Tải bản đầy đủ - 0trang
II. Methods of Estimating Fixation by Crops
THOMAS A . LARUE AND THOMAS G . PATTERSON
Experiments on legumes growing on artificially impoverished soils are not
uncommon in the literature. The substrate used by Lyon and Bizzell was an
artificial one and for that reason the data they obtained should not be extrapolated
to estimate fixation by farm crops.
Long-Term Nitrogen Balance Studies Using Lysimeters
A lysimeter is an enclosed soil system in which the addition and removal of
nutrients, water, and plant material can be controlled and measured. Although
many designs have been described (Kohnke et al., 1940), the basic construction
consists of a tank or box placed in the ground and filled with soil. Leachate from
the soil in the tank is collected from the bottom of the lysimeter. Lysimeters have
been used in long-term studies on nitrogen balance in crops under different
management systems (Chapman et al., 1949). In addition, lysimeters can be
used to compare and calibrate different techniques for estimating N fixation
under controlled conditions (Williams et al., 1977).
Several disadvantages are inherent in lysimeter studies. They are expensive to
install because they require the excavation of large volumes of soil and must be
constructed of materials resistant to corrosion. Due to their expense, the size and
number of lysimeters, and thus the number of treatments and replications, is
In some lysimeter studies, long-term nitrogen balances have shown unaccounted losses of nitrogen (Collison et al., 1933). This is attributed to volatilization of N from the lysimeters (Chapman et al., 1949; Patwary and Raikavich,
1979), although this assumption has been challenged (Craswell and Martin,
An adjusted measure of fixation by the nitrogen accumulation technique is
obtained when the contribution of soil N to the total N of legumes is estimated.
This correction for the contribution of soil N is obtained by growing a nonfixing
plant in comparison with the N-fixing legume. Total N content of the nonfixing
crop (derived solely from soil N) is subtracted from the total N content of the
N-fixing legume. The difference between the values is assumed to be the quantity
of N derived by N fixation. This procedure is often referred to as the “difference” method (Williams et al., 1977). Three versions of the difference
method are commonly used.
1 . Comparison of a Legume with a Nonlegume
Soil N contribution to a fixing legume is estimated by growing a nonlegume
concurrently with the legume. Annual small grains such as wheat and oats have
HOW MUCH NITROGEN DO LEGUMES FIX?
been used (Rizk, 1966). In studies of forage legumes, however, perennial grass
species are used. This is preferred in studying N fixation in grass-legume mixed
pastures. Sprague (1936) used wheat as a control crop to estimate the soil N
contribution. The wheat contained 38 kg N/ha and the vetch 149 kg N/ha. By
subtraction, he estimated the net fixation of the vetch to be 111 kg N/ha.
Using a nonlegume as a control in the difference method requires two major
assumptions. First, the nitrogen contained in the nonlegume is assumed to arise
solely from the soil N pool. Secondly, the legume and nonlegume will take up
soil N in proportion to the amount available, and differences due to growth
patterns and root morphology will not be significant. The validity of the second
assumption is open to question. No evidence exists that shows the exploitation of
soil N is equal between a legume and its nonlegume control.
The fixation estimated will obviously depend on the arbitrary choice of nonlegume control. Wagner (1954) calculated that the N fixation of ladino clover
was 165 kg/ha when orchard grass was used as a control or 189 kg/ha when tall
fescue was the control.
2 . Comparison of a Legume with a Nonnodulating Legume
Another approach to the difference method utilizes the existence of nonnodulating legume genotypes. Soybean genotypes with nonnodulating characters
have been described (Williams and Lynch, 1954). The nonnodulation character
is controlled by a single recessive gene, labeled rj, . This nonnodulating gene has
been backcrossed to produce near-isogenic lines. The advantages of using nonnodulating lines as a control species are that the growth pattern, root
morphology, and N uptake patterns are assumed to be nearly identical. However,
the nodulated isoline may absorb more soil N than the nonnodulated isoline
(Ruschel et al., 1979). This suggests a synergistic effect of fixed N on N derived
from other sources.
Weber (1966) used nonnodulating isolines of soybean to estimate N fixation
under different levels of applied fertilizer nitrogen. By subtracting the N content
of the nonnodulating isoline from that of the nodulated soybean, N fixation rates
of 14.7-84 kg/ha were calculated.
In soybeans, isolines are available for only a few cultivars. Unfortunately,
near isogenic nonnodulating and nodulating cultivars are not yet available for
other legume species.
3 . Comparison of Inoculated and llninoculated Legumes
Another difference method is a comparison of single cultivars grown on inoculated or uninoculated soil. A variation of this procedure uses separate plots
inoculated with effective or ineffective strains of rhizobium (Nutman, 1973). In
this situation, the identical variety is used for the control and treatment, with the
THOMAS A . LARUE AND THOMAS G. PATTERSON
only variable being the presence or absence of nodules (or effective and ineffective nodules). This method requires a soil free of native rhizobium species
capable of establishing an effective symbiosis with the legume being studied.
Bezdicek et al. (1978) compared inoculated and uninoculated soybeans under
field conditions. Uninoculated soybeans accumulated 76 and 105 kg N/ha in the
2 years, while nodulated soybeans yielded 387 and 368 kg N/ha. By difference,
the net fixation was calculated as 31 1 and 263 kg N/ha. In areas where legumes
are routinely grown, soils lacking the appropriate rhizobium species may not be
available. When using this technique, considerable care is required to prevent
contamination of the uninoculated soils with rhizobium. For these reasons this
procedure has limited application in the routine evaluation of N fixation.
Using the difference method with nonnodulating legumes or uninoculated
plots requires the assumption that the root structure and function are not altered
by the presence of nodules. This assumption has yet to be proven experimentally.
4 . N Fertilizer Equivalence
In some studies, an indirect estimate of nitrogen contribution from a legume to
a nonlegume is used to estimate N fixation. Forage legumes and grasses are often
grown together in pasture systems. The legumes have been observed to stimulate
the productivity of the grass species (Lyon and Bizzell, 1911). If the response of
the grass in the mixed stand is compared to the response of the grass in a pure
stand with different amounts of applied fertilizer N, then an estimate of the N
contributed to the grass by the legume is obtained.
This method has been used in the studies of pasture management systems.
Cowling (1961) calculated the “indirect effect” of clover on grass as the amount
of fertilizer N applied to a grass plot to give a yield equivalent to the grass
component of a clover-grass mix. He found that the effect of clover on cocksfoot
was equivalent to 117-134 kg fertilizer N/ha.
Another variation of the nitrogen contribution technique is to estimate the
value of a legume to a following crop. Green manure crops have often been used
to increase soil N (Kroontje and Kehr, 1956; Kamprath et a l . , 1958; Sprague,
1936). The N added to the soil by the legume crop gives a yield in the successive
crop that is compared to the response to different levels of added N fertilizer.
Kroontje and Kehr (1956) estimated that vetch added the equivalent of 107 kg
N/ha to the soil. This resulted in a significant increase in barley yields following
the vetch crop. Similarly, Sprague (1936) estimated that a winter crop of vetch
was equivalent to 671 kg NaNO,/ha.
In comparing the response of follow crops, urea can be used as the nitrogen
fertilizer (de Souza, 1969) because it contains no other nutritional element that
might contribute an effect. A N fertilizer is subject to leaching and volatilization,
whereas the decomposing forage crop releases N more slowly and efficiently to
HOW MUCH NITROGEN DO LEGUMES FIX?
the nonlegume crop. Therefore the “equivalents N ” is likely an overestimate of
N fixation. The beneficial effects of legumes on nonlegumes are not a direct
measure of N fixation. However, they do show the utility of green manures in
adding N to the soil, and the cost of the fertilizer spared may serve as an
indication of the economic value of N fixation.
C. ISOTOPIC METHODS
The use of 15N isotopes for studying N uptake by plants was recently reviewed
here (Hauck and Bremner, 1976). Very few of the articles reviewed related to the
use of I5N for estimating fixation by legume crops. The technique is likely to
become more common, however, because of the lowered cost of the isotope and
improvements in methodology. These include relatively inexpensive mass spectrometers designed for low-mass compounds (e.g., Micromass) and increased
precision by using double introduction and double collection.
Because it is a direct method, fixation of I5N2 remains the method of choice for
checking the validity of other estimates of fixation (Burris, 1972). The cost of
I5N2 and the volumes that would be required to test replicate plant samples over a
growing season preclude estimates based on that gas. However, the technique
has been used to test assays based on other methods. The methods are essentially
those developed by Burris and Wilson (1957; Bums, 1972). Typically, nodulated roots (Saito et al., 1980) or nodules (Hudd et al., 1980) are briefly incubated in a chamber with a gas phase enriched in 15N1. The fixation is stopped by
the addition of strong acid, the tissue is analyzed for total N (Kjeldahl) and I5N
by mass spectrometry (Smith et al., 1963) or emission spectrometry (LloydJones et al., 1977).
The major limitation to mass spectrometer methods is the high cost involved
for the instruments and the isotopes. The procedures are fraught with potential
errors. Although the proper techniques are well documented (Bremner, 1965), it
remains true that the method is demanding and requires skilled operators.
When N gas is excited by a high-frequency oscillator, the wavelengths of
emitted light depend on the isotope composition. Commercial emission spectrometers, less expensive than mass spectrometers, are available for measuring
15N. The instruments are not as precise as mass spectrometers. Their major
advantages are that they require only small samples (10 p g N) and they are not so
demanding of operators.
I . Methods Based on Isotope Dilution
Nitrogen fixation can be estimated by isotope dilution. In this method the
fixing crop and a nonfixing control are grown in soil to which I5N has been added
THOMAS A. LARUE AND THOMAS G . PATTERSON
as a small amount of labeled nitrate or ammonium (McAuliffe et al., 1958; Legg
and Sloger, 1975). The N in the control plant should have the same 15Ncontent
as the available soil N. The plant obtaining part of its N from the atmosphere will
have less of the isotope. The percentage of total N from fixation is calculated as
at. % 15N excess test crop
at. % I5N excess control crop
) x 100
This technique involves the assumption that the isotope incorporated into soil
N is equally available to both crops. Moreover, if the soil N is low, the nonfixing
crop may grow less well than the test crop and thus not be a comparable control.
Fried and co-workers (1975, 1977) proposed the “ A N value” modification.
Adequate labeled fertilizer N can be added to the soil for the nonfixing control to
promote growth. I5N in fertilizer N is added to soil for test plants if the effect of
fertilizer N on fixation is required. The 15Ncontent of the plants is measured after
harvest. For each crop and treatment, the “AN” value is determined. The A N
value (Legg and Stanford, 1967) is an indication of how much of a nutrient is
available to a crop and is expressed as
AN = B ( l - y ) / y
where B is the amount of fertilizer N added (kilograms N per hectare) and y
is the proportion of N in plant derived from that added:
excess 15N in plant
= excess 15N in fertilizer
The fraction of legume N from fixation is calculated as
Nitrogen fixed (kg/ha) = (AN test plant - A N nonfixing control)
N in crop x excess 15N
N in fertilizer x excess I5N
It should be possible to calculate for each crop and treatment the contribution
made by soil, fertilizer, and symbiotic fixation. Procedural details and potential
sources of error in these methods have been reviewed by Hauck and Bremner
(1976) and Rennie et al. (1978).
3 . Methods Based on Natural Isotope Abundance
There are very slight isotope effects during chemical or biological processes
involving N compounds.
Symbiotic fixation seems to favor a very slight increase of 15N. This can be
determined by growing a plant on N-free nutrient so that all its N arises by
fixation. The isotope discrimination of fixation for soybean has been calculated
as 1.0038 (Bardin, 1977) and 1.0014 (Amarger et al., 1979) but also as 0.999
(Kohl and Shearer, 1980).
HOW MUCH NITROGEN DO LEGUMES FIX?
Denitrification produces N 2 lowered in 15N and the nitrate remaining in
soil slightly enriched. The degree of enrichment is measured as parts per thousand difference from the l5NP4N ratio in a standard (usually atmospheric
nitrogen) (Hauck and Bremner, 1976)
615N = (15NP4N) - (l5NP4N) atm
where one 6
unit equals 0.00037 at. %
A plant obtaining all its N from soil (e.g., nonlegume or nonnodulated
control) will have a slightly enriched I5N relative to the atmosphere. A plant
obtaining N from symbiosis will have a lower I5N composition. The I5N enrichment is not uniform throughout the plant, although it is claimed that the seed
content is representative of the whole soybean plant (Shearer et al., 1980). The
percentage of N from fixation is calculated (Domenach et al., 1979):
% N fixed = 100 x
(% 15N control - % I5N legume)
% I5N control - (% 15N airlb)
where b is the isotope discrimination of fixation.
An appealing advantage of this method of estimating fixation is that it does not
necessitate the purchase of isotopic N to add to soil. The major disadvantage is
that the enrichment may not be uniform in the soil (Rennie et al., 1976).
Karamanos and Rennie (1980) found that the P 5 N dropped with depth in welldrained profiles but was constant to a depth of 5 m in upper slopes. Pedogenic
processes produced marked differences in the 6I5N profile within short distances.
These results extended and confirmed those of previous workers that there is an
unpredictable lack of uniformity of enriched N in soils.
The possible isotope discrimination due to symbiotic fixation may favor the
fixation of I4N. If this is not included in the calculation, the level of fixation will
appear to be higher than it is (Rennie et al., 1978).
Experienced critics (Hauck and Bremner, 1976) view the technique based on
natural abundance as giving only qualitative or semiquantitative information.
However the standard errors of the procedure when applied to field-grown
nodulating and nonnodulating soybeans compared favorably with the errors of
estimation based on N difference (Kohl et al., 1980).
Nitrogenase reduces acetylene to ethylene and, so far, it is the only biological
agent reported to do so. Typically, freshly excised roots are incubated in a
chamber with 1-20% C2H, for 30-120 min. A sample of the gas phase is then
THOMAS A . LARUE A N D THOMAS G . PATTERSON
removed and the ethylene produced is measured by gas chromatography. Innumerable variations of the method have been described (Hardy et al., 1973).
The principal assumption in the method involves the ratio of acetylene reduced
to nitrogen fixed. The reduction of nitrogen to ammonia uses six electrons, while
the production of ethylene requires two. Therefore, the ratio 3 : 1 was originally
assumed, i.e., a mole of ethylene was equivalent to % mole of N 2 reduced. It is
now realized that the reaction of nitrogenase approximates
N2 + 8 H+ + 8e ---t 2 NH,
Protons and acetylene compete for electrons, and, with the amounts of GH,
usually used, only small amounts of H2 may appear. In the intact nodule the
hydrogen may be metabolized by hydrogenase in appropriate strains of rhizobia
and therefore not be detected.
There is no adequate method now for calibrating ethylene formation with
nitrogen fixation, but it seems that the ratio is approximately 4 : 1 for legumes
(Hardy et al., 1973).
The acetylene reduction method has the advantages of sensitivity, speed, and
economy. A detection limit of pmole C,H,/ml gas permits estimation of nitrogenase activity even when only a few nodules are forming. It is possible to
measure 40-80 samples/operator-day.
The operation of a gas chromatograph can be easily learned, and the analysis
may be conducted by semiskilled staff. No expensive reagents are necessary. If
sensitivity is not required, there are inexpensive alternatives to measuring
ethylene (LaRue and Kurz, 1975).
Because the plant is usually destroyed in the assay, acetylene reduction is a
one-time estimate of fixation. The incubation must be done immediately after the
root is harvested. There is a diurnal variation in fixation by many legumes
(Sloger et al., 1975). Much plant-to-plant variability is observed, and the entire
nodulated root may not be recovered from the soil. Therefore an estimate of
fixation over a growing season by a crop requires a mathematical summation of
many frequently obtained assays on replicate plots.
E. OTHERMETHODSOF COMPARING
Several indirect methods have been used for estimating the nitrogen fixing
ability of a legume. These include indices of nodulation-number of nodules,
fresh or dry weight of nodules, and leghemoglobin concentration in nodule or
amount per plant. Within a single cultivar these may be closely related to nitrogen fixation and will continue to be useful in screening rhizobial strains or
scoring the effects of environment on nodulation. We have seen no evidence,
however, that these nodule-related characters can be used to calculate the amount
HOW MUCH NITROGEN DO LEGUMES FIX?
of fixation by crops. Nodules that are pink or red internally are only an indication, and not a proof, of nitrogen fixation (de Souza, 1969).
The concentration of the ureides allantoic acid and allantoin in the shoot is
positively correlated with the amount of fixation by soybeans grown in the
greenhouse under defined conditions (McClure er al., 1980). This suggests that
estimates of fixation might be made by assays, nondestructive to the plant, of N
compounds in the shoot parts. Several laboratories are pursuing this topic. While
there are indications that it may aid in the qualitative ranking of cultivars, the
procedure seems unlikely to measure the amount fixed by a crop.
Ill. ESTIMATES FOR MAJOR CROPS
A. THE FORAGES
Most published estimates of fixation by forages assume that plant N arises only
from N fixation. Excluding these, there remain very few estimates, most of
which are based on long-term lysimeter studies (Table I). Except where noted, all
studies used pure stands of legumes.
Chapman et al. (1949) studied the nitrogen balance of vetch (Vicia villosa)
and sweetclover (Melilotus alba) used as winter cover crops at Riverside,
California, over a 10-year period. Only one lysimeter per treatment was used, due
to their large size (3.04-m diameter) and small number (12). The legumes were
grown in a winter rotation and incorporated into the soil prior to planting a
summer crop of barley or sudan grass. Nitrogen fixation of the legumes was
estimated by the increase in soil N due to the incorporation of the green manure
into the soil. A 10-year average of N added to the soil gave an annual fixation
rate of 140 kg N/ha for sweetclover and 184 kg N/ha for vetch.
Estimates for alfalfa (Medicago sativa), white clover (Trifolium repens), red
clover (Trifolium pratense), and Korean lespedeza (kspedeza stipulacea) came
from an 1 1-year lysimeter study by Karraker et al. (1950) in Lexington, Kentucky. They studied the nitrogen balance in a continuous cropping system with
forage legumes and Kentucky bluegrass (Poa prarensis). The lysimeters were 57
cm in diameter, and each treatment was replicated twice.
Although the authors presented only the total N accumulation over the 1 1-year
experiment, the presence of the Kentucky bluegrass control permits our calculation of the average annual N fixation by the difference method. Averages for
each legume, minus the bluegrass control, are 212, 128, 154, and 193 kg N/ha
for alfalfa, white clover, red clover, and Korean lespedesa, respectively.
Jones et al. (1977) used a lysimeter-difference technique to estimate N fixa-
Estimates of Nitrogen Firation by Forage Crops
R m m o u n t , MN
Collison cr d.,1933
Karralicr cr al.. 1954
Heichel cr d.,I 9 8 1
Karraker cr ol.. 1950
Halliday and Pate, 1976
Karraker er 01.. 1950
S p g u e . 1936
chapman cr al.. I949
Phillip and Bennca. 1978
Joms er al.. 1977
New Brunswick. NJ
New Brunswick, NJ
Giza,E g y p
New Bmnswick. NJ
New Brunswick. NJ
New Brunswick, NJ
chapman e r a / . , 1949
(Trifoliurnrepens var. Ladino)
(Mclibrn a h )
K o m kspedeza
aDuration given in months.
*Winter crop rotation.
Orchard grass-call fescue