South Dakota State University Open PRAIRIE: Open Public Research Access Institutional Repository and Information Exchange
Electronic Theses and Dissertations
1981
Populations of Nitrogen-Fixing Bacteria in Soil and Rhizospheres in Eastern South Dakota
Edward B. Anderson
Follow this and additional works at: https://openprairie.sdstate.edu/etd
Recommended Citation Anderson, Edward B., "Populations of Nitrogen-Fixing Bacteria in Soil and Rhizospheres in Eastern South Dakota" (1981). Electronic Theses and Dissertations. 4022. https://openprairie.sdstate.edu/etd/4022
This Thesis - Open Access is brought to you for free and open access by Open PRAIRIE: Open Public Research Access Institutional Repository and Information Exchange. It has been accepted for inclusion in Electronic Theses and Dissertations by an authorized administrator of Open PRAIRIE: Open Public Research Access Institutional Repository and Information Exchange. For more information, please contact [email protected]. POPULATIONS OF NITROGEN-FIXING BACTERIA
IN SOIL AND RHIZOSPHERES IN EASTERN SOUTH DAKOTA
by
Edward B. Anderson
A thesis submitted in partial fulfillment of the requirements for the degree Master of Science, Major in Microbiology, South Dakota State University 1981
S U:H POPULATIONS OF NITROGEN-FIXING BACTERIA
IN SOIL AND RHIZOSPHERES IN EASTERN SOUTH DAKOTA
This thesis is approved as a creditable and independent investi gation by a candidate for the degree, Master of Science, and is accept able for meeting the thesis requirements for this degree. Acceptance of this thesis does not imply that the conclusions reached. by the candidate are necessarily the conclusions of the major department.
\.7 Dr. Rcfl:!�rt Pengr¥ "Date Thesis Advise<
Dr. T .- R. Wilkinson � Date Head, Microbiology Dept. iii
ACKNOWLEDGEMENTS
I wish to express my sincere thanks to Dr. Robert Pengra for his helpful suggestings during this investigation and manuscript prepara tion.
I thank Dr. Carl Westby for his comments and suggestions for pre paring this report.
I would like to thank Mrs. Jerry Siegel for typing the manuscript.
EBA .iv
TABLE OF COITTENTS
page
INTRODUCTION 1
LITERATURE REVIEW • . • • . • • • • • • • • • • • • • • • • • . • • • • • • • • • • • • . • • • . • • • • • • . 3
MATERIALS AND METHODS • • • • • • . • • • • • • • • • • • . • • • • • • • • • • • • • . • • • . • • . • . • 11
I . Medi a ...... •. . . . . • . • • • ...... •...... • 11
II. Sunnner Project .._...... 13
III. Greenhouse Project . . . . •. . . . •. . •...... • ...... 15
IV. Nitrogen-Fixing Isolates 16
V. --MPN --vs ---Pour ---Plate Counts 18 RESULTS AND DISCUSSION . • • • • • . • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • 19
I. MPN vs Pour Plate Counts ...... � .. ...•...... � 19
II . S unnner Pro j e ct ...... 2 0
III. Greenhouse Project ...... • ...... 27
IV. Nitrogen-Fixing Isolates ...... •...... •...... 29
SUMMARY AND CONCLUSIONS • . • • • . • • • • • • • • • • • • • • • • • • • . • • • • • • • • • • • • • • • 31
REFERENCES . . . • • . . . • • • • • • • • • • • • • • • • • • • • • . • • • • • • • • • . • • • • • • • • • . • • • • 32 V
LIST OF TABLES
Table page
1 Comparison of mpn and pour plate counts of isolate P75 in log phase� optical density 1.1, 5 60 nm ...... 20
2 Extremes of the soil conditions of the sampling site ...... 21
3 Ranges of pH and water content in which the 3 populations larger than 2 x 10 nitrogen fixers per g were -clustered ...... 23
4 Characteristics of nitrogen-fixing isolates ...... 30 vi
LIST OF FIGURES
Figure page
1 Facultatively anaerobic nitrogen fixers versus water content of the soil. Open circles are bacteria per g soil; closed circles are bacteria per g root. The �umbers in parenthesis are populations ( x 10 ) ...... 24
2 Facultatively anaerobic nitrogen fixers versus soil pH. Open circles represent bacteria per g of soil, and closed circles represent bacteria per g of root . N�mbers in parentheses are populations (x 10 ) ...... 25
3 Azospirillum number versus soil pH. Open circles are bacteria per g soil; closed circles are bacteria per g root ...... 26
4 Facultatively anaerobic nitrogen fixers versus soil concentration in crocks planted to wheat (open bar), barley (cross-hatched) and Reed's canary grass (solid) ...... 28 1
INTRODUCTION�
A search for alternate sources of fixed nitrogen for crop produc tion has been caused by the'rising cost of petroleum used in manufac turing nitrogen fertilizers. Nitrogen in its elemental form, N , is 2 quite stable, and all organisms except a few of the prokaryotic micro organisms can utilize nitrogen only in some chemically combined form.
Those microorganisms which can utilize elemental nitrogen are a poten tial source of fixed nitrogen to crops.
Rhizobium is a genus of nitrogen-fixing bacteria which form nod ules on the roots of leguminous plants. This is a symbiotic relation ship; the plants supply photosynthate to the bacteria which use that energy to grow and fix nitrogen for the plant as well as themselves.
These bacteria can fix enough nitrogen to eliminate the need for· nitrogen fertilizer for these crops .
Rhizobia will fix nitrogen only 1n association with nodulated plants. Other bacteria do not need to form this type of relationship to fix nitrogen. However, they will use N only under certain con 2 ditions. Nitrogen-fixing bacteria will use combined form of nitrogen rather than N because of the energy required to reduce N to ammonia, 2 2 the form which 1s incorporated into the cells. This energy may be sup plied by material leaking from. the roots of the plants and would stimu late nitrogen-fixing bacteria in the rhizosphere. The free-living nitrogen-fixers may be able to supply nitrogen for non-leguminous crops 2 if the proper conditions exist in the soil. Finding conditions of the
soil which favor these bacteria is a part of the search for alterna tives to commercial nitrogen fertilizer. 3
LITERATURE REVIEW
In the search for alternate sources of nitrogen for crop pro duction, it is necessary to learn more about the free-living nitrogen
fixing bacteria and the conditions which affect them. These bacteria can use N for their nitrogen source, but do not form nodules in 2 symbiotic relationships with legumes or other plants. Only a few of the prokaryotic microorganisms are able to use nitrogen gas as a source of nitrogen; all other cells, including eukaryotic microbes
must have some form of comb ined nitrogen .
Nitrogen fixation by species of Clostridium and Azotobacter was demonstrated d�ring the last decade of the nineteenth century by
Bertholet (7), Winogradsky (53), and Beijerinck and van Delden (5).
In 1925 Beijerinck (4) reported possible nitrogen fixation by Spirillum lipoferum, and then Skinner (43) in 1928 confirmed reports of nitrogen
fixation by bacteria that are now in the family Enterobacteriaceae .
The nitrogenase enzyme system consists of more than one protein and appears to be similar in all species of bacteria .
To follow activity of these bacteria, convenient methods for measuring nitrogenase activity and estimating population density are 15 needed. The preferred standard for nitrogen fixation is N enrichment 15 from N , since this shows the actual incorporation of nitrogen 2 from.N into its combined form in the cells, but it is expensive to 2 use . Acetylene reduction has become the most popular assay f�_r nitrogenase activity. Acetylene is an alternate substrate for 4
nitrogenase, and ethylene is the sole product of the reaction. These
two gases can be easily separated by gas chromatography. Since it is
. . . . 3 4 · · · convenient, inexpensive, and sensitive. ( 10 to 10 times more sensitive 15 . tah n N incorporation. ), the acetylene-ethylene assay is the one chosen
by most workers. This as say procedure was first produced by Hardy and
Knight (24). Hardy et al. (23) have evaluated the acetylene-ethylene
assay and reviewed its application.
Appreciable nitrogen fixation associated with the roots of grasses was first noticed in Paspalum notatum in Brazil by Dobereiner (15).
The search has been carried into our own agricultural regions of the
temperate climate. . These surveys in the temperate zone have found nitrogen fixation associated with some grasses, but significant activ- ity has been found only sporadically.
In Oregon, Nelson et al . (33) did a survey of natural grasslands and grain crops. They reported more than a dozen grasses which had some acetylene-reducing activity in the rhizos phere. The estimated rate of nitrogen fixation ranged from 0. 03 g N fixed per hectare 2 per day to 37 g N fixed per hectare per day for Agrostis tenuis 2 . (colonial bentgrass). The roost active of these had N fixation rates 2 comparable to those found in the grasslands in California (45), Wis consin (47, 48), and Saskatchewan (35, 51), which is about 25 g of
N fixed per hectare per day. Certain assumptions must be made to 2 extrapolate quantities of acetylene reduced to nitrogen fixed. There 5
must be a conversion factor for moles of acetylene to moles of nitro-
gen. This is usually 3, since c H is reduced only one step to c H , 2 2 2 4 and N is reduced 3 steps to NH . Another assumption is that the en 2 3 zyme activity is linear over 24 hat the rate observed during the
incubation period for the assay. Incubation periods are kept short to
minimize any effect of acetylene on the cells (23).
Pederson and associates (36) processed approximately 1200 samples
of winter wheat roots from 109 sites throughout Nebraska and found
appreciable activity at only two sites, both located near Chappel in
the western part of the state. One wheat variety was sampled at other
locations, but gave no exceptional results. They also found that of
400 grain lines and crosses, grain sorghums exhibited higher nitro-
genase activity than forage sorghums or winter wheats.
Plants can cause various effects on bacteria in the rhizosphere
by the substances exuded from their roots (31, 39, 40, 28). Wheat
and other temperate climate grasses use the C-3 photosynthetic pathway
for CO 2 fixation, while sorghum, corn, and tropical grasses have the
more efficient C-4 photosynthetic pathway . Light intensity is known
to influence the qualitative and quantitative nature of root exudates by affecting the photosynthetic process (41, 2). Nitrogen fixation in
the rhizosphere has been found to increase with light intensity and
day l_ength (19 ,21). This may be one reason why some of the tropical grasses have much higher nitrogen fixation rates than the grasses of 6
the temperate climates (18� 9, 49, 29, 6). There is also evidence that
the various genotypes within a species support different populations
of bacteria or different rates of nitrogen fixation (13, 16, 42, 17).
The nitrogen-fixing bacteria which have been isolated from rhizo
spheres of the temperate grasses have been members of the Enterobac
teriaceae, such as Klebsiella pneumoniae, Erwinia herbicola, Entero
bacter cl oacae, Bacillus species, or Azospirillum. Nitrogen-fixing
bacteria found associated with grasses of the tropics are Azobacter
paspali, Azospirillum, Beijerinckia, and Derxia (9, 22, 30, 51, 36, 27,
52, SO, 3, 26, 14, 20). Azospirillum is the new genus name for
Spirillum lipoferum, which has been split into two species, A. brasi
lense and A. lipoferum (46) .
In order for these bacteria to fix significant amounts of nitrogen,
their population density must be high, and this would require favorable
soil conditions. Kapustka and Rice (28) in their study of acetylene
reduction during successional stages of revegetation of abandoned
fields in Oklahoma found acetylene reduction to correlate with soil
moisture. The analysis of soil samples from Chappel, Nebraska, where
Pederson et al. (36) found nitrogenase activity, revealed very low
amounts of itrogen (less than 1 ug N per g of soil), but
levels of potassium, phosphorous, zinc, organic matter, and pH were normal for that area . Exactly what conditions are rieeded to c use or 7 all ow bacteria to be actively growing and fixing nitrogen 1n the soil and rhizosphere are still unknown.
It 1s possible to develop procedures to estimate the popul ations of some of these bacteria. Direct plate counts can be used to count
Azotobacter because of their distinctive large colonies on the ap propriate nitrogen-free medium incubated aerobicall y. Azotobacter species generally have low populations in the soil and have not been found particularly associated with plant roots except for the Paspalum notatum-Azotobacter paspali relationship. The facultativel y anaerobic bacter ia, Bacillus species and members of the Enterobacteriaceae, must be grown anaerobically, and Azospirillum, a microaerophilic bacterium, must be incubated at a reduced oxygen tension to have functioning nitrogenase. The nitrogenase is destroyed when the cell s are exposed to oxygen. Plate counts from mixed cultures are difficult to do for the facultatively anaerobic nitrogen fixers because microorganisms are efficient enough at absorbing traces of fixed nitrogen from the medium or inoculum to form colonies. Azospirilla can be grown aerobically on plates of the right medium if a trace of yeast extract is added, but with a mixed inocul um such as a soil dilution, it is outgrown by other bacteria.
Okan et al. (34) have described a most probable number (mpn) method for estimating azospirillum populations. This method uses the semi-solid nitrogen free medium of Dobereiner and Day (18). The .low concentration of malic acid (0. 5%) for a carbon and energy source in this medium favors cells of azospirillum (12 ). Some researchers have 8
had better results with even lower concentrations of malic acid, for
example in Egypt (25) where high azotobacter populations ?ccur. Azo
spirillum forms a characteristic thin layer of growth 1 to 4 mm below
the surface of this medium where the p0 is correct to balance the 2 need for oxygen by its aerobic metabolism and the poisoning effect of
oxygen on its nitrogenase. When a culture exhibits this type of growth,
it is checked for nitrogenase activity by acetylene reduction . These
two characteristics are considered presumptive identification of Azo
spirillum sp. , and those culture tubes which have both are counted
positive for growth in the mpn series. The mpn of azospirillum cells
can then be read from mpn tables (34).
Most probable number techniques can also be used for facultatively
anaerobic or anaerobic bacteria. Campbell and Evans (10) described use of Pankhurst or H tubes for this method. A Pankhurst tube is two tubes, one regular length and one short, connected by a cross tube packed with cotton. Both tubes are sealed with serum stoppers. The culture is in the long tube and alkaline pyrogallol to scrub out the oxygen is in the short tube. The cotton filter allows the atmosphere to be sampled through the short side without contaminating the culture.
These tubes work well, but are more elaborate than what is needed for growing facultatively anaerobic bacteria even under nitrogen-fixing con ditions. The facultative anaerobes can scrub the remaining oxygen from the vessel and grow on N if the atmosphere is replaced with culture 2 nitr�gen gas (37). This led to using the medium Pengra and Wilson (37) developed for studying nitrogen fixation by Klebsiella pneumoniae and 9
standard test tubes sealed with serum stoppers for the mpn's of
nitrogen-fixing, facultatively anaerobic bacteria.
Since media are available for selecting the types of bacteria of
interest, the next steps are to plan the dilution series and evaluate
the accuracy of this technique. Cochran (11) discusses the principal
assumptions and some of the mathematics of the mpn method, as well as
planning the dilution series. As the name suggests, this technique is
not a direct count of cells, but is an estimate derived by probability
mathematics from the number of culture tubes showing growth in a series
which are given a measured inoculum from the sample to be treated. This
involves a serial dilution of the sample. Culture tubes are inoculated
from more than one dilution to increase the limits of the rnpn count.
Two factors increase the accuracy of this estimate. The steps of the dilution series can be narrowed, for example, from 10-fold to 2-fold dilutions, or the number of culture tubes at each dilution can be in creased. Narrowing the dilution steps will also narrow the effective
range of the mpn unless culture tubes are inoculated from more dilution
steps. Ultimately, over a given range, the total number of tubes in an mpn count determine its accuracy; closer dilutions with fewer tubes at each dilution has equal accuracy to larger dilution steps with more tubes per dilution. For this study, three 10-fold dilutions and 5 tubes at each dilution were used.
·There are two principal assumptions behind the mpn method (11).
One is that the organisms are distributed randomly throughout the liquid. The other, which is especially important here, is that only one 10 cell is required to initiate growth in� a culture tube. Although scrubbing the trace of oxygen remaining in a culture tube would re quire more than one cell, t�ere are more bacterial cells than just nitrogen fixers in a dilution of soil. These may help to absorb the oxygen.
In this study, numbers of nitrogen fixers were established by these mpn methods in various soil conditions. The purpose was to gather information on what conditions are needed to encourage these bacteria. 11
MATERIALS AND METHODS
I. Media
The strongly buffered medium of Pengra and Wilson (37) was used for the mpn counts of facultative anaerobes. It contains per liter of de- ionized water:
select for the microaerophilic azospirilla, the semi-solid malate medium of Dobereiner and Day (18) was used. It contains the following per liter of water: KH Po , 0. 4 g; K HP0 , 0.1 g; Mgso ·7H o, 0. 2g; NaCl, 2 4 2 4 4 2 0. 1 g; CaC1 , 0.02 g; FeC1 , 0. 01 g; Na Mo0 0, 0. 002 g; sodium 2 3 2 4 ·zH2 malate, 5.0 g; agar, 3. 5 g. These media are free of fixed nitrogen.
Both media were measured into 16 x 15 0 nun culture tubes at 7 ml per tube. Since the tubes of malate semi-solid medium needed air diffusion, they were capped with plastic caps. After autoclaving they were ready for use. Each rack of Pengra and Wilson's medium was covered with aluminum foil for sterilization. After autoclaving these tubes were sealed with sterile serum stoppers. Working in a hood, serum stoppers handled with ethanol-sanitized forceps were placed on the tubes replacing the foil. The atmosphere in the tubes was then changed to nitrogen gas by evacuating to 0.15 atm or less and refilling three times. This was done through a 12-port manifold. The barrels of 1 ml tuberculin syringes were filled with cotton, sterilized, and connected by rubber tubing to the manifold. With this apparatus
12 tubes could be evacuated and refilled with nitrogen at once. The 12 serum stoppers were swabbed with 70% ethanol before inserting the syringe needles.
For the mpn procedure, 15 tubes of medium were used for each count by having 5 tubes at 3 ten-fold dilutions. For the tubes inoculated with 0.1 ml, a 1.0 ml syringe was used, and for the tubes receiving 1.0 ml, a 10 ml syringe was used. This way, one syringe could inoculate both media at a dilution without refilling. The serum stoppers were swabbed with alcohol before inoculation, and the syringes were also rinsed with alcohol between san�les. The tubes for Azospirillum cultures were incubated for 48 has described by Okan et al. (34) and the anaerobic cultures were incubated for 5 days to allow for longer lag phases from low inoculum concentrations. Incubation was at room temperature.
Growth of nitrogen-fixing organisms 1n the mpn tubes was checked by assaying for nitrogenase activity by the acetylene reduction method.
One ml of tank acetylene was injected into each tube; this makes approx imately 10% acetylene 1n the head space. Those tubes of Pengra and
Wilson's medium which had enough growth to push the stoppers off by gas pressure from fermentation were considered positive. The tubes of
Dobereiner and Day's medium were first examined for typical azospirillum growth which is a clearly defined band 1-4 mm below the surface of the medium. The caps of cultures showing this type of growth were replaced with serum stoppers and 1 ml of acetylene was added. These cultures under.acetylene were incubated overnight (12-17 h) before being checked for ethylene production by gas chromatography. The gas chromatograph 13 used for this assay was a Varian Aerogr�ph series 15 20 equipped with a flame ionization detector and a column of alumina, 180 cm x 1. 2 mm. The carrier gas was nitrogen. If ethylene was present, that tube was counted positive for the presence of nitrogen-fixing organisms. Counts of or ganisms per ml were then read from the mpn tables used by Postgate (38) and calculated to cells per g of soil or root.
II. Summer Project
In the summer of 1978 the mpn techniques described were used to estimate numbers of nitrogen-fixing bacteria in the rhizospheres of grasses and soil adjacent to those grasses. This was done from June through August of that year. Samples were collected first from Clay
County on the banks of the Missouri River, then from experimental grass plots in Faulk County, and at the end of the season from Brookings
County. Altogether 71 sites were sampled.
At each sample site duplicate samples of both roots and soil were taken for the counts of bacteria and one larger soil sample was col lected for soil tests. If more than one grass species was present at a site, the dominant or healthiest species was selected for the root samples. The plants were dug with a spade. The soil was shaken from the roots, and the sterns were trimmed just above the crown. If the plant was corn, a section of ·the root system was cut away with the spade so that some of the larger roots with laterals could be collected with- out destroying the plant.
Soil samples were taken with a 2 x 20 cm soil probe. This soil was either from between rows or from within a few feet of the selected
365356 14 grass in uncultivated or pasture areas. This was to avoid having roots in the soil samples. All samples were kept on ice in a styrofoam chest until brought back to the laboratory. The samples to be sent to the university soil testing laboratory were then spread on aluminum pie plates to air dry, and the other samples for bacterial counts were kept refrigerated until processed . Samples were collected in the morning and processed in the afternoon.
The first dilutions of soil were 10-fold and subsequent dilutions
100-fold. From each sample 11. 0 g of soil were placed in a 99 ml water blank and shaken vigorously by hand for 1 min. Then for azotobacter counts these dilutions were allowed to settle for 30 s before making duplicate spread plates at dilutions of 1:20 and 1:100 on Brown's medium
(8). The 10-fold dilutions were then resuspended. Further dilutions were shaken for 1 min between transfers. Mpn tubes for both media were inoculated at dilutions of 1:100, 1:1000, and 1:10,000.
From the rhizosphere samples 2. 0 g of root material with closely adhering soil were suspended in a 99 ml water blank and shaken for 10 min. This was done by a mechanical shaker having a 4 cm throw operating at 200 cycles/min. From this 1:50 dilution of root wash water, a 1:5000 dilution was made. Inoculation of the rnpn tubes was at dilutions of
1:500; 1:5000, and 1:50,000 for both media. No counts were done for azo tobacter populations in the rhizospheres.
To determine water content, about 10 g of soil were weighed in dupli�ate from the same sample used for bacterial counts. The soil was
° weighed into dry crucibles and dried at 80 C for 24 h. Percent water was 15
then calculated as (wet weight - dry weight)/(wet weight) x 100. These
weights were measured to 0.01 g.
The Soil Testing Laboratory at SDSU detennined the texture, pH,
organic matter, potassium, phosphorous, ammonium and nitrate content of
the air-dried samples (44).
III. Greenhouse Project
A black loam was worked through a 1 cm screen and mixed with
washed sand in ratios of 0, 5, 10, 15, 20, 35, and 50% soil. One quart
sealed-bottom crocks were used. Each grass was planted in three repli
cates of each soil/sand ratio. All crocks were over planted and later
thinned to 30 to 40 plants per crock, depending on the size of the
grass plants and germination rate. The crocks were first watered with
Bond's nitrogen-free stock salt solution (1) and all subsequent watering was with deionized water. The crocks were checked daily for moisture
and watered when the surface appeared dry. The grasses were Morex barley, Solar hard red spring wheat, Proso millet, foxtail millet, sudan
grass , creeping foxtail, timothy, Reed's canary grass, smooth brome
grass, intermediate wheat grass, and western wheat grass. Rambler al falfa inoculated with Rhizobium meliloti was a positive control for nitrogen deficient symptoms. These were planted December 1, 1978 and grown in a greenhouse under natural light.
To harvest, the plants in each crock were clipped at the soil sur- face. The total weight and number of plants in each crock were re corded. After clipping the plants, a 2 cm core top to bottom was collected from near the center of each crock. This could be considered 16 a rhizosphere sample since the root mas-ses filled th e crocks. These core samples were refrigerated overnight and processed the next day.
The first dilution of the core samples was 10-fold. A sample was first mixed in its container, then 11.0 g were weighed into a 99 ml sterile dilution blank and shaken for 10 min. Furth er dilutions were
100-fold. These were mixed for 1 min. All mixing was done by a mechan ical shaker. The mpn tubes of Pengra and Wilson's medium were inocu lated at dilutions of 1:1000, 1:10,000, and 1: 100,000. No counts were done for azospi rilla o r azotobacter.
IV. Nitrogen-Fixing Isolates
Nitrogen-fixing strains were isolated from the mpn tubes which had acetylene-reducing activity. Selected cultures were streaked on plates of Pengra and Wilson's medium. These streak plates were incu bated at room temperature in brewer jars or dessicato rs made anaerobic by evacuating and refilling with nitrogen gas three times. Larger colonies were streaked again and incubated under the same conditions.
Colonies picked from these were kept on nutrient agar slants to be characterized later.
These cultures were then checked for nitrogenase activity by acetylene reduction. Duplicate cultures were grown on slants of Pengra and Wilson's medium suppleme�ted with 10% soil extract (1) and 0.01%
r h yeast extract. The slants we e sealed under N2 wit serum stoppers by the procedure described above for preparing the mpn tubes, except that the slants were streaked before being sealed. After 6 day s 10% acetylene was injected over the slants and gas chromatog raphy was done 17
the next day (24 h incubation) for ethylene production. Stock cultures
of the active strains were streaked again on plates of Pengra and Wilson's
medium and then Gram stained to check for purity. After this assay
stock cultures were maintained on Pengra and Wilson's medium, since
many cultures had been outgrown on nutrient agar by non-nitrogen fixers.
Biochemical tests, including ability to utilize some of the Kreb's
cycle intermediates, were done on the isolates. The composition of the
basal medium for testing utilization of the organic acids was
KH 0.75 g 2Po4 Na HP0 6. 25 g 2 4 MgS0 -7H 0 0.10 g 4 2 NH Cl 2. 0 g 4 CaC1 0.01 g 2 Feso -7H 0 0.0005 g 4 2 H 1 liter. 20 Thi s is Monad and Wellman's synthetic medium for Eschericia coli (3 2), with the buffer and magnesium sulfate cut in half to minimize the pre cipitate. The organic acids tested were oxalic, malic, acetic, succinic, and fumaric. Ten g of the sodium or potassium salts of these acids and lactose and glucose were separately dissolved in 100 ml of deionized water. These solutions were dispensed, 1 ml into 9 ml, into the tubes sat of basal medium before sterilization. The fumarate solution was ° ml at 25 C). The final pH's of the urated (solubility 0.7 g per 100 -· s pH 6. 7. The medium media .were 7.3 to 7.5, except the fumarate; it wa medium with 2. 0 g for the inoculum cultures was Pengra and Wilson's 18
ammonium chloride and 5.0 g sucrose per liter. The isolates were in
oculated into 4 tubes of each medium and sterile mineral oil was floated
over 2 of those to make them anaerobic.
The requirement for supplementation of the medium with yeast ex
tract was also noted. No specific test was done to check for this re
quirement; it was observed by visible growth on the slants. Some of the
isolates did not grow well after several transfers, but responded when
the medium was enriched with soil extract. After this observation,
10% soil extract and 0.01% yeast extract were added to the medium to
maintain the stock cultures.
Isolates were numbered by plant root (P) or soil (S) sample,
sample site number, and dilution factor in the mpn count.
V. MPN vs Pour Plate Counts
This experiment was designed to check the consistency of the mpn
technique and its relation to a viable plate count. A culture of
isolate P75, a facultatively anaerobic bacterium, was grown to log
phase (optical density 1.1, 560 nm) in nitrogen free Pengra and Wilson's -10 -medium. A 100-fold dilution series was carried to 10 . This series
was used for 9 replicates of mpn's and 5 replicates of pour plate
counts. The mpn technique was done as described above for the facul
tatively anaerobic nitrogen fixers. Nu trient agar was used for the -6 -7 pour plates. Pour plates were inoculated at dilutions of 10 and_lO , -8 3 -9 -l and the mpn tubes were inoculated at 10 10 , and 10 o. Inc�bation
0 temperature was 27 C. 19
RESULTS AND DISCU SSION
I . MPN vs Pour Plate Coun ts
To check the accuracy of the mpn technique for facultatively anaerobic nitrogen fixers, mpn' s of a known culture were compared to viable plate counts. The primary concern was whether the mpn's would give estimates below the ac tual cell population , since a certain in oculum size is possibly needed to absorb oxygen and initiate growth in the mpn tubes . At the same time it was possible to check if the n�n 's were within the expected limits of con fidence.
The data in Table 1 are the counts from a culture of isolate p75 grown in liquid culture to opt ical density 1. 1. Although the range of the mpn ' s was much greater than the range of the pour pl ate counts,. this was expected , and the average of the mpn's was only 5% less than the average of the plate counts . The percent of error introduced by the number of cells required to initiate growth in the mpn tubes was still smaller than the error in the mpn estimation itself.
Cochran (11) gave the factor for 95% confidence limits to be 3. 30.
This means that the mpn estimate should be 95% certain to lie within 9 that factor from the true population density. Considering 1. 5 x 10 cells per ml to be the true population density and 3. 3 to be the factor for the confidence limits, then 95% on the mpn's were expected to be 9 between 0.5 and 5.0 x 10 cells per ml. All of the mpn's were within this range . 20
Line and Loutit (30) did report a discrepancy between rnpn and plate methods for counting clostridia in soil , and gave the multiple number of cells required to initiate growth as the cause. They did not report comparing the methods with a pure culture. The mpn tech n1que was also different than the one described in this study : the mpn tubes were incubated together in anaerobic chambers, and acetylene reduction was done on sub cultures instead of the original mpn cultures.
Table 1. Comparison of mpn and pour plate counts of isolate P75 in log phase , optical density 1.1, 560 nm
9 7 (mpn) cells / ml x 10 colonies / plate, 10- dilution 1.10 139 1. 30 180 0.80 142 3.50 153 1. 30 142 0. 50 0 . 70 2.50 1. 30 9 9 averages 1.44 x 10 cells/ml 1.51 x 10 cells / ml
9 X 9 ranges 0. 50 - 3. 50 X 10 1. 3 9 - 1. 80 10
II. Su mner Project
Samples were taken from 71 different sites. Rhizosyhere popu- lations generally were larger than populations in the soil , but this
caution , since they are based on weights comparison must be made with -· samp l es taken of different materials . The soil populations were from they were very fine with a probe ; if any roots were in these samples, 21
and could not be avoided. The samples for rhizosphere populations were
roots and the soil that adhered to the roots . The populations of
3 facultative nitrogen fixers in the soil were more than 2 x 10 cells per
g of soil at 8 of the sites. In the rhizosphere the populations of
3 these bacteria were more than 2 x 10 cells per g of root at only 18
3 sites. Koch and Oya (29) found at least 2 x 10 cells of facultative
anaerobes per g of soil where they found significant nitrogenase
activity by acetylene reduction. This was in Hawaiian pasture soils
which are different from those found here.
Table 2 list s the ranges of the soil characteristics. The values of each were distributed fairly evenly over the ranges found except for the percent organic matter and the parts per million amnonia and nitrate nitrogen. The values of percent organic matter were clustered in the
center and ppm of ammonia and nitrate were grouped in the lower half of their respective ranges.
Table 2. Extremes of the soil conditions of the sampling sites
Soil Characteristics Range
pH 6.1 - 8.2
water 5 - 27%
organic matter 0. 6 - 6. 0% 4 - 45 ppm
80 - 600 ppm
p 1 - 45 ppm
texture sand - loam 22
The numbers of bacteria were �lotted against each of the soil characteristics. There were no actual linear correlations that appeared between bacterial numbers and any of the soil characteristics, but bacterial populations in the soil are certainly not dependent on only 3 one variable. Populations less than 2 x 10 per g occurred through the ranges of all characteristics. What became more apparent were the pH and water content at which higher numbers appeared. This is shown in
Table 3 . Populations in both rhizo sphere and soil samples appeared in
3 similar ranges of pH and moisture content. More than 2 x 10 facultative anaerobes per g were found in th e rhizo sphere above 13% water, and in the soil above 25% water (Figure 1). The first group of samples (11 -
17) were not assayed for content of water, but were from the moist sand 3 along the Missouri River, and were among those with more than 2 x 10 cells per g. This group is not plotted on Figure 1. This trend of higher water content favoring nitrogen fixing populations or nitro genase activity has been reported by other researchers (29, 49).
3 Among all populations of facultative anaerobe s above 2 x 10 cells per g were found in alkaline soils with pH's from 7.5 to 8. 2. There were two sites at pH 6. 9 (Figure 2). The pH values of all these soils were in the growth range of these bacteria, but the optimum pH of
Klebsiella is 7. 2. In th e f�elds where Pederson et al. (36) found ex ceptional nitrogenase activity the soil was also alkaline, pH 7.7.
Although the optimum pH of facultative anaerobes in pure culture is _ close to neutrality, in the field of alkalinity or some other conditions related to alkaline pH appears to favor these nitrogen fixers. 23
Table 3. Ranges of p� and water content in which the populations larger than 2 x 10 nitrogen fixers per g were clustered
nitrogen fixer pH % water
facultative anaerobes rhizosphere 7.5 - 8.2 13 - 27 soil 25 - 27
azospirilla 7.8 - 8.2 13 - 23
azotobacter 7.5 - 7. 7 16 - 27
Azospirilla were not found at many sample sites. There were 18
sites with azospirilla in the rhizosphere and 16 sites where they were
found 1n the soil. Eight of these were the same sites. No soil sample
3 had more than 2 x 10 azospirilla per g. The rhizosphere samples 1n th e 3 Missouri River sand (11 - 17) had as many as 18 x 10 cells per g of
root. The range of azospirillum populations can be seen in Figure 3
where they were plotted against the soil pH. These conditions were
generally the same as for the facultatively anaerobic bacteria. How
ever, even in low numbers, azospirilla were found at very few sites
'below pH 7. 5 . The alkaline pH of soil or some factor rel ated to it
seems to favor both azospirilla and facultatively anaerobic nitrogen
fixers. In their isol ation of azospirilla from diverse regions , Tyler et
al. (SO) found azospirilla in only a small per cent of their samples,
and no more often in the rhizosphere of tropical than in temperate
plants. They also found that the band of growth just below the sur
face of the soft agar medium is not a reliable ·characteristic for 24
•(85) • (45) · •(2 9) • (2 5)
(15)0
.-,g' -0 ><8 E c7... C) '-6
10 5[ '- • c4 C • ....>Q) 3 0 • • • et- 2 • -::, • 0 • • 0 0 • � • 0 • LL • 0 0 • 0 0 • • • 0 • e O •g O 0 •o • • • 0 0 o 7 •oo• e2+.L •o'-1,AD,,& oo o•0•o co onoc• �o40 0 :.; - ��I - I I I 10 15 20 25 % H 2 0
Figure 1. Facultatively anaerobic nitrogen fixers versus water con tent of the so il. Open circles are bacteria per g so il; closed circles are bacteria per g root. The numbers 3 in parenthesis aLe populations (x 10 ). 25
e(a5) e(4 5} e(29) e(25) 0(l5 ) 0(l6) • • 0 0 • 0 • • •
0 o • . • 0 • • • • • �2 • • • • • • •• • • • • • • 0 0 •o • 0 • • • 0�0 0• 0 t•o • • • 0 Oo O O O O O O e O O o 0 0 0 0 • e oo 0e oo ,.,o o • eo o O • •o o 0 oooo o o oo o o�"\ - o 0 o ...... 0 6.0 6.5 7. 0 7.5 8.0 8.5 Soi l pH
Figure 2. Facultatively anaerobic nitrogen fixers versus soil pH . Open circles represent bacteria per g of soil, and closed circles represent bacte�ia per go! root. Num bers in parentheses are populations (x 10 ). 26 1 • • s •• t 16 • rt')o 14
)( • 12 E
o> 10 ' 8 0
'- 6 •
4. • 0 2 0 • • o• 0 0 • ��� 0 • • C 0 Q eo• � I I r 6. 0 6.5 7.0 7.5 8.0 8.5 Soi I pH
il pH . Open circles are Figure 3. Azospirillum number versus �o e bacteria per g bacteria per g soil; closed circles ar root. 27 identification. This would af fect the accuracy of the mpn technique used in this study.
Azotobacter were found at 26 of the 71 sites that were sampled.
These were populations in the soil ; no counts were made of azotobacter populations in the rhizosphere samples. At only 4 sites was the azo- 3 b 1 . to acter popu ation more th an 2 x 10 cells per g. The pH an d percent water at these 4 sites were similar to the sites where populations of azospirilla and facultative anaerobic nitrogen fixers were greater than 3 2 x 10 cells per g (Table 3). The 4 sites had a narrow range of pH , from 7. 5 to 7. 7, and the water content varied from 16 to - 27%. From these few sites it appears that azotobacter are affected by similar conditions of pH and water as azospirilla and facultatively anaerobic nitrogen fixers.
All three of these types of nitrogen fixers had populations above 3 2 x 10 cells per g at similar conditions of pH and moisture, but not necessarily at the same sites . These types of bacteria are apparently affected differently by other factors. There was also no correlation between populations in the soil an d rhizosphere of each site for the azospirill a and facultatively anaerobic nitrogen fixers.
III. Greenhouse Project
The populat ions of facultatively an aerobic nitrogen fixers in the crocks of increasing soil-to-sand ratios were plotted versus the soil concentration. The histograms for 3 of the grasses, wheat, barley and
Reed's can ary grass are in Figure 4. The populations represented here 28 6
5 )( E C '- C) 4 '
(/) Q) .c 3 0 '- (1) 0 C: C 2 Q) >
0
:::, 0 0
0 5 10 15 20 35 50 % SO i
Figure 4. Facultatively anaerobic nitrogen fixers versus soil concen tration in crocks planted to wheat (open bar) , barley (cross- hatched) and Reed's canary grass (solid). 29
are the averages of the 3 replicates of each treatment. The only 3 populations much greater than 2 x 10 cells per g were found in the
crocks of wheat and barley with low soil-sand ratios . There were large
variations amo ng the replicates with high averages, but these were the
only crocks where high numbers of these bacteria appeared. The popu
lations in the crocks of the other grasses followed plots similar to
the one for Reed 's canary grass, increasing gradually with the percent
soil in the crocks.
The wheat and barley had the least increase in plant weight with
increase in soil-sand ratios, which could be caused by the nutrient
reserve in their relatively large seeds.
IV. Nitrogen- fixing Isolates
All of the isolates were facultatively anaerobic, �ram negative
rods. By the fermentation of the carbohydrates and utilization of
organic acids, the 17 isolates fit into 3 main groups (Table 4). Eight
of the isolates (group 1) produced gas from the fermentation of glucose ,
sucrose, and manitol. This group can also use citrate, malate, acetate,
•and succinate as their sole carbon source. Two of these required soil
extract. Of the remaining isolates, some coul d use the organic acids
(group 2) and the rest could not (group 3). Groups 2 and 3 produced
acid but not gas from fermentation of the carbohydrates . All of g roup
3 required soil extract for growth. Only two isolates , one in group 1
and the other in group 2, could grow with only fumarate as its c;rbon
source and none showed growth on oxalate. 30
Table 4. Characteris tics of nitrogen-fixing isolates
* Group 1 ( 8) and Characteristic K. pneumoniae Group 2 (5) Group 3 ( 4) fennentation (glucose, sucrose s manitol) AG A A ut ilization of organic acides as sole C source (cit rate, mal tc , + + acetate, succinate) requirement fr + soil extract * The number of isol ates 1n each group are in parentheses. 31
SUMMARY AND CONCLUSIONS
Populations of three diffe ent types of nitrogen-fixing bacteria
were surveyed at 71 sit es in eastern South Dakota. The azospirilla
and facultatively anaerobic nitrogen fixers were counted in soil and
rhizosphere samples of each site and azotobacter were counted from the
soil samples. These populations were compared to various soil con
ditions. Adequate moisture and alkaline pH or some factor which in
fluences th e pH were found necessary for these bacteria to grow in large
numbers. Thus these factors are necessary for the bacteria studied here
to be a pot e ntial source of nitrogen for crops. However, alkaline pH
and adequat .. moi sture do not assure high populations of nitrogen fixers.
Those were the two conditions which were found consistently with the
higher populations. Other conditions which affect these bacteria must be
found before they can be manipulated for a significant source of nitrogen
for crop production.
For further investigation, control led irrigation on a soil with pH
7.5 - 8.2 would be needed to determine the optimum moisture content.
,When adequate moisture for that particular soil can be assured then the
optimum pH may be determined and other factors such as ground temperature
and plant species coul d be investigated. 32
REFERENCES
1. Allen, 0. N. 195 1. Experiments 1n soil bacteriology. Burgess Pub 1ish ing Co. , Minneapolis.
2. Barber, D. A. , anc J. K. Martin. 1976. The release of organic substances by cereal roots into soil. New Phytol ogist. 76: 69-80 .
3. Barber , Lynn E. , and H. J. Evans. 1976 . Characterization of a nitrogen-fixing bacterial strain from the roots of Digitaria sanguinalis. Can. J. Microbial. 22: 254-60 .
4. Beij eri ck, M. W. 1925. Uber ein Spiril lum welches freien Sticko ff binden kann? Zentralbl. Bakteriol. Parasitenkd. Infectionskr. Hyg ..bt . 2. 63:353-359.
5. Beijeriuck J M. W. and A. van Delden. 1902. Uber die assimilation des freien Stickstoffs 6. van Berkum, P., and J. M. Day. 1980 . Nitrogenase activity associated with soil cores of grasses in Brazil. Soil Biol·. Biochem. 12: 137-140 . 7. Bertholet, M. 1894. Novelles recherches sur les microorganismes du sol fixatears de l ' azote. Bol. Soc. Chim. France. (ser. 3) 11 : 7 84- 7 9 3 . 8. Brown, M. D. , S. K. Burlingham, and R. M. Jackson. 1962. Studies on Azotobac ter species in soil . I. Comparison of media and techniques for counting Azotobacter in soil. Plant Soil. 17: 309-319. 9. von Bulow, J. F. W. , and J. Dobereiner. 1975 . · Potential for nitrogen fixation in ma ize genotypes in Brazil. Proc. Natl. Acad. Sci. U. S. A. 72: 2389-2393 . 10 . Campbell, N. E. R., and H. J. Evans. 1969. Use of Pankhurst tubes to assay acetylene reduction by facultative and anaerobic nitrogen-fixing bacteria. Can. J. Microbial. 15 :1342-1343. 11 . Cochran, W. G. 1950 . Estimation of bacterial densities by means of the "most probabl e number". Biometrics. 6: 105 . 12 . Day, J. M. and J. Dobereiner. 1976 . Physiological aspects of N -fixation by a Spirillum from Digitaria roots. Soil. Biol. Bfochem. 8:45-50. 33 13. Day, J. M. , M. C. P. Neves ., and J--; Dobereiner. 1975. Nitro genase activity on the roots of tropical forage grasses. Soil Biol. Biochem. 7: 10 7-112. 14. Dobereiner, J. 19 1. N"trogen fixing bacteria of the genus Beijerinckia Derx in · the rhizosphere of sugar cane. Pl. Soil. 14:211-217. 15. Dobereiner, J. 196" . Azotobacter · paspali sp. n. -, uma bacteria fixadora de nit ·ogenio na rizosfere de Pa� palum. Pesq. agropec. bras. 1: 35 7-365 . 16. Dobereine r, J. 1 9 70 . Fu rther research on Azotobact e r ·paspali and its variety specific ccurence in the rhizosphe re of Paspalum not a t um Fl ugge. Zentbl . Bakt . Parasitkde. 11. 124: 224-230. 17. Dobereiner, J. , and V. L. D. Baldani. 1979. Selective infection of m8 i :� e root s by streptomy cin resistant Azospirillum lipo fe rum adoth er ba ct _ria. Can. J. Microbial. 25: 1264-1269. 18. Dobereiner, J. , and J. M. Day. 1976. Associative symbiosis in tropical grasses : characterization of microorganisms and N fixing sites,. p. 518-538. In W. E. Newton and C. J. Nyman ted. ) Sympo ium on nitrogen fixat ion, vol. 2. Washington Stat e Uni versity Press, Pullman. 19. Dobereiner, J. , J. M. Day , and P. J. Dart. 1972. Nitrogenase activity and oxygen sensit ivity of the Paspulum notat um Azot obact er paspali association . J. Gen. Microbial. 71: 103- 116. 20. Dobe riner, J., J. M. Day, and P. J. Dart. 1972. Nit rogenase activity in the rhizosphere of sugar cane and other tropical grasses. Plant Soil. 37:19 1-196. 21. Donrrnergues, Y., J. Balandreau, G. Rinaudo, and P. Weinhard. 19 73. Nonsym biotic nit rogen fixation in the rhizospheres of rice, maize , and different tropical grasses. Soil Biol . Biochem. 5. 83-89. 22. Evans, H. J. , N. E. R. Campbell, and S. Hill. 1972. Asymbiotic nitrogen-fixing bacteria from the su rfaces of nodul es and roots of l egumes. Can. J. Microbial. 18:13-21. 23. Hardy, R. W. F. , R. C. Bu rns, and R. D. Holsten. -. 1973. Applica tions of the acetylene-ethylene assay for measurement of � nitrogen fixat ion. Soil Biol. · Biochem. 5:47-81. 34 24. Hardy, R. F. W. , and E. Knight Jr. 1967. ATP dependant reduction of azide and HCN by N2-fixing enzymes of Azotobacter vinelandii and Clostridium pasteurianum. Biochim. Biophys. Acta . . 139: 69- 90. 25. Hegazi , N. A. , H. A. Amer, and M. Monib. 1979. Enumeration of N fixing spirilla. Soil Biol. Biochem. 2- 11: 437-438. 26. Jensen, V. 195 ?. A new nitrogen fixing bacterium from a Danish watercourse. Arch . Mikrobiol. 29:348-35 3. 27. Jurgensen , M. F. , and C. B. Davey. 1970. Nonsymbiotic nitrogen fixing organisms in acid soils and the rhizosphere. Soils and Fertilize rs. 33 :435-446. 28. Kapustk. , L. A. , and E. L. Rice. 1976. Acety lene reduction (N - 2 fixat i m) in soil and old field su ccess ion in central Oklahoma. Soil Biol. Biochem. 8: 497-503 . 29. Koch, B. L. and J. Oya. 1974. Nonsymbiotic nitrogen fixation in some Hawai ian pa sture soils. Soil Biol. Biochem. 6:363-367. 30. Line, M. A. , and M. W. Loutit. 1971. Non-symbiotic nitrogen fixing orga nisms from some New Zealand Tu ssock-grassland solls. J. Gen. Microbial. 66:309-318. 31. Martin, J. K. 1971. Influence of plant species and plant age on the rhizosphere microflora. Aust. J. Biol. Sci. 24:1143-1150. 32. Monad , J. , and E. Wol lman. 1947. L'inhibition de la croissance et de : l'adaption enzymatique chez les bacteries infectees par le ba·cteriophage. Ann. Inst. Pasteur. 73: 937-956. 33. Nelson, A. D. , L. E. Barber, J. Tjepkema, S. A. Russel l, R. Powel son , H. J. Evans, and R. J. Seidler. 1976. Nitrogen fixation associated with grasses in Oregon. Can. J. Microbial. 22:523- 530. 34. Okon, Y. S. L. Albrecht, and R. H. Burris. 1977. Methods for growing Spirillum lipoferum and for counting it in pure culture and in association with plants. Appl. Environ. Microbial. 33 85-88. 35. Paul, E. A. , R. J. K. Myers, and W. A. Rice. 1971. Nitrogen - fixation in grasslands and associated cultivated ecosyste�s. Plant Soil. Special Vol . 495-507. 35 36. Pederson , W. L. , K. Chakrabarty , R. V. Klucas , and A. K. Vidaver. 1978. Nitrogen fi ation (C H associate with 7oots of winter � 2 2 � � _ wheat and sorghum :m NebrasRa. App l. Environ . M1crob1ol. 35 : 129-1 35 . 37. Pengra , R. M. , and . W. Wilson. 1958. Physiology of nitrogen fixation by Ae robacter aerogenes. J. Bacterial. · 75 : 21-25 . 38. Postgate , J. R. 1969. Viable counts and viabil ity , p. 611-628 . In J. R. No rris and D. W. Ribbons (ed.) Methods in Micro biology . Academic Press Inc. , New York. 39. Rice > E . L. 1964 . Inhibition of nitrogen-fixing and nitrifying bact eria by seed plant s (I). Ecology . 45 : 824-827. 40 . Rice, E. L. 1965 . Inhibition of nitrogen-fixing and nitrifying bacte ria by seed plants (II). Characterization and identifi cat ' on of inhibitors . Physiol . Plantarum. 18:255-268. 41 . Rovira A. D. 1969. Plant root exudates. Bo t. Rev . 35 : 35-57. 42. Shearman, R. C. , W. L. Pederson , R. V. Klucas , and E. J. Kinbacher. 1979 . Nit rogen fixation as sociated with 'Park ' Kentucky blµe grass (Poa pratensis). Can. J. Microbial. 25 :1197-1200. 43. Skinner, C. E. 1928. Th e fixation of nit rogen by Bacterium aerogene s and related species . Soil Sci. 25 : 195-205 . 44 . Soil testing procedures in use at the South Dako ta State Soil Test ing Laboratory. Plant Science Pamph let 415 5 ( South Dakota State Univers ity Agricultura l Experiment Station. July 1980). 45 . Steyn , P. L. , and C. C. Delwiche . 1970 . Nitrogen fixation by nonsymbiotic microorganisms in �ome California soil& . . Environ. Sci. Technol . 4:11 22-1128. 46 . Tarrand , J. J. , N. R. Krieg , and J. Dobereiner. 1978. A taxo nomic study of the Spirillum lipo ferum group , with descriptions of a new genus , Azospirillum gen. nov. and two species , Azo spi rillum lipoferum (Beij erinck) comb . nov . and Azospirillum brasilense sp. nov. Can. J. Microbial. 24 :967-980 . 47. Tjepkema, J. 1975 . Nit rogenase activity in the rh izosphere of Panicum virgatum. Soil Biol. Biochem. 7:179-80 48. Tj epkema , J. , and R. H. Burris . 1976. Nit rogenase activity" associ ated with some Wis consin prairie grasses. Plant Soil . 45 : 81-94 . 36 49. Tow , P. G. , and G. J. White. 1976. Non-symbiotic nitrogen fixation in the rhizosphere of Digitaria smutzii Stent. Plant Soil. 45 : 637-646. 50. Tyler , M. E. , J. R. Milam, R. L. Smith, S. C. Schank , and A. Zuberer. 1979 . Isolation of Azospirillum from diverse geo graphic regions . Can. J. Microbial. 25: 693-697. 51. Vlassak, K., E. A. Paul, and R. E. Harris. 1973. Asses sment of biological nitrogen fixation in grassland and associated sites. Plant Soil. '38:637-649. 52. Wahab, A. M. A. 1975. Nitrogen fixation by Bac illus strains isolated from the rhizosphere of · Amrnophila arenaria. Plant Soil . 42: 703-708. 53. Winogr Jd sky, S. 1893. Su r l'assimulation de l'azote gazeur de l'atn�sphere par les microbes. Compt. Rend. Acad. Sci. 116: 1 385 -1388.