly to the inoculation with Glomus
‘“N excess for the 17 kg N ha-’ application.
mosseae, indicating that no limiting factor occurred
Analysis of the plants
in preventing the establishment of G. mosseae.
Plants were carefully harvested avoiding contami-
Table 1. VAM infection frequency of soybeans grown in
nation with soi1 N; leaves, stems, husks, and grains
nonsterile soils A, B, and C from the Sefa experimental
were sampled and analyzed separately. The samples
station (preliminary pot experiment)
were dried at 6S”-70°C for 24 h, weighed, ground into
a 40-mesh powder, and anaiyzed for total N content
VAM infection frequency (%) at
according to the Kjeldahl method. ‘“N analyses were
Soils
Treat-
20th day
38th day
40th day
carried out at the Seibersdorf Laboratory (IAEA)
ments
using Dumas’ method (the combustion performed in
A
0
9
17
5 5
this technique converts total N directly to NJ and
G
14
19
57
emission spectrometry. For the sake of simplification
B
0
19
20
55
figures related to leaves. stems, and husks were
G
18
25
54
pooled under the term shoot, but data related to
C
0
4
12
56
G
18
24
61
grains were presented separately.
The amount of N? fixed was evaluated according to
0, no inocultion; G, inoculation with G. mosseae
the A value method (Fried and Broeshart 1975).
Each value is the mean of the results from tïve replicates
Root samples from each treatment were stained
with trypan blue in lactophenol using the method of
Phillips and Hayman (1970).
Field experiment
Frequency and intensity of VAM infection were
then assessed according to Ollivier et al. (1983). The
Infection by VAM. Table 2 shows that there was no
interpretation of the whole set of data was performed
significant interaction between the main treatments
according to the test proposed by Quidet and Mas-
and the subtreatments. The only significant main
mejean (1982), which indicated the level of sigifi-
effect on infection frequency was that of inoculation
cnnt (P = 0.05) of the main treatments and their
with Glomus mosseae within subtreatment Super
possible interaction. Within each treatment, sub-
(application of superphosphate): without G. mosseae
‘ireatments were compared and within each subtreat-
inoculation, the infection frequency was 65% and
ment the main treatments were also compared. Re-
with inoculation it was 87% at the 26th day. This
lated LSDs are indicated at the bottom of each table
effect disappeared when plants were older (40th
of data.
day). There were significant effects of G. mosseae
-..
-,
18
F. Ganry et al.: G1omu.s rnosseae improves NJ fixation by soybeans
Table 2. Nodule weight, frequency, and intensity of VAM ir faction of soybean roots 26 and 40 days after in-
oculation witb Rbizabium japonicnm (USDA 138) alone (RI or with Rhizobium japonicum (USDA 138) plus
Glomus mosseae (RG)
--~
-
Treatments
26 day.
40 days
1
~--
Main treatments’
Subtreat-
Nodule
VAM infection
Nodule
VAM infection
ments
-.
-
Inocu-
N ferti-
(P ferti-
dry wtb
Frequency
Intensity dry wtb Frequency Intensity
lation
lizer
lizer)
(%)
w
(96)
(%)
-
-
-
-
R
17N
OP
14
70
25
74
98
42
R
17N
Super
27
65
25
103
96
42
R
17N
Rock P
2 1
68
27
98
96
45
RG
17N
OP
22
73
34
76
91
39
RG
17N
Super
61
87
35
98
92
40
R G
17N
Rock P
45
75
28
87
93
42
LSD between the main treatments
10
6
5
N S
within the same subtreatment
LSD between the subtreatments
1 2
5
4
N S
within the same main treatment
~-
a 17N: N ferrilizer added at the rate of 17 kg N ha-’
b Geometrical mean for dry wt. at the 26th day
inoculation on infection intensity: one wi
RG-17N) and form of P fertilizer (Super versus rock
subtreatment Super (25% versus 35%) and t
phosphate) were observed for (1) grain yield express-
within the subtreatment OP (25% versus 34%).
ed as kilograms dry weight per hectare, (2) grain yield
The coefficient of variation of infection f
expressed as kilograms N per hectare, (3) grain N
was 4.5% for the whole set of plots inocul
concentration, and (4) grain and shoot total N ex-
G. mosseae, which is a much lower figure
pressed as kilograms N per hectare.
for plots not inoculated with G. mosseae (
Harvest index (grainlshoot ratio): Comparing soy-
bean inoculated with Rhizohium USDA 138 and
NJ Fixation. Nodulation (Table 2): The
G. mosseae (treatment RG), we note (Table 5) that
significant interaction between the main treat
in the absence of P fertilizer, dual inoculation (treat-
and the subtreatments. There were two
ment RG) increased only slightly the harvest index
main effects:
expressed as dry weight (+2%), total N (6%). and
1. Inoculation with G. mosseae incr
total P (+9%): in the presence of P fertilizer (Super)
weight of nodules per plant in the ear
the effect of dual inoculation (treatment RG) on
stage (from 21 to 45 mg). This effect
harvest index was much more marked: the increase of
ed when plants were older (40th day)
harvest index expressed as dry weight, total N, and
2. Effect of P addition, whatever form
total P was t>%, 13%, and 22% respectively.
Table 5 shows that inoculation of soybean with
Percentage of N derived from NJ fixatio
Rhizobium USDA 138 (R) instead of the ineffective
There was no significant interaction
strain of Rhizohium (1) improved the harvest index
main treatments and the subtreatments
when expressed on dry weight and total N basis.
two significant main effects:
1. Effect of G. mosseae inoculation on
Concluding remarks: When the level of soluble P in
N derived from NZ fixation within t
the soi1 was raised by adding 22 kg P ha-‘, inoculation
of soybean with
ment Super (R, 69.8%; RG, 75.9%)
Glomus mosseae increased
2. Effect of P addition, whatever form
NZ fixation (+28%). the grain yield (f 13%). and the
harvest index based on dry weight (+12%). The
Grain yield (Table 4): Significant interacti
increase in grain protein resulting from G. mosseae
tween inoculation with G. mosseae (R-17N
inoculation was relatively modest (+16%) but the
1 9
Table 3. Sources of N (expressed as a percentage of total plant N or in kg N ha-‘) in soybean
inoculated with Rhizobium japonicum (USDA 138) alone (R), or with Rhizobium japonicum
(USDA 138) plus Glomus mosseae (KG)
--_.l-l_l~
Treÿtments
Main treatments
Subtreat-
Fertilizer N
Soi1 Nb
Fixed N2
-
ments
Inocu-
N ferti- (P ferti-
%
kgN ha-l
%
kg N ha-l
%
kg N ha-l
lation
lizera
lizer)
-
R
17N
O P
2.6
2.8
30.1
32.8
67.3
73.1
R
17N
Super
2.0
3 . 1
28.2
44.0
69.8
109.0
R
17N
Rock P
2.0
2.9
26.2
39.2
73.8
110.3
RG
17N
OP
2.6
3 . 1
30.6
36.7
66.9
80.2
RG
17N
Super
1.6
3.0
22.5
41.3
75.9
139.3
RG
17N
Rock P
2 . 1
3.2
26.3
40.0
71.6
108.6
LSD between the main treat-
0.40
NS
5 . 3
14.5
ments within the same sub-
treatment
LSD between the subtreat-
0.44
NS
5.7
13.4
ments within the same
main treatment
a Applied at the rate of 17 kg N ha-l
b Calculated from fixation and labeled fertilizer data
Table 4. Grain yield (expressed as k dry wt ha-l or kg total N ha-l), grain N content (%), grain and shoot total
-B
N and total P (expressed as kg ha ) of soybean inoculated with an ineffective strain of Rbizobium japonicum
(I), with Rhizobium japonicum (USDA 138) alone (R), or with Rhizobium juponicum (USDA 138) plus
C~O~US mosseae (KG)
Main treatments
Sub-
Grain yield
Grain N
Grain and
Grain and
ments
content
shoot
shoot
total N
total P
Inocu-
N ferti-
(P ferti-
kg dry wt ha-la
kg N ha-’
W)
kg N ha-l
kg P Ila-’
lation
lizer
lizer)
~-
1
90N
OP
1093 (16)
65.7
6.01
84.2
5.6
1
90N
Super
1725 (15)
101.6
5.89
127.2
11.1
1
90N
Rock P
1482 (11)
86.2
5.81
109.1
8.2
R
17N
OP
1423 (21)
90.3
6.47
112.0
6.0
R
17N
Super
2017 (16)
133.8
6.58
161.0
10.0
R
17N
Rock P
1888( 9 )
124.6
6.62
150.2
8.2
RG
17N
OP
1431(21}
98.2
6.60
120.0
6.9
RG
17N
Super
2290 (11)
154.7
6.76
183.7
11.8
RG
17N
Rock P
1892 ( 7)
126.0
6.69
152.0
8.3
LSD between the main treatments
192
12.8
0.14
16.2
0.79
within the same subtreatment
L S D between the subtreatments
197
13.5
0.14
17.0
0.81
within the same main treatment
---__--
a In parentheses, coefficient of variation (%)
--
--
---
2 0
Table 5. Harvest index of soybean inoculated with an i n : H:ective
have been used successfully on a small scale, but the
strain of Rhizobium japonicum (l), with Rhizobium
0Jzicum
ia. 7’
method would require nearly 2 t ha-’ of soi1 inoculum
(USDA 138) alone (R’j, or with Rhizobium j~p()ZkZ rw (USDA
and the technology of pellets production needs to be
138) plus Glomus mosseae (RG)
-~~-
- --.-
improved (Hayman et al. lY81). Witty and Haymnan
Main treatments
Subtreat-
Harvest index expie ssed
(1978) used the fluid drilling of mycorrhizal soi1
ments
on the bases of - -_- wet-sicvings suspended in methylcellulose. Unfortu-
Inocu-
N ferti-
(P ferti-
Dry
Total
Total
nately such a technique is not suitable when drought
lation
lizera
lizer)
w e i g h t N
,P
conditions coincide with sowing (Hayman et al.
-
.-
-
-
1981).
1
90N
O P
0.33
3.55
3.8
R
17N
O P
0.43
4.18
4.3
In the field experiment reported in this paper we
R G
17N
OP
0.44
4.42
4.7
used the same type of alginate bead inoculum as the
1
90N
Super
0.36
3.98
4.4
one used earlier (Diem et al. 1978; Ganry et al. 1982).
R
17N
Super
0.48
4.71
4.5
The results presented here clearly confirm the conclu-
R G
17N
Super
0.51
5.33
5 . 5
__--.---
. ..-
.-
sion of this previous tria1 that VAM inoculum made
Harvest index: grainkhoot ratio
of spores, mycelium. and homogenized mycorrhizal
a 9ON, 17N: N fertilizer applied at the rate of 90 or 1’ 7 ki;; N ha‘l
roots entrapped in alginate beads cari successfully be
used for large-scale field inoculation. The alginate
absolute value was + 132 kg ha-‘, which is SUI ) it. :mtial
bead inoculum presents the following advantages:
gain if we consider that this figure is equivi 16:nt to
1, Simple preparation of a large amount of inocu-
1300 kg Pearl millet grain (assuming that the1 p1.otein
lum
content of this cereal is 10%). Inter( :st!ingly ,
2. Easy storage and transportation of the inoculum
G. mosseae inoculation reduced the coeffi(xi6 tnt of
to the field
variation of the grain yield expressed in Weight,
3. Easy incorporation into the soi1 at planting time
which confirms previous observations (Gan‘Y et al.
1982). In the 1980 experiment (Ganry et 211. 1982)
when no Super was added, total P content >f grain
Forms of’ phosphate to he used if2 combination with
and shoots expressed as kg P ha-’ was not affl:Cfed by
VA M inoculatiorz
G. mosseue inoculation since it was 6.9 in th 1 ,1.’ mino-
culated plots and 63 in the inoculated ones, t U’! when
A number of greenhouse experiments have shown
Super was added total P content was raised fi,011 n 10.6
that in sterile P-deficient soils addition of soluble P UP
in the uninoculated plots to Il .2 in the in1 )Cl Ilated
to an appropriate level is required to obtain a signi-
ones. Similar effects were observed in the 198 2,E :xper-
ficant response to VAM inoculation (Gianinazzi-
iment (present report), the related figure.S heing
Pearson and Diem lY82). In other words, in these
respectively 6.9 (uninoculated plots) and 6.0 ( iqlc Jcula-
soils we cari expect a positive interaction between
ted plots) in the absence of Super, and 10.0 (, Ir,.i nocu-
VAM inoculation and fertilization with soluble P. but
lated plots) and 11 .X kg P ha-’ (inoculated ~P’C )ts) in
such an interaction does not occur if soluble P is
the presence of S’uper.
replaced by insoluble P except when insoluble P
/
(rock phosphate) is solubilized subsequently tu its
application. These conclusions were confirmed in the
Discussion
field by a previous study (Ganry et al. 1982) and
reconfirmed here. thus it cari be claimed that, in some
Field inocrrlum preparation
tropical P-deficient soils. an appropriate addition of
soluble P may be required to optimize the response to
Although procedures usually propos4 for rflycorr-
VAM inoculation. Besides this concept our experi-
hizal inoculum preparation and field inl)culation
ment indicated that the form of added rock phospha-
have been reviewed by Hayman et al. (1981) and
te should be carfullyconsidered. Taiba rock phospha-
Menge and Timmer (19X2), only a few rep rts have
te that was used in our experiment was net taken up
dealt with field trials performed with the new forms
by mycorrhizal plants growing in the fie]d aithou&
of inoculum that have been proposed in the past
the soil wx sli&tly acidic (pH 6.2). T&s resu/tis
Years. soi] Pekts tht were devised by HiII1 (197~))
apparentlj nOt consistent with those ohtained by
Mosse et al. (1976), who found that inoculation with
Based on several papers on the effect of mycorr-
VAM fungi greatly improved the utilization of Gafsa
hizal inoculation (Powell et al. 1980: Abbott and
rock phosphate in some acidic soils (pH, 5-6.4). The
Robson 1978; Owusu-Bennoah and Mosse 1979), it is
reason for the discrepancy is probably that the rock
suggested that inoculated plants benefit especially
phosphate w’e used was very resistant to solubiliza-
from an early mycorrhizal infection compared with
tion or that some limiting factors occurred hindering
uninoculated ones. Early mycorrhizal infection al-
activity of rhizospheric phosphate-dissolving bacteria
lows the host plants, especially legumes, to increase P
(e.g., unfavorable water regime; lack of energetic
uptake in the first growth stage when P requirements
substrates).
are high. Carling et al. (1979) wrote that soybean
derives maximum benefit from a VAM fungus only if
it receives maximum exposure at the early seedling
Kesponse of field-gros plants to VA M inoculation
stage.
The effect of earlier VAM infection resulting from
In contrast with pot experiments, large-scale field
inoculation upon nodule weight of soybean is obvious
experiments on VAM inoculation have seldom been
when comparing the data of Table 2, subtreatmem
successful. Thus the satisfactory response to VAM
Super (the most favorable to the expression of the
inoculation we obtained here for the second time in
beneficial effect of VAM). In the early growth stage
Senegalese soils should be discussed. Probably the
(26th day) Glomus infection (frequency and intensi-
main cause of this response is that the experimental
ty) and nodule weight in plants inoculated with Glo-
design was set up in a soi1 chosen for its low VAM
mus mosseae (treatment RG-17N, Super) were 87%.
infection potential, but one cannot totally exclude
35%, and 61 mg respectively. whereas the corre-
the hypothesis that the introduced strain of Glomus
sponding figures for plants uninoculated with Rhizo-
mosseae was more effective than the native VAM
bium alone (treatment R-17N, Super) were only
strains and that it was competitive enough to outclass
65%, 25%‘. and 27 mg respectively. Later (40th day)
them.
there were no differences between the treatments.
suggesting that assessment of infection at a late stage
Low irzfkction potential oJ‘ the native VAM popula-
in plant growth is inadequate to ascertain the effect of
tion. Actually this expression refers to the inability of
inoculation. This is in agreement with the suggestion
the native VAM population to infect the plant in its
of Abbott and Robson (1981) that only an early
early growth stage. the late infection being possibly
assessment of infection is valuable in studies of plant
attributed either to the low number of native VAM
growth response to mycorrhizal inoculation.
propagules or to some intrinsic characteristics (such
Whereas the “rhizobiologist” is familiar with the
as low infectivity ability) of these propagules.
idea of evaluating the best locations for successful
Late infection by native VAM populations proba-
inoculation trials, the “mycorrhizologist” very sel-
bly explains why Howeler et al. (1982) found that
dom cares about that problem. which explains most
cassava -- a mycorrhiza-dependent plant which is
failures of VAM inoculation that are a11 the more
most easily infected by VAM fungi - responded
frequent as VAM fungi are ubiquitous organisms.
positively to inoculation only when VAM popula-
From this point of view preliminary inoculation exper-
tions were low. Similarly we obtained a positive
iments in the laboratory of greenhouse for selecting
response of soybean to VAM inoculation in soi1 C
sites of inoculation in the field are recommended and
chosen for its low VAM infection potential (Table 1).
methods for assessing indigenous VAM popuIations
Lt is also necessary to indicate that in our experiment
in soi1 (Porter 1979: Wilson and Trinick 1983) should
inoculation was performed by placing the Glomw
be developed.
rnosseae propagules (entrapped in the alginate be-
ads) in the seedbed (that is close to the germinating
Effectivity and competitive abiiity of the introduced
soybeans), which is a strategic position conferring a
strain of Glomus mosseae. The double concept of
marked advantage to Glomzu mosseae upon the na-
VAM effectivity (or capability to enhance absorption
tive VAM fungi (probably irregularly and widely
of nutrient by the host plant) and competitiveness is
scattered throughout the soi]) and also making possi-
still not very well established. However, a number of
ble an early infection of the soybean roots.
experiments support this concept. Thus Barea et al.
“~I”~‘,,,l~,ir---
_~.-
-.--
-.-
22
F. Ganry et al.: Glomus moueue improves NI fixation by soyheans
(1980) and Kucey and Paul (1982) have shown tha
Influence of the distribution qf prec-ipitation
Glomus mosseae improved P uptake more than indig
enous VAM fungi.
When comparing the data obtained in 1980 (Ganry et
Similary Powell et al. (1980) found that indigeno
al. 1982) and 1982 (present note) in P-fertilized plots
VAM .fungi were ineffective in many soils and t
inoculated with Clornus mosseae, we see that the
inoculation by more effective VAM fungi wo
amoung of NZ fixed was more than twice as high in
result in positive responses even in nonsterile so
1982 (139.3 kg ha-‘) than in 1980 (63 kg ha”). Similar-
containing a high indigenous VAM population,
ly the percentage of plant N derived from NI fixation
Kucey and Paul (1982) related the effectiveness of
was much higher in 1982 (75.9%) than in 1980
the introduced VAM fungus in relation to its CO
(41.4%). These differences cari be attributed to the
tibility with the host. They have shown that
harmful influence of drought periods on N- fixation
beans growing in a field previously planted wi
that occurred in 1980, a year when the distribution of
wheat would benefit from inoculation wi
precipitation was very irregular.
G’. mosseae because indigenous VAM species spo
By contras& in the P-fertilized plots the increase in
taneously selected by wheat may not be efficient f
N2 fixation resulting from Glomus mosseae inocula-
promoting growth of Faba beans. Furthermore,
tion was identical in the 1980 (+29 kg ha“) and in the
ling and Brown (1980) and Schenck and Smith (1
1982 (+30 kg ha-‘) experiments. Similarly the yield
have found that G. mosseae was more stimulatory tc
increase due to Glomus mosseae inoculation was
soybean than many other species of VAM fungi an
roughly the same in 1980 (-t-302 kg dry weight ha“)
response of soybean to G. mosseae inoculation wa
and 1982 (t273 kg dry weight ha-‘).
particularly marked at a high temperature (36°C).
In 1982, the beneficial effect of VAM inoculation
Since our experimental design was set up in a SO
was attrihuted to the fact that it had induced early
which had lain fallow for many years, we cari assum
infection in the roots of soybean grown in a soi1 with a
that the introduced strain of C/ornus mosseue w
low VAM infection potential whereas in 1980 the
more adapted to soybean than the native VA
inoculation effect was mostly attributed to the fact
microflora. However, this speculation should
that during drought periods, which had occurred
ther explored, using reliable experimental tools
throughout the growth cycle of soybean, VAM pro-
as labeled strains of VAM and “‘P-labeled fertih
tected the host plants against the harmful effects of
In any case, data from the literature and this p
water stress (Sprent 1979; Gianinazzi-Pearson and
suggest that G. mosseae could be one of the
Diem 1982).
VAM fungi promoting maximum growth of soybe
in different soi1 types and climatic conditions.
Acknowledgments. We thank Youssouph Ndiaye and Moussa
Niang for technical assistance. The investigation was funded in
inoculation versus N fertilization
part by IAEA (joint FAOIIAEA coordinated Research Pro-
gramme. Contrat n”DI-SEN 2375).
Inoculation of soybean with effective strain Rhi;
bium USDA 138 increased the yield and grai
References
ratio more than fertilizer application. The a
N? fixed due to Rhizobium inoculation alone
Abbott LK, Robson AD (1978) Growth of subterranean clover in
estimated to be ca 109.0 kg N ha-‘; when soy
relation to the formation of endomycorrhizas by introduced
inoculated simultaneously with Rhizobium
and indigenous fungi in a field soil New Phytol 81575-585
mus mosseae, N? fixation was 139.3 kg N ha.‘,
Abbott LK. Robson AD (1981) Infectivity and effectivcness of
is the highest figure reported in a field expe
vesicular-arhuscular mycorrhizal fungi: effcct of inoculum
Senegal to date. From Tables 3 and 4 we
type. Aust J Agric Res 32:h31-639
Bagyaraj DJ, ManJunath A, Patil RB ( 1979) Interaction hetween a
that increasing N? fixation made possible i
vesicular-arbuscular mycorrhiza and Rhizohiurn and their ef-
yield without depleting soi1 N content.
fects on soybean in the field. New Phytol X2:141-136
F. Ganry ct al.: Glotnus mo.~sscuc improves N2 fixation by soybeans
23
Barea JM. Escudero JL, Azcon-G de Aguilar C (1980) Effects of
Menge JA, Timmer LW (1980) Procedures for inoculation of
introduced and indigenous VA mycorrhiza fungi on nodulati-
plants with vesicular-arbuscular mycorrhizae in the laboratory,
on. growth and nutrition of Medicago sativa in phosphate-fixing
greenhouse, and field. In: Schenck NC (ed) Methods and
soils as affected by P fertilizers. Plant Soi1 5432X3-296
principles of mycorrhizal research. The American Phytopatho-
Carling DE, Brown MF (19X(I) Relative effect of vesicular-arbus-
logical Society, St. Paul, Minnesota, pp 59-68
cular mycorrhizal fungi on the growth and yield of soybeans
Masse B, Powell CL. Hayman DS (1976) Plant growth responses
(G‘&5w max). Soil Sci Soc Amer J 44:528--532
to vesicular-arbuscular mycorrhiza. IX. - Interactions between
Carling DE, Brown MF. Brown FA (1979) Colonization rates and
VA Mycorrhiza, rock phosphate and symbiotic nitrogen fixa-
growth responses of soybean plants infected by vesicular-arbus-
tion. New Phytol76:331-342
cular mycorrhizal fungi. Can J Bot 57: 1769-1772
Ollivier B. Bertheau Y. Diem HG. Gianinazzi-Pearson V. (1983)
Diem HG, Jung G, Mugnier J, Ganry F. Dommergues Y.R.
Influence de la variété de V@a unguiculutu C. Walp dans
(1981) Alginate-entrapped Glomus mosseae for trop inocula-
l’expression de trois association endomycorrhizienne a vesicu-
tion. In: Proc 5th Nacom. University Laval, Quebec, p 23
les et arbuscules. Can J Bot 61:354-358
Fried M, Broeshart H (1975) An independent measuremcnt ofthe
Owusu-Bennoah E, Mosse B (1979) Plant growth responses to
amount of nitrogen fixed by a legume trop. Plant Soi1 43:
vesicular-arbuscular
mycorrhiza. XI. Field inoculation respon-
707-7 I 1
ses in barley, lucerne and onion. New Phytol 83:671-680
Ganry F. Diem HC, Dommergues YR (1982) Effect of inoculation
Phillips JM. Hayman DS (1970) Improved procedure for clearing
with Giomus mosJeue on nitrogen fixation by field grown
roats and staining parasitic and vesicular-arbuscular
mycorr-
soybeans. Plant Soi1 68:321-329
hizal fungi for rapid assessment of infection. Trans Brit Mycol
Gianinazzi-Pearson V, Diem HG (19X2) Endomycorrhizae in the
Soc 551%-161
tropics. In: Dommergues YR. Diem HG (eds) Microbiology of
Porter WM (lY79) The ‘most probable number’ method for enu-
tropical soils and plant productivity Nijhoff/Junk, The Hague,
merating infective propagules of vesicular-arbuscular mycorr-
pp 2flY-2s 1
hizal fungi in soil Auçt J Soi1 Res 17:5lS-519
Hall IR (1979) Soi1 pellets to introduce vesicular-arbuscular my-
Powell CL, Groters M, Metcalfe D (19X0) Mycorrhizal inoculation
corrhizal fungi into soil. Soi1 Biol Biochem 1 l:R5<-86
of a barley trop in the field. NZJ Agr Res 23: 107-109
Hayman DS, Morris El, Page RJ (1981) Methods for inoculating
Quidet P, Masmejean T ( 1962) Organisation. conduite et interpré-
field crops with mycorrhizal fungi. Ann Appl Biol 99:247-253
tation des essais fumure. Services Agronomiques de la Société
Howeler RH, Cadavid LF. Burckhardt E (1982) Response of
Commerciale des Potasse d’alsace. SEDA, Paris, p 108
Cassava to VA mycorrhizal inoculation and phosphorus appli-
Schenck NC, Smith GS (1982) Responses of six species of vesicu-
cation in greenhouse and field experiments. Plant Soi1 69:327-
lar-arbuscular mycorrhizal fungi and their effects on soybean at
339
four soi1 temperatures. New Phytol92:19-3-201
Islam R, Ayanaba A ( 1981) Growth and yield responses of cowpea
Sprent JI (1975) Nitrogen fixation by legumes subjected to water
and maize to inoculation with C/ornus mosseae in sterilized soi1
and light stresses. In: Nutman PS (ed) Symbiotic nitrogen
under field conditions. Plant Soi1 63:505-509
fixation in plants. Cambridge University Press, London. New
Islam R, Ayanaba A. Sanders FE (1980) Response of cowpea
York, pp 405-420
(I/ign~ unguiculata) to inoculation with VA-mycorrhizal fungi
Wilson JM. Trinick MJ (1983) Infection development and interac-
and to rock phosphate fertilization in some unsterilized Niger-
tions between vesicular-arbuscular mycorrhizal fungi. New
ian soils. Plant Soi1 54: 107-l 17
Phytol 93:543-553
Kucey RMN. Paul EA (1982) Vcsicular-arbuscular mycorrhizal
Witty JF, Hayman DS (1978) Slurry inoculation of VA mycorrhi-
populations in various Saskatchewan soils and the effect of
za. Rothamsted Experimental Station Report for 1977. Part
inoculation with Glomus mosseae on Faba hean growth in
1:239-240
greenhouse and field trials. Can J Soi1 Sci 63:87-95
Lagacherie B. Hugot R, Amarger N (1977) SekCtiOn of Rhizo-
hium japonictim strains on their competitiveness for infection.
Ann Agron 2X:379-390
Received September 20, 1984
-.___--
--c.-
__C_.---m-
-.“ml-n---