MODELING ECOLOGICAL DETERMINANTS OF THE SYMBIOTIC

LIST OF TABLES (continued)

Table Page

4.1 Kendall tau b correlation coefficients for

environmental factors influencing nodule

occupancy by inoculant rhizobia and size of

indigenous rhizobial populations........................73

4.2 Summary of equations to describe the

relationship between total nodule occupancy by

inoculant rhizobia in all trials, number of

indigenous rhizobia, and inoculant application

rate....................................................77

4.3 Competitive success of inoculant strains in

relation to indices of the size and competitive

strength of indigenous rhizobial populations............81

4.4 Relative effectiveness of cowpea nodules crushates

obtained from 3 Maui field soils on 4 legumes that

nodulate with Bradyrhizobium sp.........................85

4.5 Effectiveness of 38 cowpea nodule crushates from

site 1 soil on cowpea and their corresponding

effectiveness on lima bean, peanut, and siratro.........88

4.6 Effectiveness of 37 cowpea nodule crushates from

site 3 soil on cowpea and their corresponding

effectiveness on lima bean, peanut, and siratro.........89

4.7 Effectiveness of 35 cowpea nodule crushates from

site 4 soil on cowpea and their corresponding

effectiveness on lima bean, peanut, and siratro.........90

5.1 Elevation, planting date, days to first flower (R1),

growing degree days, and daylength at (R1), and

average soil and air temperature during crop growth

of soybean and bush bean at 4 field sites on

Maui, HI................................................111

5.2 Effect of N source on crop growth rate during

vegetative and reproductive growth of soybean

and bush bean at 3 sites on Maui, HI....................124

5.3 Effect of N source on N assimilation rate during

vegetative and reproductive growth of soybean

and bush bean at 3 sites on Maui, HI....................125

CHAPTER 1

Dissertation Introduction

Rhizobia are symbiotic N₂ fixing soil bacteria that form nodules on the roots of leguminous plants. The association between rhizobia and legumes results in the biological transformation of atmospheric N₂ to plant protein. The ability of legumes to obtain the N required for their growth and reproduction from both soil and symbiosis sets them apart from other economically valuable crops, such as cereals, that rely solely on soil N assimilation to satisfy their N requirements.

Nitrogen is the most common nutrient limiting plant growth, particularly in the tropics (Atkins, 1986). Increasing yield through application of nitrogenous fertilizers is costly, may have adverse environmental consequences, and is often not a viable option for farmers in developing countries. The legume‑Rhizobium symbiosis has been exploited for many years to reduce dependence on N fertilizers without compromising crop yield (Fred et al., 1932).

Rhizobia are commonly inoculated onto legume seeds prior to planting in the hope of increasing plant protein content and seed yield. Despite improvements in inoculation methods (Boonkerd et al., 1978: Sparrow and Ham, 1983: Jensen, 1987; Torres et al., 1987) and selection of rhizobial strains for increased nitrogen fixation capacity (Kishinevsky et al., 1984), competitive ability (Berg et al., 1988), and ability to withstand environmental stress (Munns et al., 1979; Keyser et al., 1979; Lowendorf, 1980), inoculation frequently fails to increase crop yield.

Several inoculation trials have been conducted to identify the factors that contribute to the success or failure of rhizobial inoculants to improve legume yield (Weaver and Frederick, 1974b; Elkins et al., 1976; Harris, 1979). However, failure to correctly identify or quantify the primary independent variables determining inoculation response has hampered use of these results to generate predictions regarding performance of inoculants under varying environmental conditions.

Symbiotic performance of rhizobia introduced into different environments can be evaluated in several ways: by their ability to increase yield above that of uninoculated crops (inoculation response); their ability to compete successfully both among themselves and with indigenous rhizobia for nodule occupancy; and their ability to promote a yield similar to that of N fertilized legumes. All of these aspects of symbiotic performance are mediated by environmental influences. The objective of this study was to identify quantifiable environmental factors that determine and can be used to predict the symbiotic performance of introduced rhizobia in tropical soils.

Determining need to inoculate is an important consideration in the cultivation of leguminous crops. Often the decision of whether or not to use inoculants is not predicated on any measurable factors of the environment, but divined through analysis of legume cropping history or from previous success in improving yields using inoculants. While these methods may provide a good basis for decision in individual instances, they do little to elucidate the underlying mechanisms that determine inoculation response. Without an understanding of the environmental factors that contribute to achieving a response to rhizobial inoculation, successful use of inoculants will remain a site‑specific phenomenon. The ability to predict locations and legume species that will most likely respond to inoculation will enable decision‑makers to make broader recommendations and direct resources where they are needed most.

Cropping history (Elkins et al., 1976): magnitude and effectiveness of indigenous rhizobial populations (Singleton and Tavares, 1986); soil N availability in relation to legume N requirement (Gibson and Harper, 1985); and environmental constraints, which interact with management inputs to determine legume yield potential and N requirement (Singleton et al., 1985), all significantly influence inoculation response. Therefore, the interaction between these factors should ultimately determine the likelihood and magnitude of an inoculation response (Singleton et al., 1985).

Competition between strains of rhizobia for nodule occupancy is influenced by environmental variables, intrinsic characteristics of the rhizobia themselves, and genetic determinants of the host. Environmental factors reported to affect competition for nodule occupancy include presence of indigenous rhizobia (Ireland and Vincent, 1968; Bohlool and Schmidt, 1973; Weaver and Frederick, 1974a,b), soil type (Damirgi et al., 1967; Ham et al., 1971), temperature (Caldwell and Weber, 1970; Weber and Miller, 1972; Kvien and Ham, 1985; Kluson et al., 1986), moisture (Boonkerd and Weaver, 1982), pH (Damirgi et al., 1967; Dughri and Bottomley, 1983,84), nitrogen availability (McNiel, 1982), and microbial antagonism (Schwinghamer and Brockwell, 1978; Triplett and Barta, 1987). Characteristics of rhizobia that may influence the outcome of competition are host genotype compatibility (Johnson et al., 1965; Caldwell and Vest, 1968; Diatloff and Brockwell, 1976; Materon and Vincent, 1980; Kvien et al., 1981; Keyser and Cregan, 1987), motility and chemotactic responses (Hunter and Fahring, 1980; Wadisirisuk et al., 1989), and ability to attach to host roots and initiate nodule formation (Dart, 1977).

Much attention has been paid to factors that affect the ability to establish inoculant strains in a significant proportion of nodules formed on plants in the presence of indigenous rhizobia. This is due to the concept that successful establishment of strains superior in N₂ fixing ability should lead to yield improvement through inoculation. This perspective presupposes that indigenous rhizobia are symbiotically less effective than inoculant strains. While this has been shown to be true in some cases (Ireland and Vincent, 1968), the average effectiveness of populations of indigenous rhizobia may be comparable to that of inoculant strains (Bergersen, 1970; Singleton and Tavares, 1986). While researchers agree that indigenous rhizobia have a tremendous impact on competition for nodule occupancy by inoculant rhizobia, considerable disparity exists in the literature concerning the influence of other environmental variables.

Several mathematical models have been proposed in the literature to describe and quantify competition for nodule occupancy (Ireland and Vincent, 1968; Weaver and Frederick, 1974a; Amarger and Lobreau, 1982; and Beattie et al., 1989). In all of these models, nodule occupancy by inoculant strains is some function of numbers of indigenous rhizobia and application rate of inoculant strains. None of these models has integrated other environmental factors that may influence the outcome of competition.

Numerous legume crop models have been developed in recent years to try to predict phenology (timing of developmental stages) and yield under varying environmental conditions (Major et al., 1975; Wann and Raper, 1979; Hadley et al., 1984; Hodges and French, 1985; Salado‑Navarro et al., 1986a,b; Sinclair et al., 1987; Jones et al., 1989). Few of these have considered N dynamics. Because N is present in numerous essential compounds, effects of N deficiency on crops are dramatic. Most legume crop models assume that plants have sufficient N for maximum growth. This assumption is not problematic if growth and yield predictions are to be made for crops grown under high management conditions. However, for these models to be of broader applicability and address problems common to crop production in the developing world, the effects of nutrient insufficiencies, particularly N, on crop growth need to be addressed. Developing models that can simulate crop growth under varying sources and supplies of N requires an understanding of the effects of different sources of N on plant development and yield.

The natural diversity in rhizobial population size, soils, and climates present at five sites on the island of Maui, Hawaii was used to examine, under field conditions, the impact of environmental factors on the symbiotic success of inoculant rhizobia in tropical soils. Sites in the University of Hawaii's Maui Soil, Climate, and Land Use Network (MauiNet) (Soil Conservation Service, 1984) provided a unique opportunity to study these relationships as sites lacked indigenous rhizobia for some legumes, but provided a range from less than 1 to more than 3.5 x 10⁴ g^-1 soil for others. The diversity of soils and climates at the MauiNet sites allowed measurement of the impact of varying crop yield potential and soil N availability on the interaction between indigenous rhizobia, legume inoculation response, and competition for nodule occupancy. Effect of N source on growth and development of two legumes was also examined at 4 of the sites. Collection of minimum data sets required to run the crop model, SOYGRO (Jones, et al. 1989), in these trials allowed comparison of field results to model simulations.

The goal of this study was to identify and quantify the primary environmental determinants of legume inoculation response and rhizobial competition for nodule occupancy. And, to use these variables to develop mathematical models that can be used to predict the symbiotic performance of rhizobia introduced into different environments.

CHAPTER 2

Environmental Factors Determining the Inoculation Response of Field‑grown Legumes

Introduction

Inoculation of legumes with exotic strains of rhizobia is a common agricultural practice intended to promote nitrogen fixation and increase crop yield. Despite improvements in inoculation methods (Boonkerd et al., 1978; Sparrow and Ham, 1983; Jensen, 1987; Torres et al., 1987) and selection of rhizobial strains for increased nitrogen fixation capacity (Harris, 1979; Kishinevsky et al., 1984), competitive ability (Berg et al., 1988), and ability to withstand environmental stress (Munns et al., 1979; Keyser et al., 1979; Lowendorf, 1980), inoculation often does not increase plant growth and crop yield.

Plant response to inoculation is determined by a variety of factors. The presence and quality of indigenous rhizobial populations (Ham et al., 1971; Diatloff and Langford, 1975; Boonkerd et al., 1978; Singleton and Tavares, 1986), soil N availability (Sutton, 1983; Gibson and Harper, 1985), soil physicochemical constraints (Holding and Lowe, 1971; Singleton et al., 1985), and climatic conditions (Caldwell and Weber, 1970) all significantly influence our ability to achieve increased crop yield through inoculation.

Population density, effectiveness, and competitive ability are the primary characteristics of indigenous rhizobial populations that affect inoculation response. In greenhouse studies, Singleton and Tavares (1986) demonstrated that statistically significant inoculation responses can be eliminated when there are as few as 20 indigenous rhizobia g^‑1 of soil as long as the population contains some effective strains. Strains within populations of rhizobia differ significantly in their ability to supply the host plant with fixed N (effectiveness) under greenhouse conditions (Singleton and Stockinger, 1983; Singleton et al., 1985; Singleton and Tavares, 1986). Differences in the effectiveness of inoculant strains can also be demonstrated under field conditions as long as the soil is free of indigenous rhizobia (Ham, 1980). In the presence of an indigenous population, however, improved crop yield through inoculation with more effective inoculant strains is difficult to demonstrate (Ham et al., 1971; Diatloff and Langford, 1975; Meade et al., 1985).

Successful competition for nodule sites from indigenous rhizobia has been suggested as one reason for failure to achieve a response to inoculation with elite rhizobial strains (Johnson et al., 1965; Meade et al., 1985; Weaver and Frederick, 1974a,b). Both pot experiments (Bohlool and Schmidt, 1973) and field trials (Weaver and Frederick, 1974b) demonstrated that to achieve nodule occupancy greater than 50%, the inoculant must be applied at a rate per seed at least one thousand times greater than the estimated number of indigenous rhizobia g‑1 soil. However, even when a highly effective inoculum strain forms the majority of nodules, failure to improve yield through inoculation is common (Weaver and Frederick, 1974b: Diatloff and Langford, 1975).

High concentrations of soil N affect response to inoculation by inhibiting nodulation thereby decreasing the proportion of plant N that is derived from N₂ fixation (Gibson and Harper, 1985). Available soil N, therefore, must be less than the legume crop N requirement for an inoculation response to be measured.

Environmental stresses that limit yield potential and hence, the crop N requirement, also affect the nitrogen fixation potential of the symbiotic association (Singleton et al., 1985). Environmental constraints include soil physicochemical factors such as acidity, toxicity, salinity, and low fertility (Holding and Lowe, 1971; Singleton and Bohlool, 1983; Singleton et al., 1985); climatic stresses such as low rainfall, inadequate soil and air temperatures, and insufficient solar radiation (Caldwell and Weber, 1970); insect predation; and disease. Consequently, the ability to improve crop yield through inoculation involves an interaction between soil N availability and other environmental conditions affecting crop yield.

The natural diversity in rhizobial population size and composition present at five sites on the island of Maui, Hawaii (Woomer et al., 1988) was used to examine the role indigenous rhizobia play in obtaining a legume yield increase from rhizobial inoculation. The hypothesis that inoculation response is a function of the size of the indigenous rhizobial population and soil N availability in relation to crop N demand was tested. Sites in the University of Hawaii's Maui Soil, Climate, and Land Use Network (MauiNet) (Soil Conservation Service, 1984) provided a unique opportunity to study this relationship as sites lacked indigenous rhizobia for some legumes, but provided a range from less than 1 to more than 3.5 x 10⁴ g^‑1 soil for other legumes. MauiNet sites also have a diversity of soils and climates which allowed measurement of the impact of varying crop yield potential and soil N availability on the interaction between indigenous rhizobial population size and legume inoculation response. Understanding the role of indigenous rhizobial populations in determining host response to inoculation should help to identify locations where inoculation will succeed in improving crop yield. Such knowledge can help determine where and when to use inoculants, appropriate locations for inoculum production facilities, and their production requirements.

Materials and Methods

General experimental approach. A series of field inoculation trials was installed at five ecologically diverse sites on the island of Maui, HI (Table 2.1), using legume species for which the number of soil rhizobia varied between sites (Table 2.2). Each legume species received three N‑source treatments: (i) uninoculated, no N applied: (ii) inoculated at 1(16‑107 rhizobia per seed; and (iii) fertilizer N applied as urea at a rate of 100 kg N ha^‑1 wk^‑1 beginning at planting for sites 1, 2, 3, and 3a and at week 2 for sites 4, 5 and 5a for a total of 800‑2000 kg N ha^‑1 over the cropping cycle. Yield of the fertilizer N treatment estimated the maximum yield potential of each legume species at each site. The uninoculated treatment measured both soil N available for crop growth and, where present, the effect of native rhizobial populations. Rates of inoculation used ranged from 11 to 68 times recommended farmer rates (FAO, 1984) and represented maximum rhizobial numbers that could be successfully applied to the seed. A non‑nodulating isoline of soybean was also planted at each site to provide a biological measurement of soil N available for plant growth during the cropping cycle. Each site was equipped with a Campbell Scientific CR‑21 micrologger (Campbell Scientific, Inc., Logan, UT) to record climate and soil data.

Chapter 4

Environmental Effects on Rhizobial Interstrain Competition for Nodule Occupancy

Introduction

Competition between strains of rhizobia for nodule occupancy is a complex and controversial area in the study of the legume‑Rhizobium symbiosis. Many environmental variables, intrinsic characteristics of the rhizobia themselves, and genetic determinants of the host contribute to the success or failure of rhizobial strains to occupy a significant proportion of nodules formed under a given set of conditions (for review see bowling and Broughton, 1986).

Environmental factors reported to affect competition for nodule occupancy include presence of indigenous rhizobia (Ireland and Vincent, 1968; Bohlool and Schmidt, 1973; Weaver and Frederick, 1974a,b), soil type (Damirgi et al., 1967; Ham et al., 1971), temperature (Caldwell and Weber, 1970; Weber and Miller, 1972; Kvien and Ham, 1985; Kluson et al., 1986), moisture (Boonkerd and Weaver, 1982), pH (Damirgi et al., 1967; Dughri and Bottomley, 1983,84), nitrogen availability (McNiel, 1982), and microbial antagonism (Schwinghamer and Brockwell, 1978; Triplett and Barta, 1987). Characteristics of rhizobia that may influence the outcome of competition are host genotype compatibility (Johnson et al., 1965: Caldwell and Vest, 1968; Diatloff and Brockwell, 1976; Materon and Vincent, 1980; Kvien et al., 1981; Keyser and Cregan, 1987), motility and chemotactic responses (Hunter and Fahring, 1980; Wadisirisuk et al., 1989), and ability to attach to host roots and initiate nodule formation (Dart, 1977). While researchers agree that indigenous rhizobia have a tremendous impact on competition for nodule occupancy by inoculant rhizobia, considerable disparity exists in the literature concerning the influence of other environmental variables.

Interstrain competition for nodule occupancy has been studied in both the greenhouse and the field from a variety of perspectives: among strains comprising the indigenous population (Caldwell and Weber, 1970; Weber and Miller, 1972; Klubek et al., 1988); between one or several introduced strains and the indigenous population (Read, 1953; Johnson et al., 1965; Ireland and Vincent, 1968; Bohlool and Schmidt, 1973; Weaver and Frederick, 1974a,b; Roughley et al., 1976; Brockwell et al., 1982; Berg et al., 1988; Klubek et al., 1988), and among introduced strains in the absence of an indigenous population (Caldwell, 1969; Kosslak and Bohlool, 1985; Brockwell et al., 1987; George et al., 1987: Abaidoo et al., 1990). Much attention has been paid to factors that affect the ability to establish inoculant strains in a significant proportion of nodules formed on plants growing in soil with indigenous rhizobia. This emphasis on competitive ability of inoculant strains is due to the expectation that successful establishment of strains superior in N₂ fixing ability will lead to yield improvement. This perspective presupposes that indigenous rhizobia are symbiotically less effective than inoculant strains. While this has been shown to be true in some cases (Ireland and Vincent, 1968), the average effectiveness of populations of indigenous rhizobia may be comparable to that of inoculant strains (Bergersen, 1970; Singleton and Tavares, 1986).

Some evidence indicates that, in the absence of indigenous rhizobia, competitive ability is a stable characteristic of rhizobial strains as long as plant growth conditions are agriculturally favorable (Brockwell et al., 1982; George et al., 1987; Beattie et al., 1989; Abaidoo et al., 1990). In other words, that the competition pattern exhibited among several introduced rhizobial strains remains constant as long as the environmental conditions remain within the ecological amplitude (range of tolerance) of the strains in question. Implicit in this concept is that competitive competence may indeed be influenced by more extreme environments, some of which may be within the ecological amplitude of the crop. It is generally thought that crops are more sensitive to environmental adversity than are rhizobia (Lowendorf, 1980), however, certain aspects of competition such as bacterial motility, attachment, and nodule initiation may be more sensitive to changes in the environment than either crops or rhizobia living saprophytically.

Several mathematical models have been proposed in the literature to describe and quantify competition for nodule occupancy. Ireland and Vincent (1968) found that nodule occupancy by inoculant rhizobia (log₁₀) was related to inoculant application rate (log₁₀) and number of indigenous rhizobia (log₁₀) by a multiple linear equation. Weaver and Frederick (1974a) reported a similar relationship between these variables. Amarger and Lobreau (1982) studied the effect of varying ratios of inoculant rhizobia applied to soils containing indigenous rhizobia on nodulation competitiveness of strains of Rhizobium leguminosarum. They found that the ratio of nodules formed by the inoculant strain to nodules formed by indigenous rhizobia was related to the ratio of cells in the inoculum to those in the soil by a power function (y = axⁿ). This relationship was used to quantitatively compare the relative competitiveness of inoculant strains in different soils (Amarger, 1984). Beattie et al. (1989) studied the relative nodulation competitiveness of two strains of R. leguminosarum biovar phaseoli by varying the ratio of their application rates. They found that the ratio of the proportion of nodules occupied by each strain (log₁₀) was linearly related to the ratio of the cells of each strain in the inoculum (log₁₀). They evaluated the competitiveness of strains by comparing the value of the y intercept from each regression equation which they defined as a competitiveness index. They found a modification of the equation was useful for comparing competitiveness of inoculum strains against an indigenous rhizobial population.

In this study, environmental effects on competition for nodule occupancy between several introduced rhizobial strains and indigenous rhizobia and among the introduced strains both in the presence and absence of indigenous rhizobia were investigated. Outcome of competition between indigenous and inoculant rhizobia was described and quantified and the ability of several competition models to predict results was evaluated.

I took advantage of the diverse environments present at

5 well‑characterized sites in the Maui Soil, Climate and

Land Use Network (MauiNet) (Soil Conservation Service, 1984) which provided a suitable database to correlate environmental factors with competition for nodule occupancy in different legumes. Many studies have been done and conclusions drawn regarding competition from sites with relatively narrow ecological amplitudes. The diversity of soils and climates in the MauiNet allowed evaluation of the influence of many environmental variables on rhizobial interstrain competition. Identification of factors that strongly influence the outcome of competition can be used to help match rhizobial strains to particular environments and identify environmental variables that may be manipulated to give the balance of the advantage to inoculant strains.

Materials and Methods

Field inoculation trials. Eight field inoculation trials, using 2‑4 legumes in each trial chosen from among 9 legume species, were conducted at five diverse sites in the MauiNet (Soil Conservation Service, 1984) on the island of Maui, HI. Design, installation, harvest, and analysis of these trials inoculum strains used, inoculation procedure, and determination of nodule occupancy enumeration of indigenous rhizobia; site characteristics: and collection of climatic data have been described previously (Chapter 2).

Assay for the effectiveness of indigenous Bradyrhizobium sp.. Soil was collected from unplanted areas adjacent to the field trials at sites 1, 3, and 4 (Table 2.1). Most‑Probable‑Number of indigenous rhizobia (MPN) was determined on 4 test hosts: Vigna unguiculata, Phaseolus lunatus, Arachis hypogaea, and Macroptilium atropurpureum. Method of soil sampling and MPN determination have been described previously (Chapter 2). A representative sample of nodules was taken from the MPN assays performed on V. unguiculata. Nodules were selected from all dilutions where present. Nodules formed by inoculant strains TAL 644 and TAL 658 (Table 2.3) were used as positive controls. Nodules were surface‑sterilized by immersion in 70% ethanol for 1 minute followed by several rinses in sterile water. Individual nodules were crushed in 0.1 ml of yeast‑extract mannitol broth (YMB) (Vincent, 1970), nodule remnants removed, and 4 ml of YMB added. After 2 days incubation at room temperature, 1 ml of each nodule crushate was inoculated onto each of the 4 test hosts growing in plastic growth pouches (Somasegaran and Hoben, 1985). No less than

7 uninoculated control plants were maintained for each test host. Observations on abundance, size, and interior color of nodules and plant vigor were recorded and leaf chlorophyll content (chl a + chl b) determined 32 days after inoculation (DAI) for V. unguiculata and P. lunatus and 41 DAI for A. hypogaea and M. atropurpureum. Relative effectiveness of the crushates was determined by comparing the chlorophyll content of 6 leaf discs (dia. = 0.635 cm) taken from the most recently fully expanded trifoliate leaf on each of the test hosts (Mirza et al., 1990). Crushates forming nodules on the test hosts were divided into 4 effectiveness groupings: highly effective, effective, moderately effective, and ineffective. Crushates were considered to be ineffective if the chlorophyll content of host plant leaf discs was within the 95% confidence interval for the chlorophyll content of uninoculated control plants. Crushates were deemed moderately effective if leaf disc chlorophyll content was higher than the upper confidence limit for uninoculated control plants, but less than the lower confidence limit for chlorophyll content of plants nodulated by known effective strains. Crushates were termed effective if leaf disc chlorophyll content was within the 95% confidence interval and highly effective if higher than the upper confidence limit for chlorophyll content of plants inoculated with known effective strains.

Data analysis. Kendall tau b rank correlation and multiple linear and stepwise regression analyses were used to evaluate the relationship between nodule occupancy by inoculant strains and details of the environment. Soil variables used in the analyses were: most probable number of indigenous rhizobia; organic C and N content (%); C:N ratio; N mineralization potential (as described in Chapter 3, Table 3.1); sum of the base nutrient ions (meq 100 g^‑1 soil) in the CEC (Ca²⁺, Mg²⁺, K⁺, and Na⁺) ; clay, silt, and sand content (%); P retention (%); bulk density; water holding capacity; and pH. Climate variables used were: mean annual rainfall; maximum, minimum, and average soil temperature at 10 cm for the first 10 days following planting and for the interval between planting and nodule harvest; maximum and minimum air temperature for the first 10 days after planting; average soil temperature at 50 cm during the interval between planting and nodule harvest; and the Julian date of planting.

Significance of differences in interstrain competition for nodule occupancy by inoculant rhizobia was determined by a Chi‑square test for deviation from a 1:1:1 ratio. Significance of differences in nodule occupancy by the two more similar of the three inoculant strains was determined using a paired t‑test. In these analyses, double occupancy by inoculant strains was scored as positive for each strain, therefore, total nodule occupancy exceeded 100 in some cases. However, nodule occupancy by inoculant strains for each legume species was adjusted to total 100 prior to correlation analysis.

Multiple linear regression analysis was performed using the MGLH module of SYSTAT v 4.0 (Wilkinson, 1988). All other analyses were performed using PC‑SAS procedures (Statistical Analysis System for personal computers, SAS Institute, 1986).

Results and Discussion

The influence of environmental factors on competition for nodule occupancy by rhizobia was investigated from 2 perspectives in this study: (i) competition between inoculant and indigenous rhizobia for up to 8 legume hosts grown in 5 environments: and (ii) competition among three select inoculant strains for each legume host grown in the different environments. These aspects of competition for nodule occupancy were differentially affected by factors of the environment.

Competition for nodule occupancy between inoculant and indigenous rhizobia. The influence of environmental factors on total nodule occupancy by inoculant rhizobia could be investigated in detail only for the legumes; lima bean (P. lunatus), bush bean (P. vulgaris), and cowpea (V. unguiculata), because only these species had enough data points across sites that had indigenous homologous rhizobia. For each of these species, maximum soil temperature at 10 cm depth during the first 10 days following planting was most strongly related to nodule occupancy by inoculant strains (Table 4.1). Following maximum soil temperature, the relationship between nodule occupancy by inoculant strains and log₁₀ 1 + number of indigenous rhizobia (LOGR) was the most significant. These variables were inversely correlated for lima bean and cowpea and positively correlated for bush bean. Decreasing nodule occupancy by inoculant strains with increasing number of indigenous rhizobia observed for lima bean and cowpea is consistent with other reports (Ireland and Vincent, 1968; Weaver and Frederick, 1974a). Positive correlation between these variables observed for bush bean may have resulted from presence of highly non‑competitive indigenous populations of R. leguminosarum biovar phaseoli or difficulty in estimating size of the effective population (Singleton and Tavares, 1986).

In agreement with the results of bloomer et al. (1988), LOGR for all three species was significantly inversely related to average and maximum soil temperature at 10 cm depth and positively correlated with mean annual rainfall. In this study, LOGR was also significantly correlated with soil organic C and N content and soil N mineralization potential (Table 4.1). Significance of these correlations most likely reflects the impact of these environmental variables on the ability of indigenous rhizobia to persist at these sites.

Correlation coefficients between environmental variables and nodule occupancy by inoculant strains for lima bean and cowpea were the converse of those observed for LOGR. In agreement with the positive correlation observed between nodule occupancy by inoculant bush bean rhizobia and LOGR, correlation coefficients between environmental variables and bush bean nodule occupancy were similar to those observed for LOGR. In a stepwise regression procedure performed for the dependent variable percent nodule occupancy by inoculant

occupancy by inoculant rhizobia, achieved in these trials across a wide range of environments. Inoculants were applied at realistic economic rates, which, indicates that existing inoculation technology may be adequate for successful nodule establishment of inoculant rhizobia. However, while nodule occupancy by inoculant rhizobia was significantly correlated with percent increase in yield due to inoculation (r = 0.43, p < 0.02), greater than 50% nodule occupancy by inoculant strains did not guarantee a significant yield response to inoculation (Figure 4.2, Table 4.3). This was perhaps due to the high effectiveness of indigenous rhizobial populations. A significant inoculation response was achieved in all trials where the ratio of applied to indigenous rhizobia exceeded 1000 to 1, and in only 2 trials, both with bush bean, where this ratio was less. These results support conclusions reached previously (Chapter 2) that where yield is limited by insufficient soil N, size of the indigenous rhizobial population is the primary environmental factor determining the ability of inoculation to increase yield.

Competitive success of inoculant strains was inversely and significantly correlated (r = ‑0.59, p = 0.001) with the competitive ability of indigenous rhizobial populations as well as their size. Competitiveness of indigenous rhizobial populations can be expressed as the ratio of nodule occupancy by indigenous rhizobia to their number in the soil (percent occupancy by indigenous rhizobia/LOGR). This ratio provides both a measure of the strength of the competition barrier presented by the indigenous population and a means to compare the relative competitiveness of rhizobial populations across sites

Table 4.3 Competitive success of inoculant strains in relation to indices of the size and competitive strength of indigenous rhizobial populations.
Site No.	Legume Species	Log (1+MPN Indigenous Rhizobia)	Indigenous^a Competition Barrier	Nodule Occupancy by Inoculant Rhizobia (%)	Ratio of^b Applied to Indigenous	Inoculation Response (p < 0.10)

1	P.lunatus	0.26	31.6	91.7	6002	*
2		1.79	11.0	80.2	137
3		0.26	23.6	93.8	10016	*
4		2.49	20.4	49.0	18
5		1.38	10.6	85.4	132
1	V.	1.74	19.1	66.7	94
2	unguiculata	3.36	13.6	54.2	2
3		1.28	3.3	95.8	227
4		4.56	11.4	47.9	<1
5		2.45	13.6	66.7	3
1a	A. hypogaea	0.78	88.3	31.3	2479	*
3a		0.78	44.2	65.6	1892	*
1a	L.	3.22	28.8	7.3	2
3a	leucocephala	3.77	24.3	8.3	1
5a	L. tingeatus	1.20	9.7	88.3	130
5a	T. repens	0.26	16.0	95.8	294
1	P. vulgaris	0.90	6.9	93.8	456	*
2		1.97	5.8	88.5	17
3		0.48	35.0	83.3	1009	*
4		2.64	1.6	95.8	24	*
5		1.51	3.3	95.1	44
1	G. max	0	0	100	na	*
2		0	0	100	na	*
3		0	0	100	na	*
3a		0	0	100	na	*
4		0	0	100	na	*
5		0	0	100	na
a b	Percent nodule occupancy by indigenous rhizobia/log (1 + MPN of indigenous rhizobia). Number of inoculant rhizobia applied/MPN of indigenous rhizobia per g soil.

(Table 4.3). For example, at the two sites where peanut was grown (sites la and 3a), numbers of indigenous rhizobia were equal, yet, inoculant strains occupied less than half the number of nodules at site 1a as they did at site 3a. Using the ratio defined above, it can be seen that indigenous rhizobia at site 1a were twice as competitive as those at site 3a and presented a much stronger competitive barrier to nodule occupancy by inoculant strains. Indeed, the Bradyrhizobium sp. population present at site 1 was also more competitive on cowpea and on lima bean than that at any other site (Table 4.3). With the exception of indigenous rhizobia nodulating cowpea, the next most competitive indigenous population was that present at site 3. Environmental conditions at sites 1 and 3 were harsher than at the remaining sites (higher soil temperatures and lower mean annual rainfall) (Table 2.1) indicating that better adaptation to prevailing environmental conditions by indigenous rhizobia may also contribute to their competitiveness. With the exception of site 3, populations of Rhizobium Leguminosarum bv phaseoli presented a comparatively weak competition barrier across sites. This may help to explain the consistently anomalous results obtained with bush bean at these sites (discussed above).

Sub‑groups of the cowpea miscellany, Bradyrhizobium sp.

Considerable diversity exists in the relative effectiveness of populations of indigenous Bradyrhizobium sp. on different host legumes (Singleton and Tavares, 1986). This diversity is reflected in differences in the size of bradyrhizobial populations capable of nodulating homologous hosts and their competitiveness with the

Table 4.5 Effectiveness of 38 cowpea nodule crushates from site 1 soil on cowpea and their corresponding effectiveness on lima bean, peanut, and siratro.

Distribution of effectiveness of 38 crushates on V. unguiculata:

Highly Effective (HE) Effective (E) Moderate (M)

Ineffective (I)

8 21 7

Nodulates Nodulates Nodulates

Yes^a No Yes No Yes No

HE E M I HE E M I HE E M I

Nodulates

Yes No

HE E M I

M. atropurpureum^b

0 0 7 1 0 0 1 11 3 0 0 0 4 1 0

0 0 1 1 0

P. lunatus

1 0 0 5 2 0 3 1 14 3 0 1 0 5 1

0 0 0 1 1

A.hypogaea

0 1 3 4 0 0 3 7 11 0 0 3 1 1 2

0 0 1 0 1

a HE = highly effective; E = effective; M = moderately effective; and I = ineffective.

b Of the total number of crushates, 8 were not tested on this species.

Table 4.6 Effectiveness of 37 cowpea nodule crushates from site 3 soil on cowpea and their corresponding effectivenesson lima bean, peanut, and siratro.

Distribution of effectiveness of 37 crushates on V. unguiculata:

Highly Effective (HE) Effective (E) Moderate (M)

4 17 7

Ineffective (I)

Nodulates Nodulates Nodulates

Nodulates

Yes^a No Yes No Yes No

HE E M I HE E M I HE E M I

Yes No

HE E M I

M. atropurpureum

0 2 2 0 0 0 3 11 3 0 0 2 5 0 0

0 0 3 5 0

P. lunatus

1 0 0 2 1 1 4 3 5 4 0 3 1 2 1

0 1 2 0 6

A.hypogaea

0 1 0 2 1 0 2 2 1 12 0 1 0 2 4

0 1 3 3 2

a HE = highly effective; E = effective; M = moderately effective; and I = ineffective.

Table 4.7 Effectiveness of 35 cowpea nodule crushates from site 4 soil on cowpea and their corresponding effectiveness on lima bean, peanut, and siratro.

Distribution of effectiveness of 35 crushates on V. unguiculata:

Highly Effective (HE) Effective (E) Moderate (M)

6 18 5

Ineffective (I)

Nodulates Nodulates Nodulates

Nodulates

Yes^a No Yes No Yes No

HE E M I HE E M I HE E M I

Yes No

HE E M I

M. atropurpureum^b

0 3 3 0 0 0 9 9 0 0 0 2 1 2 0

0 2 2 1 0

P. lunatus^b

0 0 0 6 0 3 1 0 13 0 0 1 0 4 0

0 2 0 4 0

A.hypogaea^b

0 3 0 1 2 0 6 6 1 4 0 3 1 0 1

0 3 2 0 1

a HE = highly effective; E =effective; M = moderately effective; and I = ineffective.

b Of the total number of crushates, 1 was not tested on this species.

CHAPTER 5

Effect of Nitrogen Source on the Growth and Phenology of Soybean and Bush Bean

Introduction

Soybean (Glycine max) and bush bean (Phaseolus vulgaris) are twoeconomically important grain legumes that are grown in diverse environments throughout the world. Both are able to form symbiotic relationships with the soil inhabiting, N₂‑fixing bacteria, rhizobia, in the groups Bradyrhizobium japonicum and Rhizobium leguminosarum biovar phaseoli, respectively. The symbiosis between these legumes and their homologous rhizobia results in the conversion of atmospheric N₂ to plant protein. The ability of leguminous plants to obtain the N required for their growth and reproduction from both soil and symbiosis sets them apart from other economically valuable crops, such as cereals, that rely solely on soil N assimilation to satisfy their N requirements. Soils are more often deficient in N than in any other element, consequently, N is the most common nutrient limiting plant growth, particularly in the tropics (Atkins, 1986). Increasing yield through application of nitrogenous fertilizers is costly, may have adverse environmental consequences, and is often not a viable option for farmers in developing countries due to its limited availability. The legume‑Rhizobium symbiosis has been exploited for many years to try to reduce dependence on N fertilizers without compromising crop yield (Fred et al., 1932). While yield of symbiotic plants may often be comparable to that of N fertilized plants (Summerfield, et al., 1977; Imsande, 1989; Kucey, 1989), it has been shown that plants relying on soil and symbiotic N for growth may achieve only 80‑90% of the yield possible through N fertilization (Table 2.7; Silsbury, 1977; Ryle et al., 1979). Bush bean, in particular, is notorious for symbiotic inefficiencies (Graham, 1981; Piha and Munns, 1987). This crop can respond significantly to fertilizer N application in low‑N soils in the absence of other limitations to yield, but, yield of symbiotic crops in the same soils frequently falls short of expectations (Figure 2.1 and Appendix 2).

Numerous soybean and a few bush bean models have been developed in recent years to try to predict crop phenology (timing of developmental stages) and yield under varying environmental conditions (Major et al., 1975; Wann and Raper, 1979; Hadley et al., 1984; Hodges and French, 1985; Salado‑Navarro et al., 1986a,b; Sinclair et al., 1987; Jones et al., 1989). Few of these models consider N dynamics. The development, calibration, validation, and refinement of models to predict performance of field crops is an immense undertaking requiring information from, and collaboration between, researchers from many scientific disciplines. These models, by their nature, are simplified representations of real cropping systems that are designed to study, understand, and make predictions about the complex interactions that take place between plants and their environment. Because of the complexity of the cropping system and our inability to measure all variables and their interactions, decisions must be made about which processes will be considered in a model, the detail with which each process is described, and level of interaction between processes. These will ultimately be determined by the purpose for which the model is intended (model objectives). SOYGRO V5.42 (Jones et al., 1989) is one such model. The SOYGRO model was originally designed to predict crop yield as a function of irrigation management, hence, weather, crop genetic potential, and soil water relations have been most extensively modeled.

Because N is present in numerous essential compounds, effects of N deficiency on crops are dramatic. In general, N deficiency causes a reduction in growth rate and general chlorosis, often accompanied by early abscission of older leaves (Salisbury and Ross, 1985). Recently, it has been shown that N deficiency hastens crop maturity in soybean (George et al., 1990). Most legume crop models, including SOYGRO, assume that plants have sufficient N for maximum growth. This assumption is not problematic if growth and yield predictions are to be made for crops grown under high N conditions. However, for these models to be of broader applicability and address problems common to crop production in the developing world, the effects of nutrient insufficiencies, particularly N, on crop growth should be addressed.

When modeling the development and yield of legumes, incorporating subroutines to handle N assimilation are complicated by the need to model the symbiotic process. The metabolic cost of N assimilation differs for root uptake and N₂ fixation primarily due to the high energy requirement of the nitrogenase enzyme and cost involved in developing and maintaining nodule tissue (Imsande, 1988; Lynch and Wood, 1988). Increased cost of N₂ fixation in symbiotic plants may result in differences in developmental and growth rates due to diversion of energy to fix N₂ that might otherwise have been used for growth. In N deficient soils, this cost would be amortized by the benefit derived from obtaining fixed N. Developing models that can simulate crop growth under varying sources and supplies of N requires an understanding of the effects of N source on plant development and yield.

This work was undertaken to investigate the effect of N source on growth and yield of soybean and bush bean. The objectives of this study were to ascertain whether: (i) crops relying on soil, symbiotic, or fertilizer N differed in their growth characteristics: (ii) symbiotic plants developed similarly to N fertilized plants; (iii) any effects of N source on crop development were related to final yield; and (iv) the growth simulation model, SOYGRO V5.42, could accurately predict phenology and yield of soybean grown in different environments.

Sites were selected and dates of planting varied in this study to provide differences in both temperature and photoperiod in order to establish whether any differences in development caused by N source were independent of climatic effects. Well characterized sites, equipped with weather stations to record climatic data, were selected from among those in the Maui Soil, Climate, and Land Use Network (MauiNet) (Soil Conservation Service, 1984) on the island of Maui, Hawaii. Weather, site, and soil information were entered into the SOYGRO crop growth simulation model (Jones et al., 1989) and soybean crop growth was simulated. Predictions of the timing of phenological events, duration of growth phases, biomass accumulation, and seed yield were compared with field data. The model was assessed for its ability to simulate growth under non‑N limiting conditions. The need to consider N nutrition and symbiotic status of leguminous crops in order to generate realistic predictions of crop growth and development was ascertained.

Materials and Methods

Field inoculation trials. Effect of N source on biomass and N accumulation, phenology, and seed yield of soybean (G. max cv Clark IV, P. Cregan, USDA Nitrogen Fixation Laboratory, Beltsville, MD) and bush bean (P. vulgaris cv Bush Bountiful) was assessed in field trials conducted at sites 1, 3a, 4, and 5 (Table 5.1). General experimental approach, soil amendments, planting density, inoculation procedures, enumeration of indigenous rhizobial populations, and early and final harvest protocols have been described previously (Chapter 2).

In these trials, additional biomass harvests were performed at growth stages V4 (4 nodes on the main stem), R5/R6 (mid pod‑fill), and R7 (physiological maturity) (Fehr, et al., 1971) at sites 1, 3a, and 5. For each plot, plants were cut at the soil surface from 3.0 linear m of row (1.8 m²) for the V4 and R5/R6 harvests and from 4.5 linear m of row (2.7 m²) for the R7 harvest. Fresh weight of the sample was determined immediately. A subsample of 10‑15 plants was taken to determine moisture content and a 5 plant subsample taken for determination of leaf area and dry weight of component parts. Fresh weight of both subsamples was taken in the field and average number of nodes on the main stem (V stage) recorded. The larger subsamples from all plots were dried, weighed, ground, and analyzed for N content as described previously (Chapter 2). Leaves were removed from plants in the smaller subsamples and leaf area determined with a Licor LI‑3100 leaf area meter. Leaves and stems were dried at 70 C to constant weight and weighed separately.

Crop phenology and growth analysis. Crop phenology was recorded every few days in the field from emergence to physiological maturity according to the stage of development descriptions of Fehr et al. (1971). Crop growth rate (CGR) and N assimilation rate (NAR) were calculated by dividing the net increase in biomass or N assimilated by the number of days between harvests. Leaf weight ratio (LWR) equalled leaf dry weight divided by total shoot dry weight. Specific leaf area (SLA) was calculated by dividing leaf area (cm²) of the subsample by its leaf dry weight (g). Total leaf dry weight (Lw) (g nit) was determined by multiplying dry weight of above ground biomass by LWR. Leaf area index (LAI) was calculated by multiplying LW by SLA and dividing by 10,000. Seed fill duration was calculated as days to R7 minus days to R4 (Fehr et al., 1971). Growing degrees days (GDD) were determined by taking the sum from sowing to first flower and from sowing to physiological maturity of the mean daily air temperature minus a base temperature of 7.8 C (Hadley et al., 1984).

Experimental design and data analysis. These trials were incorporated into the larger field inoculation trials described in Chapter 2. Trials were part of a split‑plot design with four replications. Legume species were assigned to mainplots and N‑source treatments confined to subplots. All crop growth data were analyzed using the analysis of variance procedures of PC‑SAS (SAS Institute, 1986). Data were analyzed first by site and LSD values calculated for mean separation. Data were then subjected to combined analysis

(McIntosh, 1983) across sites to evaluate main effects of site and associated interactions.

Use of SOYGRO and interpretation of output. Soil, site, and weather data for all trials were entered into SOYGRO V5.42 (Jones et al., 1989). No specific genetic coefficient file was available for the soybean variety Clark IV, therefore, the genetic coefficient file for general maturity group IV soybean was used. Simulations were run for each site under conditions of no water stress as trials were conducted under irrigated conditions. Predicted dates of emergence and first flower were compared with field data. Two adjustments, recommended by J.W. Jones and L.A. Hunt (personal communication), were made to the model input files to adjust simulated flowering dates to match those observed. Minimum temperature for optimum crop growth, TOPT1 in the CROPPARM.SBO file, was changed from 30 C to 25 C. Duration of the photoperiod sensitive phase during vegetative growth, VARTH[4] in the GENETICS.SB9 file, was reduced from 5.88 to 3.00 days. Simulations were then run for each site using the adjusted input files. SOYGRO output was compared with development and yield of plants in the fertilizer N treatment (Chapter 2). Nitrogen fertilized plants were chosen for the comparison as these were most representative of N‑sufficient plants. Phenophases of the model output did not correspond exactly to phenophases at each biomass harvest. Therefore, biomass predicted by the model on the date of each biomass harvest of N fertilized plants was used for comparison. Where simulations ran beyond observed crop duration, dates of maximum predicted biomass accumulation and harvest maturity (R8) were plotted.

Use of the BEANGRO model. Phenology and growth analysis data for bush bean were assembled into a database for comparison with output from the BEANGRO model (J.W. Jones and G. Hoogenboom, personal communication). No appropriate genetic coefficients were currently available for bush varieties of P. vulgaris (L.A. Hunt, personal communication), hence, obtaining reliable simulation output was not possible. Comparison between simulated and observed bush bean results awaits development of these genetic coefficients.

Results and Discussion

The effect of N source on phenology and yield of soybean and bush bean was evaluated in 4 different environments. Sites were planted at different times of the year and were located at different elevations. These provided differences in both photoperiod and temperature regimes (Table 5.1). While all growth, development, and yield variables differed significantly between sites (Appendix 7), the effect of changing N source on these variables, in N limited environments, was consistent across sites (Figures 5.1‑5.3 and Figures 5.6‑5.8).

Effect of N source on crop phenology. Both vegetative and reproductive development were affected by changing N source in the two crops. In general, vegetative growth was accelerated (Figure 5.1) and reproductive development delayed (Figure 5.2 and Appendix 7.1) by N sufficiency. These results agree with those of George, et al. (1990). In the trials reported here, delay in reproductive maturity resulted primarily from an increase in seed fill duration (Figure 5.3) as the time of flowering (R1/R2) was not affected (Figure 5.2). Observed differences in time of flowering between sites were temperature related as critical daylength for the soybean genotype Clark IV was met at all sites (Table 5.1). Similar to the report of George et al. (1990) a strong relationship was observed between time of flowering and growing degree days (GDD).

Differences in vegetative growth between N source treatments in both crops were apparent by full bloom (R2) when the rate of leaf appearance in N fertilized plants was as much as 29% greater than that in uninoculated plants (Figure 5.1). Differences in vegetative development between the N source treatments were not significant for either soybean or bush bean at site 5. Available soil N was shown to be sufficient to meet the N requirement of crops grown at this site (Chapter 2), hence, there was no N source treatment effect. Lack of differences between the treatments at this site, however, indicates that developmental differences observed at the other sites can be primarily attributed to plant N status. For soybean grown at these sites, N fertilized plants had 37‑60% greater leaf production by physiological maturity than uninoculated plants (Figure 5.1 A). Leaf production in symbiotic plants was 22‑40% greater than that in uninoculated plants, but, 514% lower than that of N fertilized plants. While symbiotic soybeans were more similar in developmental pattern to N fertilized plants, they were not equivalent.

Differences in rate of leaf appearance in bush bean could not be adequately assessed beyond the R4 phase in this study. This was because the vegetative stage of development descriptions used, that were originally developed for soybean (Fehr, et al., 1971), were inappropriate for describing the growth habit of bush bean.

While there were significant differences between sites in days to full‑bloom (R2), there was no effect of N source on flowering (R1 and R2) in either legume (Figure 5.2 and Appendix 7.1). Differences in reproductive phase duration due to N source in soybean were evident by R4 at sites 1, 3a, and 4. In general, the duration of each successive phase was slightly extended in N fertilized soybean compared to uninoculated plants. This resulted in significantly extended crop duration in the fertilizer N treatment at all sites. Although there was no effect of N source on vegetative growth at site 5, crop maturity was slightly delayed in N fertilized soybean plants. Delayed reproductive maturity of symbiotic plants was also observed at all sites except site 5, but, differences in phase duration between these and uninoculated plants did not occur until the later phases in reproductive development (generally between R6 and R7).

Differences in phase duration due to N source also occurred during the later reproductive phases in bush bean (Figure 5.2 B). With the exception of site 5, crop duration of N fertilized bush bean was significantly extended over that of both inoculated and uninoculated plants. No significant difference in crop duration between inoculated and uninoculated bush bean was observed at any site. There were, however, indigenous rhizobia capable of nodulating bush bean at all sites. And, at R2, nodule mass on uninoculated plants at sites 3a and 5 was not significantly different from that on inoculated plants (Figure 2.2). Nodule mass was significantly increased by inoculation at sites 1 and 4, but, increased nodulation did not significantly increase N accumulation (Appendix 2). Lack of

any difference in phase duration between these two treatments is, therefore, most likely due to lack of any significant difference in the N status of these plants.

Because the crop growth simulation model SOYGRO V5.42 assumes no N limitation to crop yield, output from SOYGRO simulations was compared with development and yield of plants in the fertilizer N treatment. Comparison between observed phonology of N fertilized soybean and that predicted by SOYGRO is presented in Figure 5.4. Results from the first simulation run (GRO 1) indicated that the model was unable to accurately predict phonology of the soybean genotype used in these experiments with the genetic coefficients developed for a generic maturity group IV soybean. Predicted crop duration was too long at sites 1, 3a, and 4 and too short at site 5 (Figure 5.4 A). Number of nodes on the main stem (V stage) was overpredicted at all sites (Figure 5.4 B). Adjusting model coefficients to achieve a match between observed and predicted flowering dates (GRO 2), resulted in improved prediction of crop duration at sites 1 and 3a, and a poorer fit to observed values at sites 4 and 5. Coefficient adjustment exacerbated the above described problem with V stage predictions.

In general, the SOYGRO model overestimated the rate and extent of leaf appearance in all environments, overestimated one or more of the durations of phases between R4 and R7 (seed filling period) at the warmer sites (sites 1 and 3a), and somewhat underestimated this period at the cooler sites (sites 4 and 5) (Figure 5.5 A). Prior to adjustment of model coefficients, days to flowering (R1) were also overestimated by as much as 12 days at some sites. Simulated values for time between physiological (R7) and harvest maturity (R8) were similar to those observed.

Accurate prediction of time to maturity and crop yield depends on correctly predicting both the rate and extent of leaf appearance and the time to critical developmental stages. If a model cannot predict crop phenology, it cannot produce an accurate estimate of yield. In line with the overestimation of node number and seed fill duration at sites 1 and 3a, the SOYGRO model seriously overpredicted seed yield at these sites (Figure 5.5 B). Although node number was also overestimated at sites 4 and 5, seed yield of N fertilized plants was underpredicted at both sites. These contrasting effects at the warmer (sites 1 and 3a) and cooler sites (sites 4 and 5) (Figure 5.5) may result from insufficient model definitions of the effects of temperature on phase duration and final yield in the soybean genotype Clark IV.

Effect of N source on crop growth and N assimilation rates, biomass accumulation and seed yield. Differences in crop growth rate during different stages of development were observed between N source treatments except at site 5 where rates were maintained at consistently low levels during all reproductive phases for all treatments in both soybean and bush bean (Table 5.2). Crop growth rate was lowest in uninoculated soybean, but, highest rates were observed in this treatment during flowering at all sites. Crop growth rate declined thereafter as growth became N limited. Growth rate of soybean was significantly increased by inoculation and N fertilization at sites 1 and 3a. Highest rates were observed for these treatments

during the early pod‑filling phase. Imsande (1989) and George et al. (1990) report similar results with symbiotic soybean. Bush bean crop growth rate was not enhanced by inoculation at these sites. Highest growth rates were observed in uninoculated and inoculated bush bean during flowering at site 1 and during early pod‑fill at site 3a. These results were exactly reversed for N fertilized bush bean. Patterns in N assimilation rates were similar to those observed for crop growth rate in both legumes (Table 5.3).

Significant differences between N source treatments in leaf area index (LAI) were observed by the first harvest (V4) (Appendix 7.6). Effects of N source on leaf weight ratio (LWR) were not observed until mid pod‑fill (R5/R6) (Appendix 7.5), when, LWR was increased by N sufficiency in soybean, but reduced in bush bean. Little to no effect of N source on specific leaf area (SLA) was observed. In general, N sufficiency resulted in greater leaf area in both legumes. Photosynthetic capacity was, therefore, enhanced in N sufficient plants. Improved C and N nutrition, increased rate of node production (Figure 5.1), and extended seed fill duration (Figure 5.3) resulted in significantly increased biomass and seed yield in N fertilized and symbiotic plants in N limited environments (Figures 5.65.8).

Leaf weight ratio, SLA, and LAI differed significantly between both sites and legume species at the first 3 harvests (Appendix 7.5 and 7.6). For soybean, SLA and LAI were lowest and LWR highest at the coolest site (site 5). These results indicate that soybean produced smaller, thicker leaves, relatively more leaves in relation to stem, but, less total leaf area in response to cooler

Table 5.2 Effect of N source on crop growth rate during vegetative and

Reproductive growth of soybean and bush bean at 3 field sites on Maui, HI.

Species Site N source

Vegetative

Flowering

Pod fill

Early Late

Crop

Duration

______________ kg biomass/ha/d ________________

G. max 1 Uninoculated

Inoculated

N Fertilized

149

180

-45

3a Uninoculated

Inoculated

N Fertilized

112

157

160

5 Uninoculated

Inoculated

N Fertilized

P. vulgaris 1 Uninoculated

Inoculated

N Fertilized

162

-20

-2

-62

3a Uninoculated

Inoculated

N Fertilized

104

171

132

137

136

5 Uninoculated

Inoculated

N Fertilized

Analysis of Variance

Source df

Pr > F

Site (ST) 2

< 0.001

0.008

< 0.001

Species (SP) 1

0.010

0.324

0.012

0.052

0.564

N source (N) 2

< 0.001

<0.001

ST * SP 2

0.003

0.018

0.001

< 0.001

0.002

ST * N 4

< 0.001

0.013

< 0.001

SP * N 2

< 0.001

0.049

< 0.001

ST * SP * N 4

0.989

0.673

< 0.001

0.002

< 0.001

Vegetative = period from sowing to V4, Flowering = period from V4 to R2, Early pod fill = period from R2 to R5/R6, Late pod fill = period from R5/R6 to R7, Crop duration = sowing to R7.

Table 5.3 Effect of N source on N assimilation rate during vegetative

And reproductive growth of soybean and bush bean at 3 field sites on Maui, HI.

Species Site N source

Vegetative

Flowering

Pod fill

Early Late

Crop

Duration

_________________ kg N/ha/d __________________

G. max 1 Uninoculated

Inoculated

N Fertilized

0.29

0.39

0.56

3.00

5.50

6.00

0.44

5.47

5.22

0.09

-0.98

-2.04

0.41

2.04

3a Uninoculated

Inoculated

N Fertilized

0.42

0.65

0.85

1.45

3.53

3.45

1.18

5.46

4.06

0.02

-0.05

3.54

0.64

2.63

2.80

5 Uninoculated

Inoculated

N Fertilized

0.37

0.32

2.63

2.34

2.98

1.14

2.30

2.05

1.60

2.31

2.66

1.10

1.54

1.69

P. vulgaris 1 Uninoculated

Inoculated

N Fertilized

0.25

0.49

1.75

2.48

4.10

0.13

0.15

3.77

-0.67

-5.23

0.24

0.30

1.82

3a Uninoculated

Inoculated

N Fertilized

0.59

0.56

1.08

2.50

2.25

6.00

2.03

2.19

1.05

5.00

1.61

1.07

1.81

1.44

1.83

5 Uninoculated

Inoculated

N Fertilized

0.63

0.50

0.75

2.96

2.40

1.96

0.85

1.24

1.96

2.01

2.06

2.52

1.37

1.38

1.75

Analysis of Variance

Source df

Pr > F

Site (ST) 2

< 0.001

0.048

< 0.001

Species (SP) 1

0.044

0.169

< 0.001

0.024

0.001

N source (N) 2

< 0.001

ST * SP 2

0.022

0.005

< 0.001

0.103

ST * N 4

< 0.001

SP * N 2

0.003

0.013

< 0.001

ST * SP * N 4

0.248

0.005

< 0.001

Vegetative = period from sowing to V4, Flowering = period from V4 to R2, Early pod fill = period from R2 to R5/R6, Late pod fill = period from R5/R6 to R7, Crop duration = sowing to R7.

temperature. Reduced photosynthetic capacity would result, which, could explain the significantly reduced yields observed in all treatments at this site (Figure 5.6 A). Response to lower temperature was quite different in bush bean. Leaf weight ratio, SLA, and LAI were significantly higher at site 5 by the third harvest (Appendix 7.5 and 7.6). This indicates that more leaves in relation to stem, with a greater leaf area were produced. This would enhance photosynthetic capacity, which, coupled with sufficient soil N, would account for the good yield obtained in all treatments at this site (Figure 5.6 B).

Increased biomass in response to N application and inoculation was evident by R2 in most cases and remained consistent throughout the crop cycle (Figures 5.7 and 5.8 and Appendix 7.2). Biomass and yield of symbiotic plants was most similar to that of N fertilized plants, but, symbiotic soybean yielded significantly less than N fertilized plants in the cooler environments (sites 4 and 5) (Figure 5.6 A). Biomass accumulation in N fertilized bush bean was significantly higher than either uninoculated or inoculated plants at the first 4 harvests at site 3a, but, increased biomass did not result in higher yield at this site (Figures 5.6 B and 5.8). This may reflect problems with partitioning of structural biomass to seed in this species. Treatment effects on nitrogen assimilation closely resembled those on biomass accumulation (Appendix 7.3).

There were highly significant differences between sites in all measured growth parameters by the first harvest that were maintained throughout crop growth (Appendix 7.2‑7.6). Yield potential of both crops was greatest at site 4 (Figure 5.6). Despite soil N

sufficiency, low temperature limited yield of soybean at site 5, whereas, bush bean yield was not strongly affected.

Biomass accumulation simulated by the SOYGRO model is compared with observed values in Figure 5.7. Model predictions of biomass accumulation at site 5 were remarkably accurate considering model problems in predicting phenology that were outlined above. Model difficulties with predicting duration of phases in the seed filling period at sites 1 and 3a can be seen clearly in Figure 5.7. The simulation ran well beyond observed crop duration and resulted in inflated yield predictions. Rate of biomass accumulation between the 2nd and 3rd harvest dates was underestimated by the model at both sites. However, had the date of physiological maturity (4th harvest) been accurately estimated, final simulated biomass and yield would not have significantly differed from that observed in the N fertilized plants at these sites.

Because of the essential role of N in most biological processes, effects of N deficiency on plant growth are profound. Available soil N was insufficient to achieve the maximum yield potential for soybean and bush bean at 3 of the 4 sites used in this study. At these sites, non symbiotic plants had significantly lower leaf area, decreased photosynthetic capacity, lower growth rate, and lower yield than either symbiotic or N fertilized plants. The period of most rapid growth in N fertilized and symbiotic soybean was during the early reproductive phase, whereas, growth rate was highest for uninoculated plants at flowering. Growth rate of N sufficient bush bean was also accelerated between flowering and mid pod‑fill. Source

and supply of N had a significant effect on crop phenology. Nitrogen sufficiency enhanced vegetative development while reproductive development was delayed. Increased crop duration observed in N sufficient plants was attributable to an increase in seed fill duration as time to flowering was not affected. Symbiotic plants were similar, but, in many cases, not equivalent to N fertilized plants in either development or yield. Extended phase duration was observed as early as R4 in N fertilized soybean. Whereas, differences in phase duration observed between inoculated and uninoculated plants did not occur until later reproductive stages (commonly between R5/R6 and R7). Extended growth phase duration in N fertilized bush bean was also not observed until the late reproductive phase. In general, N sufficient plants were larger and had a higher number of nodes on the main stem, and consequently, more leaves, pods, and seeds. This increased sink size extended the time required to remobilize structural biomass and N to seeds, hence, as much as a 10 day increase in seed fill duration was observed in these plants.

Accurate simulation of yield under varying N sources cannot be handled in the current version (V5.42) of the SOYGRO model. However, even for N sufficient plants, problems in simulating soybean development and yield were encountered. The model overpredicted rate and extent of leaf emergence and time of flowering in all environments and overpredicted seed fill duration in the warmer environments. Adjusting model coefficients to match observed and predicted flowering dates exacerbated the problems with leaf emergence and seed fill duration. Genetic coefficients and temperature response functions need adjustment if the SOYGRO model is to accurately simulate phenology and yield of the soybean cultivar Clark IV under non‑N‑limiting conditions. In N limited environments, the tremendous impact of N source on plant growth, development, and yield demonstrated in these trials indicates the need to address both source and supply of N in future versions of the SOYGRO model.

Complete soybean data sets from these trials and those conducted at sites 2 and 3 (Table 2.1, Chapter 2) have been provided to J.W. Jones of the University of Florida at Gainesville and G. Hoogenboom of the University of Georgia, Georgia Experiment Station, for their use in validation of a recently developed version of the SOYGRO model. The new version contains N subroutines that consider both soil N assimilation and symbiotic N2 fixation as sources of N for soybean crop growth.

SUMMARY AND CONCLUSIONS

Eight field inoculation trials were conducted at 5 well‑characterized sites in the MauiNet on the island of Maui, Hawaii. No less than 4 and as many as 7 legumes were planted at each site from among the following: soybean (Glycine max), lima bean (Phaseolus lunatus), cowpea (Vigna unguiculata), bush bean (Phaseolus vulgaris), peanut (Arachis hypogaea), Leucaena leucocephala, tinga pea (Lathyrus tingeatus), alfalfa (Medicago sativa), and clover (Trifolium repens). Crops were either: inoculated at high levels with an equal mixture of three effective strains of rhizobia: fertilized at high rates with urea: or left uninoculated with no N applied. Treatments measured legume inoculation response, crop yield potential, and influence of indigenous rhizobia, when present, respectively. Crops were otherwise grown under high management conditions. Size of indigenous homologous rhizobial populations and indices of soil N availability were measured at each site. Climatic details were recorded for all sites during crop growth.

Major objectives of this study were to identify and quantify the primary environmental factors that determine and can be used to predict the symbiotic success of inoculant rhizobia introduced into tropical soils. Symbiotic success was defined in several ways: (i) ability of inoculation to significantly increase yield over uninoculated crops (inoculation response): (ii) ability of inoculant strains to compete with indigenous rhizobia for nodule occupancy (competitive competence); (iii) ability of inoculant rhizobia to compete among themselves for nodule occupancy in different environments; and (iv) ability of the symbiosis to supply the host with fixed N for maximum yield.

Numbers of indigenous rhizobia and soil N availability in relation to crop N requirement were found to be the primary determinants of inoculation response as long as there were no other serious environmental limitations to yield. Response to inoculation was inversely related to numbers of indigenous rhizobia. As few as 54 rhizobia g^‑1 soil eliminated inoculation response. When fewer than 10 indigenous rhizobia g^‑1 soil were present, inoculation significantly increased economic yield 85% of the time. A significant yield increase due to inoculation was obtained in only 6% of the observations where numbers of indigenous rhizobia were greater than 10 cells g^‑1 soil.

A significant response to N application, indicating an N limitation to maximum yield, did not guarantee a significant inoculation response. Neither did significant increases in nodule parameters. While inoculant strains were very successful in competing with indigenous rhizobia for nodule occupancy, no less than a doubling of nodule mass, and 66% nodule occupancy by inoculant rhizobia were required to significantly increase yield of inoculated over that of uninoculated crops. Lack of an inoculation response was common, however, even when inoculum strains occupied the majority of nodules formed.

The relationship between numbers of indigenous rhizobia and legume inoculation response was best described using a hyperbolic equation. Slope coefficients generated from hyperbolic regressions performed on a site basis were significantly related to indices of soil N availability. Replacing the slope: coefficient in the hyperbolic response regression with equations incorporating indices of soil N availability yielded useful models for describing, quantifying, and predicting legume inoculation response. Nitrogen derived from N₂ fixation in soybean proved to be the best indicator of crop N demand in these trials as it directly measured the crop symbiotic N requirement. The best fit between observed and predicted values was obtained from the equation that contained this N variable. A significant fit of observed to predicted values was also obtained using soil N mineralization values from the different sites to express soil N supply. Using this equation, predictions regarding inoculation response could be made directly from results of soil analyses.

Nodule occupancy by inoculant rhizobia was significantly correlated to the same environmental variables as numbers of indigenous, homologous rhizobia. Correlation coefficients for these two dependent variables were similar in magnitude, but, opposite in sign. This result suggests that environmental factors exert their influence on nodule occupancy by inoculant strains indirectly through their impact on abundance of indigenous rhizobia. And, that number of indigenous rhizobia present at a site is the primary environmental factor controlling nodule occupancy by inoculant strains. Competitive success of inoculant rhizobia was inversely related to numbers of indigenous rhizobia. Models to predict the outcome of competition for nodule occupancy between inoculant and indigenous rhizobia obtained from the literature were evaluated and compared with the best mathematical relationship obtained for observed values. Two equations from the literature were able to provide a significant fit to observed values. At the consistently high inoculant application rates used in these trials, a simplification of the equation proposed by Weaver and Frederick: (1974a) (log‑linear) provided the best fit to nodule occupancy by inoculant strains observed in this study.

Strength of the competition barrier presented by indigenous populations of rhizobia was expressed as the percent nodule occupancy by indigenous rhizobia divided by their population size (log₁₀). This index was useful for comparing the relative competitiveness of indigenous rhizobial populations across sites and indicated that the more competitive indigenous populations were observed in the harsher environments.

Effectiveness of indigenous rhizobia belonging to the cowpea miscellany, Bradyrhizobium sp., was determined on 4 legumes belonging to the cowpea 'cross‑inoculation' group. Crushates of nodules formed on cowpea following inoculation with soil from 3 field sites were tested for their effectiveness on cowpea, lima bean, peanut, and siratro. Effectiveness of nodule crushates applied to cowpea roots was approximately normally distributed. Presence of rhizobia significantly more effective than inoculant strains was found in each of the 3 soils. Effectiveness and invasiveness of the nodule crushates on siratro was similar to their effectiveness on cowpea, both legumes being very promiscuous. Effectiveness and invasiveness of the crushates varied considerably from that observed on cowpea when applied to lima, bean and peanut. Peanut was more specific for nodulation than any of the other legumes and was more specific for effectiveness than either cowpea or siratro. Whereas, lima bean was more specific for effectiveness, but, showed greater specificity for infection than either cowpea or siratro. The greatest disparity was observed for both infectiveness and effectiveness of nodule crushates between lima bean and peanut. Both legumes shared a larger proportion of crushates in common with cowpea and siratro than with each other. The presence of infective, effective rhizobia capable of nodulating each legume was demonstrated for these sites. Differences in observed infectiveness profiles helped to explain vast differences in the most probable number of indigenous rhizobia counted on these legumes at these sites.

An equal mixture of 3 serologically distinct strains of rhizobia, differing for each of the 8 legumes used in these trials, comprised the inoculant. In competition for nodule occupancy between the 3 inoculant strains of rhizobia, one of strains for each legume species (except clover) was identified as a poor competitor across environments. Competition between the other 2 inoculant strains for each legume species was correlated with environmental factors for some strain/legume combinations, but not for others. Soil minimum temperature and clay content were the 2 environmental variables most frequently correlated with competitive success of one inoculant strain over another. Nodule occupancy by TAL 1383 on bush bean was significantly correlated with soil sodium content. None of the other environmental variables examined were significantly correlated to the competitive success of inoculant strains. In these trials, fields were limed where required and irrigated at all sites. However, extent of differences between environments were still considerable. In this light, it was remarkable that so few variables were found to significantly influence the outcome of competition for nodule occupancy between inoculant strains used. This result suggested that highly competitive inoculant strains can be identified that will perform well across a range of environments. However, failure of at least one of the 3 inoculant strains for each legume species to compete well for nodule occupancy in these trials argues against the use of single strain inoculant, particularly in more stressful environments.

An in‑depth analysis of the impact of varying N source on the growth and phenology of soybean and bush bean was conducted at four sites. This portion of the study focused on 3 questions: (i) the ability of inoculation to supply the host with fixed N for maximum yield: (ii) ability of a current process‑oriented crop growth simulation model to accurately estimate soybean development and yield in Hawaii under non‑N limiting conditions; and (iii) whether N source influenced growth and developmental aspects of leguminous crops sufficiently to warrant the attention of crop modelers. Phenology, rate and extent of leaf emergence (node production), rate of biomass and N accumulation, and yield of non‑symbiotic, symbiotic, and N fertilized plants were compared.

Increases in all growth parameters were observed in symbiotic and N fertilized plants in N limited environments. However, N sufficiency delayed time to critical reproductive stages starting as early as beginning pod‑fill. Symbiotic plants were found to be similar in phenology and yield to those receiving high rates of N application, but, were not identical. In general, symbiotic plants accumulated less biomass across the crop cycle and yielded less than N fertilized plants, particularly in the cooler environments. Effects of N sufficiency on phenology were also not as pronounced in symbiotic plants.

There was significant disparity between observed phenology and yield and that simulated by the SOYGRO model for N sufficient soybean grown at these sites. Model simulations overestimated rate and extent of leaf emergence at all sites; overestimated crop duration, rate of biomass accumulation, and yield of plants grown in warmer environments; and somewhat underestimated these in the cooler environments. Time of flowering at all sites was also significantly overestimated using the current version (5.42) of the model. Crop temperature response functions and genetic coefficients require adjustment for accurate simulation of the growth and phenology of the soybean genotype Clark IV by the SOYGRO model.

Irrespective of difficulties in simulating the best case scenario in soybean, significant differences in phenology and growth observed between N source treatments in this study indicate that future versions of this soybean crop growth simulation model need to include subroutines that can integrate the effects of both source and supply of N on soybean development and yield.

The primary ecological determinants of the performance of introduced rhizobia in tropical soils were found to be number and competitiveness of indigenous rhizobia and soil N availability in relation to crop N demand. These variables can be incorporated into mathematical models and used to predict inoculation response of legumes and nodule occupancy by inoculant rhizobia. These models should reduce the need to conduct multiple inoculation trials in order to determine the inoculation requirements of legumes grown in diverse environments.

APPENDIX 2.1	Site 1 (Hashimoto Farm) field harvest data summary - PD 3/24/87 (Gm 4/8/87)
			Early Harvest - 35-40 DAP				Late Harvest at Harvest Maturity (R8)
Legume Species	N Source Trmt	Biomass (kg/ha)	N Uptake (kg/ha)	Nodule Number (/ha)	Nodule Mass (kg/ha)	Total Biomass (kg/ha)	Seed Yield (kg/ha)	Stover (kg/ha)	Harvest Index	Total N Uptake (kg/ha)	Seed N (kg/ha)
G. max	Uninoc inoc + N	494 714 808	15 29 38	- a 15834686 -	- a 42.4 -	1624 5078 5571	627 3025 3024	996 2053 2547	0.35 0.59 0.54	32 191 180	27 177 163
P. lunatus	Uninoc Inoc + N	1067 1339 2727	30 36 113	39223 5724750 50497	0.6 49.3 0.01	4111 6967 10457	1379 2531 3970	2732 4436 6487	0.33 0.36 0.38	91 135 297	40 77 186
V. unguiculata	Uninoc Inoc + N	2803 2691 4487	74 87 171	2609714 8201475 2126937	33.3 79.9 6.1	5527 5411 7764	2179 2113 2839	3348 3298 4925	0.40 0.40 0.37	153 136 197	85 100 115
LSD (0.05)(18)		416	14	b	c	1517	705	921	0.05	45	30
CV (%)		14.7	14.1			17.5	19.7	18.1	8.6	19.2	18.7
Spp Effect		***	***			**	ns	***	***	ns	*
Trmt Effect		***	***			***	***	***	***	***	***
Trmt*Spp Inter	action	***	***			**	***	**	***	***	***
P. vulgaris	Uninoc Inoc + N	576 599 1414	20 23 59	3774722 37871455 175427	1.9 15.9 0.1	888 1403 4981	400 731 2891	489 672 2091	0.45 0.52 0.58	17 27 128	11 20 100
LSD (0.05)(6)		229	10	b	c	441	257	201	0.03	7	8
CV (%)		15.1	17.2			10.5	11.1	10.7	2.8	6.1	8.7
Trmt Effect		***	***			***	***	***	***	***	***
a Confidence interval does not include zero. b LSD for P. vulgaris and P. lunatus = 1600000; LSD for V. unguiculata = 5500000. c LSD for P. vulgaris and P. lunatus = 8.3; LSD for V. unguiculata = 58.1.

APPENDIX 2.2 Site 2 (Kuiaha) field harvest data summary - PD 8/15/86
			Early Harvest - 35-40 DAP				Late Harvest at Harvest Maturity (R8)
Legume Species	N Source Trmt	Biomass (kg/ha)	N Uptake (kg/ha)	Nodule Number (/ha)	Nodule Mass (kg/ha)	Total Biomass (kg/ha)	Seed Yield (kg/ha)	Stover (kg/ha)	Harvest Index	Total N Uptake (kg/ha)	Seed N (kg/ha)
G. max	Uninoc Inoc + N	805 1387 1104	16 53 35	- a 16066167 887167	- a 76.5 4.7	1669 5073 4692	840 3120 2962	829 1953 1731	0.50 0.61 0.63	47 206 166	42 183 152
P. lunatus	Uninoc Inoc + N	1095 1408 1381	37 46 45	5285470 10998667 1971083	22.7 51.8 10.5	6101 6644 6579	2819 2698 2743	3282 3946 3836	0.46 0.41 0.42	112 124 135	69 57 58
P. vulgaris	Uninoc Inoc + N	986 1085 1663	20 29 52	2399223 30338667 4635083	2.7 36.5 3.3	2931 3142 5023	1649 1781 2858	1282 1361 2264	0.57 0.57 0.55	69 75 139	56 60 109
V. unguiculata	Uninoc Inoc + N	1578 1628 1797	54 45 67	25788322 27662167 21067494	100.6 105.2 53.4	4247 5175 5820	1443 1738 2082	2805 3436 3738	0.34 0.34 0.36	115 151 159	61 77 76
LSD (0.05)(24)^b		208	7.5	6691182	13.3	790	399	538	0.05	26	18
CV (%)		10.7	12.4	31.1	20.8	11.4	12.3	14.5	6.6	14.3	14.7
Spp Effect		**	***	***	***	***	***	***	***	**	***
Trmt Effect		***	***	***	***	***	***	***	ns	***	***
Trmt*Spp Interaction		***	***	***	***	***	"**	ns	***	***	***
a Confidence interval does not include zero. b df for nodulation data = 18

APPENDIX 2.3 Site 3 (Kula Agricultural Park) field harvest data summary - PD 9/12/86
			Early Harvest - 35-40 DAP				Late Harvest at Harvest Maturity (R8)
Legume Species	N Source Trmt	Biomass (kg/ha)	N Uptake (kg/ha)	Nodule Number (/ha)	Nodule Mass (kg/ha)	Total Biomass (kg/ha)	Seed Yield (kg/ha)	Stover (kg/ha)	Harvest Index	Total N Uptake (kg/ha)	Seed N (kg/ha)
G. max	Uninoc Inoc + N	568 1036 1122	13 34 27	- a 16536583 -	- a 87.3 -	1002 3318 4413	485 2026 2733	517 1292 1680	0.48 0.61 0.62	31 154 199	27 141 187
P. lunatus	Uninoc Inoc + N	652 928 1502	14 23 53	192409 15551424 92389	3.1 68.0 2.3	3644 6680 7424	1520 3012 3179	2123 3668 4245	0.42 0.45 0.43	82 155 199	49 103 131
P. vulgaris	Uninoc Inoc + N	697 1147 1998	14 31 67	2844249 34198583 328603	7.0 66.1 0.3	1242 2024 3275	669 1228 1853	573 797 1422	0.54 0.61 0.57	23 41 77	18 36 61
V. unguiculata	Uninoc Inoc + N	1062 1243 1714	32 33 61	4244917 23375322 2143882	26.7 90.3 4.0	3749 3886 4898	1482 1667 1926	2268 2219 2972	0.40 0.43 0.40	89 93 140	53 64 81
LSD b (0.05)(24)		243	15	2831112	12.1	736	390	408	0.04	32	27.2
CV (%)		14.6	30.7	20.6	27.3	13.3	14.7	14.1	5.3	20.3	23.6
Spp Effect		*	*	***	***	***	***	***	***	***	***
Trmt Effect		***	***	***	***	***	***	***	***	***	***
Trmt*Spp Interaction		**	**	***	ns	***	***	***	***	***	***
a Confidence interval does not include zero. b df for nodulation data = 18

APPENDIX 2.4 Site 4 (Haleakala Station) field harvest data summary - PD 6/08187
			Early Harvest - 35-40 DAP				Late Harvest at Harvest Maturity (R8)
Legume Species	N Source Trmt	Biomass (kg/ha)	N Uptake (kg/ha)	Nodule Number (/ha)	Nodule Mass (kg/ha)	Total Biomass (kg/ha)	Seed Yield (kg/ha)	Stover (kg/ha)	Harvest Index	Total N Uptake (kg/ha)	Seed N (kg/ha)
G. max	Uninoc Inoc + N	1132 969 1063	35 32 48	- a 19154223 -	- a 47.0 -	4111 6601 9042	1711 3686 4596	2400 2915 4446	0.42 0.56 0.51	109 263 354	83 221 279
P. lunatus	Uninoc Inoc + N	2403 2502 2539	74 82 98	8479012 14383124 3697178	39.8 42.2 3.3	10607 10296 11073	4117 3838 4165	6491 6458 6908	0.39 0.37 0.38	250 246 293	145 138 158
P. vulgaris	Uninoc Inoc + N	1593 1572 1867	47 43 68	4625396 41135242 871719	5.5 27.7 0.3	5423 6621 7342	2622 3489 3868	2801 3132 3473	0.48 0.53 0.53	91 133 161	70 101 128
V. unguiculata	Uninoc Inoc + N	3407 2975 3904	116 105 145	16576254 16358697 10156952	51.2 68.6 11.5	8146 7991 9052	2884 2811 2923	5262 5180 6128	0.35 0.35 0.32	195 193 236	107 102 110
b LSD (0.05)(24)		472	18	9754253	16.8	1594	793	1112	0.06	50	33
CV (%)		15.0	16.0	50.8	40.6	13.6	16.0	16.4	8.9	16.4	16.5
Spp Effect		***	***	***	***	***	*	***	***	***	***
Trmt Effect		*	***	***	***	***	***	**	*	***	***
Trmt* Spp Interaction		ns	ns	***	*	**	***	ns	***	***	***
a Confidence interval does not include zero. b df for nodulation data = 18

APPENDIX 2.5 Site 5 (Tengan Farm) field harvest data summary- PD 10/20/87 (Gm); 10/28/87 (PI and Pv); 11/18187 (Vu)
			Early Harvest - 35-40 DAP				Late Harvest at Harvest Maturity (R8)
Legume Species	N Source Trmt	Biomass (kg/ha)	N Uptake (kg/ha)	Nodule Number (/ha)	Nodule Mass (kg/ha)	Total Biomass (kg/ha)	Seed Yield (kg/ha)	Stover (kg/ha)	Harvest Index	Total N Uptake (kg/ha)	Seed N (kg/ha)
G. max	Uninoc Inoc + N	871 823 877	28 23 32	- a 8395689 -	- b 24.3 -	3082 3314 4488	1356 1233 1983	1726 2082 2505	0.44 0.37 0.44	97 116 158	73 74 109
P. lunatus	Uninoc Inoc + N	1241 1249 1472	32 31 40	4728018 9516899 2181036	6.1 13.0 0.8	9443 10727 11377	3793 4135 4627	5651 6592 6750	0.40 0.39 0.41	205 261 319	117 135 186
P. vulgaris	Uninoc Inoc + N	1121 939 1204	40 32 35	19354439 25552555 8165662	6.5 8.2 1.4	5123 6375 5644	2625 3035 2694	2498 3341 2950	0.50 0.47 0.48	106 132 136	71 82 81
V. unguiculata	Uninoc Inoc + N	3300 4855 2991	97 142 99	8237207 8718849 5167896	39.7 36.5 1.9	7281 6266 6565	1910 1801 1746	5371 4464 4819	0.27 0.29 0.26	170 153 178	79 76 74
c LSD (0.05)(24)		807	25	6435150	12.1	1726	848	991	0.04	54	34
CV (%)		32.8	33.6	42.5	64.1	17.8	22.5	16.7	7.8	21.8	24.0
Spp Effect		***	***	***	***	***	***	***	***	***	***
Trmt Effect		ns	ns	***	***	ns	ns	ns	ns	**	**
Trmt*Spp Interaction		*	ns	*	**	ns	ns	ns	ns	ns	ns
a Not significantly different from zero. b Confidence interval does not include c df for nodulation data = 18

APPENDIX 2.6 Site 1a (Hashimoto Farm - 2nd planting) field harvest data summary - PD 3/10/88
			Early Harvest - 35-40 DAP			Late Harvest at Harvest Maturity
Legume Species	N Source Trmt	Biomass (kg/ha)	N Uptake (kg/ha)	Nodule Number (/ha)	Nodule Mass (kg/ha)	Total Biomass (kg/ha)	Seed Yield (kg/ha)	Total N Uptake (kg/ha)	Seed N (kg/ha)
A. hypogaea	Uninoc Inoc + N	776 808 892	14 16 19	23215101 20176153 14163505	20.2 22.9 10.3	12309 14648 15144	5486 6579 6082	289 361 367	246 312 291
LSD (0.05)(6) CV (%) Trmt Effect		123 8.6 ns	3 11.6 *	7696136 23.2 ns	6.8 22.1 **	2020 8.3 *	1206 11.5 ns	74 12.5 ns	77 15.8 ns
L. leucocephala	Uninoc Inoc + N	1121 1236 1812	30 27 43	nd nd nd	nd nd nd	24665 21225 26772		456 339 487
LSD (0.05)(6)		404	13	-	-	5603		146
CV (%)		16.8	22.7	-	-	13.4		19.8
Trmt Effect		*	ns	-	-	ns		ns

APPENDIX 2.7 Site 3a (Kula Agricultural Park - 2nd planting) field harvest data summary - PD 5/14187
				Early Harvest - 35-40 DAP				Late Harvest at Harvest Maturity (R8)
Legume Species		N Source Trmt	Biomass (kg/ha)	N Uptake (kg/ha)	Nodule Number (/ha)	Nodule Mass (kg/ha)	Total Biomass (kg/ha)	Seed Yield (kg/ha)	Stover (kg/ha)	Harvest Index	Total N Uptake (kg/ha)	Seed N (kg/ha)
G. max		Uninoc Inoc + N	709 887 1230	18 34 44	- a 18206824 -	- a 65.8 -	2150 5629 7114	935 2782 3125	1215 2848 3989	0.42 0.50 0.44	50 203 226	42 166 179
P. vulgaris		Uninoc Inoc + N	1323 1266 2353	35 33 84	27285774 66307223 19448056	39.5 56.4 6.6	4120 4255 4402	2198 2316 2130	1638 1939 2272	0.55 0.54 0.48	75 77 101	57 60 72
A. hypogaea		Uninoc Inoc + N	536 697 978	12 19 26	16340950 14224225 18420100	11.6 22.7 13.9	14347 20487 16859	4926 5921 5679		0.48 0.41 0.51	256 332 299	217 276 247
b LSD (0.05)(17)			248	11	9869377	11.9	1191	949	1202	0.07	41	41
CV (%)			15.2	22.4	33.3	39.0	8.8	18.5	21.1	10.6	14.8	18.3
Spp Effect			***	***	***	***	***	***	***	ns	**	**
Trmt Effect			***	***	***	***	***	***	***	ns	***	***
Trmt*Spp Interaction			***	***	***	***	***	ns	***	ns	***	**
L. leucocephala	Uninoc Inoc + N		551 577 770	18 19 30	5227259 5234667 2380000	13.0 16.1 2.9	14682 17502 21998				317 323 459
b LSD (0.05)(5)			562	24			4470				101
CV (%)			48.3	58.6			13.4				14.2
Trmt Effect			ns	ns			*				*
a Confidence interval does not include zero. b P. vulgaris and L. leucocephala each had one missing replication; df for nodulation data = 22(includes Leucaena).

APPENDIX 2.8 Site 5a (Tengan Farm - 2nd planting) forage legumes field harvest data summary - PD 1/7/88
		Early Harvest - 71-74 DAP				Late Harvest - 112-117 DAP
Legume Species	N Source Trmt	Biomass (kg/ha)	N Uptake (kg/ha)	Nodule Number (/ha)	Nodule Mass (kg/ha)	Total Biomass (kg/ha)	Total N Uptake (kg/ha)
M. sativa	Uninoc Inoc + N	4130 3683 4774	156 112 196	2.7E+08 1.9E+08 81651639	93.8 65.0 23.6	4640 5129 5320	107 119 131
T. repens	Uninoc Inoc + N	3596 4847 4202	125 186 188	16083332 48333332 5300000	14.3 12.1 1.1	3562 3778 4078	97 104 133
L. tingeatus	Uninoc Inoc + N	4776 3678 4806	201 156 227	13582575 25805700 12640500	75.1 98.7 29.5	2946 3598 3050	91 111 108
LSD (0.05)(18)		665	32	68613494	17.0	906	23
CV (%)		10.5	12.1	62.9	24.9	15.2	13.8
Spp Effect		ns	**	***	***	**	ns
Trmt Effect		*	***	**	***	ns	**
Trmt*Spp Interaction		***	*	**	**	ns	ns

APPENDIX 6 Rate of N accumulation in legumes grown in 8 inoculation trials conducted

at 5 sites on Maui, HI.

Legume

Species

N Source

Trmt

Site number

1 2 3 4 5 1a 3a 5a

____________________ kg N/ha/d_____________________

G. max

Uninoc

Inoc

0.44

2.59

2.41

0.58

2.45

1.83

0.38

1.90

2.19

1.10

2.63

3.36

0.95

1.12

1.50

- 0.52

- 2.12

- 2.30

P.lunatus

Uninoc

Inoc

0.92

1.37

3.00

1.32

1.46

1.59

0.90

1.53

1.99

2.05

2.02

2.40

1.32

1.69

2.06

- -

P. vulgaris

Uninoc

Inoc

0.25

0.48

1.96

1.10

1.18

2.21

0.33

0.60

1.11

1.21

1.65

3.00

1.12

1.36

1.43

- 1.01

- 1.03

- 1.70

unguiculata

Uninoc

Inoc

1.54

1.37

1.99

1.26

1.66

1.75

0.94

0.98

1.47

1.70

1.68

2.05

1.14

1.03

1.20

- -

A. hypogaea

Uninoc

Inoc

2.09 1.92

2.64 2.49

2.68 2.25

leucocephala

Uninoc

Inoc

2.75 2.18

2.04 2.08

2.94 2.88

M. sativa

Uninoc

Inoc

- -

0.93

1.04

1.15

T. repens

Uninoc

Inoc

- -

0.85

0.91

1.17

L, tingeatus

Uninoc

Inoc

- -

0.80

0.97

0.95

LSD (0.05)

0.40

0.30

0.35

0.43

0.42

0.66 0.40

0.20

CV (%)

18.8

13.3

20.4

14.3

21.4

16.9 14.8

13.8

Spp Effect

***

ns **

Trmt Effect

***

ns ***

Spp * Trmt Interaction

***

* ***

a Calculated by dividing N accumulation (kg/ha) at harvest maturity (R8) by total

crop duration in days.

APPENDIX 7.1 Average days to critical phenological stages in soybean and bush bean grown at 4 sites on Maui, HI.

No.

Site

Name

Legume

Species

Source

Trmt

Phenological Stage

V4 R2 R6 R7 R8

Seed fill

Duration

_________________days _____________

Hashimoto

Farm

G. max

P. vulgaris

Uninoc

Inoc

FertN

Uninoc

Inoc

FertN

Kula

Agric.

Park

G. max

P, vulgaris

Uninoc

Inoc

FertN

Uninoc

Inoc

FertN

Haleakala

Station

G. max

P. vulgaris

Uninoc

Inoc

FertN

Uninoc

Inoc

FertN

106

108

113

Tengan

Farm

G, max

P. vulgaris

Uninoc

Inoc

FertN

Uninoc

Inoc

FertN

109

113

Period from R4 to R7.

LITERATURE CITED

Abaidoo, R.C., T. George, B.B. Bohlool, and P.W. Singleton. 1990. Influence of elevation and applied nitrogen on rhizosphere colonization and competition for nodule occupancy by different rhizobial strains on field‑grown soybean and common bean. Appl. Environ. Microbiol. In press.

Amarger, N. 1984. Evaluation of competition in Rhizobium spp. p. 300‑305. In: M.J. Klug and C.A. Reddy (ed.) Current Perspectives in Microbial Ecology. Amer. Soc. Microbiol., Washington, D.C.

Amarger, N., and J.P. Lobreau. 1982. Quantitative study of nodulation competitiveness in Rhizobium strains. Appl. Environ. Microbiol. 44:583‑588.

Atkins, C.A. 1986. The legume/Rhizobium symbiosis: Limitations to maximizing nitrogen fixation. Outlook on Agric. 15:128‑134.

Bauer, W.D. 1981. Infection of legumes by rhizobia. Ann. Rev. Plant Physiol. 32:407‑449.

Beattie, G.A., M.K. Clayton, and J. Handelsman. 1989. Quantitiative comparison of the laboratory and field competitiveness of Rhizobium leguminosarum biovar phaseoli. Appl. Environ. Microbiol. 55:2755‑2761

Berg, R.K. Jr., T.E. Loynachan, R.M. Zablotowicz, and M.T. Lieberman. 1988. Nodule occupancy by introduced Bradyrhizobium japonicum in Iowa soils. Agron. J. 80:876‑881.

Bergersen, F.J. 1970. Some Australian studies relating to long term effects of the inoculation of legume seed. Plant Soil. 32:727‑736.

Bohlool, B.B., and E.L. Schmidt. 1973. Persistence and competition aspects of Rhizobium japonicum observed in soil by immunofluorescence microscopy. Soil Sci. Soc. Amer. Proc. 37:561‑564.

Boonkerd, N., D.F. Weber, and D.F. Bezdicek. 1978. Influence of Rhizobium japonicum strains and inoculation methods on soybean grown in rhizobia‑populated soils. Agron. J. 70:547‑549.

Boonkerd, N., and R.W. Weaver. 1982. Survival of cowpea rhizobia in soil as affected by soil temperature and moisture. Appl. Environ. Microbiol. 43:585‑589.

Bremner, J.M., and C.S. Mulvaney. 1982. Nitrogen‑total. In A.L. Page, R.H. Miller, and D.R. Keeney (ed.) Methods of Soil Analysis. Part 2. Agronomy 9:595‑624.

Brockwell, J., R.R. Gault, M. Zorin, and M.J. Roberts. 1982. Effects of environmental variables on the competition between inoculum strains and naturalized populations of Rhizobium trifolii for nodulation of Trifolium subterraneum L. and on rhizobia persistence in the soil. Aust. J. Agric. Res. 33:803‑815.

Brockwell, J., R.J. Roughly, and D.F. Herridge. 1987. Population dynamics of Rhizobium japonicum strains used to inoculate three successive crops of soybean. Aust. J. Agric. Res. 38:61‑74.

Caldwell, B.E. 1969. Initial competition of root‑nodule bacteria on soybeans in a field environment. Agron J. 61:813‑815.

Caldwell, B.E., and G. Vest. 1968. Nodulation interaction between soybean genotypes and serogroups of Rhizobium japonicum. Crop Sci. 8:680‑682.

Caldwell, B.E., and D.F. Weber. 1970. Distribution of Rhizobium japonicum serogroups in soybean nodules as affected by planting dates. Agron. J. 62:12‑14.

Damirgi, S.M., L.R. Frederick, and I.C. Anderson. 1967. Serogroups of Rhizobium japonicum in soybean nodules as affected by soil types. Agron. J. 59:10‑12.

Dart, P. 1977. Infection and development of leguminous nodules. p. 367‑472. In R.W.F. Hardy and W.S. Silver (ed.) A Treatise on Dinitrogen Fixation, Section III. Biology. Wiley, New York.

Department of Land and Natural Resources. 1982. Median rainfall. Circular C88, Department of Natural Resources, Division of Water and Land Development, Honolulu, HI.

Diatloff, A., and S. Langford. 1975. Effective natural nodulation of peanuts in Queensland. Queensl. J. Agric. Anim. Sci. 32:95‑100.

Diatloff, A., and J. Brockwell. 1976. Ecological studies of root‑nodule bacteria introduced into field environments: 4. Symbiotic properties of Rhizobium japonicum and competitive success in nodulation of two Glycine max cultivars by effective and ineffective strains. Aust. J. Exp. Agric. Anim. Husb. 16:514‑521.

Doku, E.V. 1969. Host specificity among five species in the cowpea cross‑inoculation group. Plant Soil 30:126‑128.

Dowling, D.N., and W.J. Broughton. 1986. Competition for nodulation in legumes. Ann. Rev. Microbiol. 40:131‑157.

Dughri, M.H., and P.J. Bottomley. 1983. Effect of acidity on the composition of an indigenous soil population of Rhizobium trifolii found in nodules of Trifolium subterraneum L. Appl. Environ. Microbiol. 46:1207‑1213.

Dughri, M.H., and P.J. Bottomley. 1984. Soil acidity and the composition of an indigenous population of Rhizobium trifolii in nodules of different cultivars of Trifolium subterraneum L. Soil Biol. Biochem. 16:405‑411.

Elkins, D.M., G. Hamilton, C.K.Y. Chan, M.A. Briskovich, and J.W. Vandeventer. 1976. Effect of cropping history on soybean growth and nodulation and soil rhizobia. Agron. J. 68:513‑517.

FAO. 1984. Legume Inoculants and Their Use. Food and Agriculture Organization of the United Nations, Rome, Italy.

Fehr, W.R., C.E. Caviness, D.T. Burmood, and J.S. Pennington. 1971. Stage of development descriptions for soybeans, Glycine max (L.) Merrill. Crop Sci. 11:929‑931.

Fred, E.B., 1.L. Baldwin, and E. McCoy. 1932. Root Nodule Bacteria and Leguminous Plants. University of Wisconsin Studies in Science, No. 5, Madison, WI

George, T., B.B. Bohlool, and P.W. Singleton. 1987. Bradyrhizobium j aponicum‑environment interactions: Nodulation and interstrain competition in soils along an elevational transect. Appl. Environ. Microbiol. 53:11131117.

George, T., D.P. Bartholomew, and P.W. Singleton. 1990. Effect of temperature and maturity group on phenology of field grown nodulating and nonnodulating soybean isolines. In press.

Gibson, A.H., and J.E. Harper. 1985. Nitrate effect on nodulation of soybean by Bradyrhizobium japonicum. Crop Sci. 25:497‑501.

Graham, P.H. 1981. Some problems of nodulation and symbiotic nitrogen fixation in Phaseolus vulgaris L.: A review. Field Crops Res. 4:93‑112.

Hadley, P., E.H. Roberts, R.J. Summerfield, and F.R. Minchin. 1984. Effects of temperature and photoperiod on flowering in soya bean [Glycine max (L.) Merrill]: A quantitative model. Ann. Bot. 53:669‑681.

Ham, G.E. 1980. Inoculation of legumes with Rhizobium in competition with naturalized strains, p. 131‑138. In W.E. Newton and W.H. Orme‑Johnson (ed.) Nitrogen Fixation, Vol. II, University Park Press, Baltimore.

Ham,G.E., and V.B. Cardwell, and H.W. Johnson. 1971. Evaluation of Rhizobium japonicum inoculants in soils containing naturalized populations of rhizobia. Agron. J. 63:301‑303.

Ham, G.E., L.R. Frederick, and I.C. Anderson. 1971. Serogroups of Rhizobium japonicum in soybean nodules sampled in Iowa. Agron. J. 63:69‑72.

Harris, S.C. 1979. Planning an international network of legume inoculation trials. NifTAL Project and U.S. Agency for International Development, Paia, Hawaii.

Hodges, T., and V. French. 1985. Soyphen: Soybean growth stages modeled from temperature, daylength, and water availability. Agron J. 77:500‑505.

Holding, A.J., and J.F. Lowe. 1971. Some effects of acidity and heavy metals on the Rhizobium‑leguminous plant association. Plant Soil, Spec. Vol. 153‑166.

Hunter, W.J., and C.J. Fahring. 1980. Movement by Rhizobium and nodulation of legumes. Soil Biol. Biochem. 12:537‑542.

Imsande, J. 1988. Interrelationship between plant developmental stage, plant growth rate, nitrate utilization and nitrogen fixation in hydroponically grown soybean. J. Exp. Bot. 39:775‑785.

Imsande, J. 1989. Rapid dinitrogen fixation during soybean pod fill enhances net photosynthetic output and seed yield: A new perspective. Agron. J. 81:549‑556.

Ireland, J.A., and J.M. Vincent. 1968. A quantitative study of competition for nodule formation. p. 85‑93. In Transactions of the 9th International Congress on Soil Science, Adelaide, Vol. 2. International Society of Soil Science and Angus and Robertson, Sydney, Australia.

Jensen, E.S. 1987. Inoculation of pea by application of Rhizobium in the planting furrow. Plant Soil 97:63‑70.

Johnson, H.W., U.M. Means, and C.R. Weber. 1965. Competition for nodule sites between strains of Rhizobium japonicum. Agron. J. 57:179‑185.

Jones, J.W., K.J. Boote, G. Hoogenboom, S.S. Jagtap, and G. G. Wilkerson. 1989. SOYGRO v. 5.42: Soybean crop growth simulation model. User's Guide. Florida Agricultural Experiment Station Journal No. 8304, Gainesville, FL.

Keeney, D.R. 1982. Nitrogen‑availability indices. In A.L. Page, R.H. Miller, and D.R. Keeney (ed.) Methods of Soil Analysis. Part 2. Agronomy 9:711‑733.

Keeney, D.R., and D.W. Nelson. 1982. Nitrogen‑inorganic forms. In A.L. Page, R.H. Miller, and D.R. Keeney (ed.) Methods of Soil Analysis. Part 2. Agronomy 9:643‑698.

Keyser, H.H., D.N. Munns, and J.S. Hohenberg. 1979. Acid tolerance of rhizobia in culture and in symbiosis with cowpea. Soil Sci. Soc. Am. J. 43:719‑722.

Keyser, H.H., and P.B. Cregan. 1987. Nodulation and competition for nodulation of selected soybean genotypes among Bradyrhizobium japonicum serogroup 123 isolates. Appl. Environ. Microbiol. 53:2631‑2635.

Kishinevsky, B., R. Lobel, and Y. Friedman. 1984. Symbiotic performance and efficiency evaluation of different peanut Rhizobium strains under field conditions. Oleagineux 39:417‑421.

Klubek, B.P., L.L. Hendrickson, R.M. Zablotowicz, J.E. Skwara, E.C. Varsa, S. Smith, T.G. Islieb, J. Maya, M. Valdes, F.B. Dazzo, R.L. Todd, and D.D. Walgenback. 1988. Competitiveness of selected Bradyrhizobium japonicum strains in midwestern USA soils. Soil Sci. Soc. Am. J. 52:662‑666.

Kluson, R.A., W.J. Kenworthy, and D.F. Weber. 1986. Soil temperature effects on competitiveness and growth of Rhizobium japonicum and on rhizobium‑induced chlorosis of soybeans. Plant Soil 95:201‑207.

Kosslak, R.M., and B.B. Bohlool. 1985. Influence of environmental factors on interstrain competition in Rhizobium japonicum. Appl. Environ. Microbiol. 49:1128‑1133.

Kucey, R.M.N. 1989. The influence of rate and time of mineral N application on yield and N₂ fixation by field bean. Can. J. Plant Sci. 69:427‑436.

Kvien, C.S., G.E. Ham, and J.W. Lambert. 1981. Recovery of introduced Rhizobium japonicum strains by soybean genotypes. Agron. J. 73:900‑905.

Kvien, C.S., and G.E. Ham. 1985. Effect of soil temperature and inoculum rate on the recovery of three introduced strains of Rhizobium japonicum. Agron. J. 77:484‑489.

Lowendorf, H.S. 1980. Factors affecting survival of Rhizobium in soil. p. 87‑123. In M. Alexander (ed.) Advances in Microbial Ecology, Vol. 4. Plenum Publishing Corp., New York.

Lynch, J.M., and M. Wood. 1988. Interactions between plant roots and micro‑organisms. p. 526‑563. In A. Wild (ed.) Russell's Soil Conditions and Plant Growth. 11th ed. Longman Scientific & Technical, Essex, England.

Major, D.J., D.R. Johnson, and V.D. Luedders. 1975. Evaluation of eleven thermal unit methods for predicting soybean development. Crop Sci. 15:172‑174.

Materon, L.A., and J.M. Vincent. 1980. Host specificity and interstrain competition with soybean rhizobia. Field Crops Res. 3:215‑224.

McIntosh, M.S. 1983. Analysis of combined experiments. Agron. J. 75:153‑155.

McMiel, D.L. 1982. Variations in ability of Rhizobium japonicum strains to nodulate soybeans and maintain fixation in the presence of nitrate. Appl. Environ. Microbiol. 44:647‑652.

Meade, J., P. Higgins, and F. O'Gara. 1985. Studies on the inoculation and competitiveness of a Rhizobium leguminosarum strain in soils containing indigenous rhizobia. Appl. Environ. Microbiol. 49:899‑903.

Mirza, N.A., B.B. Bohlool, and P. Somasegaran. 1990. Nondestructive chlorophyll assay for screeening of strains of Bradyrhizobium japonicum. Soil Biol. Biochem. 22:203‑207.

Moawad, H., and B.B. Bohlool. 1984. Competition among Rhizobium spp. for nodulation of Leucaena leucocephala in two tropical soils. Appl. Environ. Microbiol. 48:5‑9.

Moawad, H., W.R. Ellis, and E.L. Schmidt. 1984. Rhizosphere response as a factor in competition among three serogroups of indigenous Rhizobium japonicum for nodulation of fieldgrown soybeans. Appl. Environ. Microbiol. 47:607‑612.

Munns, D.N., H.H Keyser, V.W. Fogle, J.S Hohenberg, T.L. Righetti, D.L. Lauter, M.G. Zaroug, K.L. Clarkin, and K.W. Whitacre. 1979. Tolerance of soil acidity in symbioses of mung bean with rhizobia. Agron. J. 71:256‑260.

Odum, E.P. 1971. Fundamentals of Ecology, 3rd ed.. W.B. Saunders Co., Philadelphia, PA.

Parkinson, M.S., and S.E. Allen. 1975. A wet oxidation procedure suitable for the determination of nitrogen and mineral nutrients in biological material. Commun. Soil Sci. Plant Anal. 6:1‑11.

Peoples, M.B., A.W. Faizah, B. Rerkasem, and D.F. Herridge. 1989. Methods for Evaluating Nitrogen Fixation by Nodulated Legumes in the Field. ACIAR, Monograph No. il, Canberra, A.C.T.

Piha, M. I., and D.N. Munns. 1987. Nitrogen fixation capacity of field‑grown bean compared to other grain legumes. Agron. J. 79:690‑696.

Read, M.P. 1953. The establishment of serologically identifiable strains of Rhizobium trifolii in field soils in competition with the native microflora. J. Gen. Microbiol. 9:1‑14.

Roughley, R.J., W.M. Blowes, and D.F. Herridge. 1976. Nodulation of Prifolium subterraneum by introduced rhizobia in competition with naturalized strains. Soil Biol. Biochem. 8:403‑407.

Ryle, G.J.A., C.E. Powell, and A.J. Gordon. 1979. The respiratory costs of nitrogen fixation in soyabean, cowpea, and white clover. II. Comparisons of the cost of nitrogen fixation and the utilization of combined nitrogen. J. Exp. Bot. 30:145‑153.

Salado‑Navarro, L.R., T.R. Sinclair, and K. Hinson. 1986a. Yield and reproductive growth of simulated and field‑grown soybean. I. Seed‑filling duration. Crop Sci. 26:966‑970.

Salado‑Navarro, L.R., T.R. Sinclair, and K. Hinson. 1986b. Yield and reproductive growth of simulated and field‑grown soybean. II. Dry matter allocation and seed growth rates. Crop Sci. 26:971‑975.

Salisbury, F.B., and C.W. Ross. 1985. Plant Physiology. 3rd ed. Wadsworth Publishing Co, Inc., Belmont, CA.

SAS Institute. 1986. SAS user's guide: Statistics. SAS Institute Inc., Cary, NC.

Schwinghamer, E.A., and J. Brockwell. 1978. Competitive advantage of bacteriocin‑ and phage‑producing strains of Rhizobium trifolii in mixed culture. Soil Biol. Biochem. 10:383‑387.

Silsbury, J.H. 1977. Energy requirement for symbiotic nitrogen fixation. Nature 267:149‑150.

Sinclair, T.R., R.C. Muchow, M.M. Ludlow, G.J. Leach, R.J. Lawn, and M.A. Foale. 1987. Field and model analysis of the effect of water deficits on carbon and nitrogen accumulation by soybean, cowpea, and black gram. Field Crops Res. 17:121140.

Singleton, P.W. 1983. A split‑root growth system for evaluating components of the soybean‑Rhizobium japonicum symbiosis. Crop Sci. 23:259‑262.

Singleton, P.W., H.M. AbdelMagid, and J.W. Tavares. 1985. Effect of phosphorus on the effectiveness of strains of Rhizobium japonicum. Soil Sci. Soc. Am. J. 49:613‑616.

Singleton, P.W., and B.B. Bohlool. 1983. The effect of salinity on the functional components of the soybean‑Rhizobium japonicum symbiosis. Crop Sci. 23:815‑818.

Singleton, P.W., and K.R. Stockinger. 1983. Compensation against ineffective nodulation in soybean. Crop Sci. 23:69‑72.

Singleton, P.W., and J.W. Tavares. 1986. Inoculation response of legumes in relation to the number and effectiveness of indigenous rhizobium populations. Appl. Environ. Microbiol. 51:1013‑1018.

Soil Conservation Service. 1972. Soil survey of the islands of Kauai, Oahu, Maui, Molokai, and Lanai; State of Hawaii. Soil Conservation Service, U.S. Department of Agriculture, Washington, D.C.

Soil Conservation Service. 1984. Soil survey, laboratory data, and descriptions for some soils of MauiNet. Soil Conservation Service, U.S. Department of Agriculture, Washington D.C.

Somasegaran, P., and H. Hoben. 1985. Methods in legume Rhizobium technology. University of Hawaii NifTAL Project, Paia, HI.

Sparrow, S.D., and G.E. Ham. 1983. Nodulation, N₂ fixation, and seed yield of navy beans as influenced by inoculant rate and inoculant carrier. Agron. J. 75:20‑24.

Summerfield, R.J., P.J. Dart, P.A. Huxley, A.R.J. Eaglesham, F.R. Minchin, and J.M. Day. 1977. Nitrogen nutrition of cowpea (Vigna unguiculata). I. Effects of applied nitrogen and symbiotic nitrogen fixation on growth and seed yield. Expl. Agric. 13:129‑142.

Sutton, W.D. 1983. Nodule development and senescence. p. 144‑212. In W.J. Broughton (ed.) Nitrogen Fixation Vol. 3: Legumes. Oxford Univ. Press, New York.

Torres, R.O., R.A. Morris, and D. Pasaribu. 1987. Inoculation methods and nitrogen fertilizer effects on soybean in the Philippines: I. nodulation and nitrogen yields. Trop. Agric. (Trinidad) 65:219‑225.

Triplett, E.W., and T.M. Barta. 1987. Trifolitoxin production and nodulation are necessary for the expression of superior nodulation competitiveness by Rhizobium leguminosarum bv. trifolii strain T24 on clover. Plant Physiol. 85:335‑342.

Vincent, J.M. 1970. A Manual for the Practical Study of Root Nodule Bacteria. Blackwell Scientific Publications, Oxford.

Wadisirisuk, P., S.K.A. Danso, G. Hardarson, and G.D. Bowen. 1989. Influence of Bradyrhizobium japonicum location and movement on nodulation and nitrogen fixation in soybeans. Appl. Environ. Microbiol. 55:1711‑1716.

Wann, M., and C.D. Raper, Jr. 1979. A dynamic model for plant growth: Adaptation for vegetative growth of soybeans. Crop Sci. 19:461‑467.

Weaver, R. W., and L.R. Frederick. 1974a. Effect of inoculum rate on competitive nodulation of Glycine max L. Merrill. I. Greenhouse studies. Agron. J. 66:229‑232.

Weaver, R.W., and L.R. Frederick. 1974b. Effect of inoculum rate on competitive nodulation of Glycine max L. Merrill. II. Field studies. Agron. J. 66:233‑236.

Weber, D.F., and V.L. Miller. 1972. Effect of soil temperature on Rhizobium japonicum serogroup distribution in soybean nodules. Agron. J. 64:796‑798.

Wilkinson, L. 1988. SYSTAT: The System for Statistics. SYSTAT, Inc., Evanston, IL

Woomer, P., P.W. Singleton, and B.B. Bohlool. 1988. Ecological indicators of native rhizobia in tropical soils. App1. Environ. Microbiol. 54:1112‑1116.

Woomer, P., J. Bennet, and R.S. Yost. 1990. Overcoming the inflexibility in most-probable-number procedures. Agron. J. 82:349-353.

Table 2.2 Most-Probable-Number counts^a of indigenous, homologous rhizobia for legumes grown in 8 inoculation trials conducted at 5 sites on Maui, HI.
No.	Legume Species Site G. P. V. P. A. L. M. T. L. Name max lunatus unguiculata vulgaris hypogaea leucocephala sativa repens tingeatus
	Rhizobial/g soil^b
1 1a	Hashimoto 0 <1 54 7 - - - - - Farm - - - - 5^c >1650 - - -

2	Kuiaha 0 61^e 2306^e 93^e - - - - -

3 3a	Kula 0 <1^d 18^c 2^d - - - - - Agric. Park 0 - - 211^d 5^c >5938^d - - -

4	Haleakala 0 311^d 35900^e 437^d - - - - - Station
5 5a	Tengan 0 23 283 31 - - - - - Farm - - - - - - 1038 <1 15
a b c d e	Calculated y the Most Probable Number Estimation System , oomer et al., 1990) Upper and lower fiducial limits are determined by dividing or multiplying by 2.7, unless otherwise noted Upper and lower fiducial limits are determined by dividing or multiplying by 2.0 Upper and lower fiducial limits are determined by dividing or multiplying by 2.9 Upper and lower fiducial limits are determined by dividing or multiplying by 3.8

Table 2.5 Summary of nodulation responses to inoculation in relation to the most probable number (MPN) of indigenous rhizobia.
	MPN of soil rhizobia Observations		Significant increase due to inoculation in nodule parameters: mass number		Ratio of Inoculated to Uninoculated yield of nodule parameters: mass number		Average nodule occupancy by inoculant strains
		----no. of trials----		----fold increase----		---- % ----
0 - 10 13		11 9		17.6^a 36.2^a		89
10 - 100 7		4 3		2.1^b 2.7^b		86
>100 9		2 2		1.1^c 1.3^c		53
Total 29		17 14
a Excludes soybean data. b Excludes bush bean at site 2. c Excludes bush bean at site 4.

Table 3.1 Summary of measures of soil N availability in the Maui inoculation trials.
Site No. Name	Soil Variables N Mineralization Total N (ug N/g soil/wk) (%)		Crop Variables Non-nodulating Soybean N Accumulation Seed Yield (kg N/ha/da) (kg/ha)	N Derived from N2 fixation (%)

1 Hashimoto Farm	7.0 (0.4)^a	0.0753 (.0004)^a	0.415 (0.156)^a 627 (282)^a	82.0 (7.9)^a
2 Kuiaha	27.4 (0.9)	0.2527 (.0052)	0.583 (0.044) 840 (53)	76.3 (3.5)
3 Kula Agric. Park 3a	17.5 (0.5) 24.3 (1.2)	0.1512 (.0022) 0.1448 (.0033)	0.382 (0.118) 485 (127) 0.523 (0.082) 935 (236)	80.3 (5.6) 75.6 (6.6)
4 Haleakala Station	44.1 (2.0)	0.3163 (.0074)	1.100 (0.051) 1711 (98)	58.2 (3.5)
5 Tengan Farm	20.9 (1.2)	0.1906 (.0008)	0.951 (0.225) 1356 (269)	15.5 (21.4)
a Standard error of the mean.

Table 3.3 Measures of soil N availability in the Maui inoculation trials and their relationship to the slope coefficient (b0) in the hyperbolic-response model: Response = b0 * 1/(1 + number of indigenous rhizobia).
Relationship to b0	Measure of Soil N Availability (MSA)	Units	Coefficients b1 b2	Residual Mean Square	Correlation of Observed vs Predicted Values r
Linear: b0 = b1 + b2(MSA)	N Mineralization	ug N/g soil/wk	314.7 -5.1	3680.1	0.83
	Total Soil N	%	335.6 -742.3	3658.5	0.83
	Seed Yield of Non-nod Soybean	kg/ha	422.4 -0.2	2329.3	0.91
	N Accumulated by Non-nod Soybean	kg N/ha/da	440.0 -364.3	2048.7	0.92
	N derived from N fixation	%	-87.5 4.5	1510.0	0.94
Logarithmic: b0 = b1 + b2(log(MSA))	N Mineralization	ug N/g soil/wk	535.2 -259.4	3680.1	0.84
Exponential: b0 = b1(exp^b2(MSA))	N derived from N fixation	%	7.3 0.05	1211.9	0.96

Table 3.4 Soil N deficit factors in the Maui inoculation trials and their relationship to the slope coefficient (b0) in the hyperbolic-response model: Response = b0 * 1/(1 + number of indigenous rhizobia).
Relationship to b0	Measures of crop N demand and soil N supply (NDEM;NSUP)	Units	Coefficients b1 b2	Residual Mean Square	Correlation of Observed vs Predicted Values r
Fractional decline: b0 = b1 + [NDEM - b2(NSUP) NDEM]	N Accumulated by N fertilized plants; N Mineralization	kg N/ha/da ug N/g soil/wk	397.6 0.05	2865.2	0.86
	N Accumulated by N fertilized plants; Total Soil N	kg N/ha/da %	360.5 4.9	3267.7	0.83
	Yield of N fertilized plants; N Mineralization	kg/ha ug N/g soil/wk	388.3 67.1	3016.2	0.89
	Yield of N fertilized plants; Total Soil N	kg/ha %	369.0 7640.7	3255.3	0.87
Fractional decline: b0 = b1 * [(NDEM - NSUP) NDEM]	N Accumulation of N fertilized plants; N Accumulation of Non-nod Soybean	kg N/ha/da kg N/ha/da	317.3	2830.9	0.88
	Yield of N fertilized plants; Yield of Non-nod Soybean	kg/ha kg/ha	326.5	2425.1	0.90