rights / license: research collection in copyright - …...importance of vegetable production in the...
TRANSCRIPT
Research Collection
Doctoral Thesis
Nitrogen fertilizer substitution for tomato by legume greenmanures in tropical vegetable production systems
Author(s): Thönnissen Michel, Carmen
Publication Date: 1996
Permanent Link: https://doi.org/10.3929/ethz-a-001616306
Rights / License: In Copyright - Non-Commercial Use Permitted
This page was generated automatically upon download from the ETH Zurich Research Collection. For moreinformation please consult the Terms of use.
ETH Library
Thesis ETHZ No. 11626
Nitrogen fertilizer substitution for tomato by legume green
manures in tropical vegetable production systems
A dissertation submitted to the
Swiss federal Institute of technology Zurich
for the degree of
doctor in Natural Science
presented by
Carmen THONNISSEN MICHEL
Dipl. Ing. Agr. ETHZ
born 14 My 1966
from Arbaz (VS) and Courtedoux (JU)
Accepted by
Prof. Dr. P. Stamp, Examiner
Dr. U. Schmidhalter, Co-examiner
Dr. D. Midmore, Co-examiner
Dr. J.K. Ladha, Co-examiner
Zurich 1996
Table of Contents
1 General Introduction 1
2 Green Manure N Accumulation and Effects of Age and
Composition on Net Mineralization 7
Abstract 7
Introduction 8
Materials and Methods 9
Results 12
Discussion 21
Conclusions 24
3 Legume Decomposition and N Release when Applied as Green
Manure to Tropical Vegetable Systems. I. In the Wet Season
in Taiwan 27
Abstract 27
Introduction 28
Materials and Methods 29
Results 33
Discussion 45
Conclusions 49
4 Legume Decomposition and N Release when Applied as Green
Manure to Tropical Vegetable Systems. II. In the Dry Season
in Taiwan 53
Abstract 53
Introduction 54
Materials and Methods 55
Results 59
Discussion 71
Conclusions 74
5 Legume Decomposition and N Release when Applied as Green
Manure to Tropical Vegetable Systems. III. Two On-Farm
Studies in the Philippines 78
Abstract 78
Introduction 79
Materials And Methods 80
Results 86
Discussion 96
Conclusions 100
6 General Discussion 104
7 Summary 109
8 Zusammenfassung 112
Acknowledgments
Curriculum Vitae
1
1
General Introduction
Importance of vegetable production in the tropics
Vegetable production is profitable and the importance of vegetables as sources of
minerals and vitamins has widely been recognized, however, both production and
consumption m most developing countnes are low The daily per capita availability of
vegetables in developing countries, except China, is, at best, only one-half that for
developed countnes
Production constraints in tropical vegetable production
The main production constraints in vegetable production are the seasonality, stress
(floods, cyclones, drought), shortage of good quality seeds, lack of improved production
technologies, lack of improved vaneties, pest and disease infestation, post harvest losses,
marketing problems, slow technology transfer (Hossain, 1992a) Many technical,
socioeconomic and institutional factors inhibit the production, distnbution and
consumption of vegetables in developing countries (AVRDC, 1992a) Natural resources
are intensively used and inputs per unit land area is usually high in vegetable production
General awareness among growers of response to inputs and their ready availability,
often leads to overuse, particularly of fertilizers and pesticides
Renewed interest in the practice of green manuring
The reduced availability of land due to rapid population growth has resulted in
over-exploitation, rapid soil degradation, and declining crop yields, particularly in the
humid tropics (Ruthenberg, 1980) where shifting cultivation was the traditional type of
farming Loss of soil fertility must be compensated by prolonged or continuous
cropping Because most shifting cultivators are subsistence farmers the possibilities of
using chemical fertilizers on a large scale is limited It is also unlikely that inorganic
fertilizers will be available in sufficient quantities in the foreseeable future, in areas where
the fertilizer to crop price ratios are relatively high Among the various ways to add
organic matenal to the soil, the use of in situ mulch and green manure appears to be a
practical proposition (Mulongoy and Akobundu, 1985)
In the past three decades, fertilizers have been the most commonly used source for
supplying nutnents to crops In developing countnes such as India, fertilizei pnces have
2
been subsidized, thereby encouraging farmers to apply fertihzeis in production-
maximizing doses The low cost, and ready availability of chemical fertilizers have lead
woild wide to a drastic decline of organic manure use, including green manures, which
were the traditionally important sources of nutnents As nonrenewable resource energy
reserves are depleted, energy pnces will increase and the negotiations to allocate scarce
supplies for various uses (including the production of chemical fertilizers) will
undoubtedly become more contentious (Lanyon, 1995) In face of a continuing eneigy
crisis, increasing fertilizer pnces, and growing concern for environmental quality, there
has been a tremendous renewal of interest in the old practice of green manuring (Singh et
al, 1992)
The ability ot certain crops to improve soil fertility or soil physical conditions has
long been recognized As early as the Chow Dynasty (1134-247 B C ) in China, there
were reports of crops whose value for soil improvement was "greater than silk worm
excrement' (Pieters, 1927) Romans around the time of Chnst likewise waxed eloquently
on the value of green manures, as in this line from Virgil "sow your wheat on land
where giew the bean, the slender vetch or the fragile stalks of the bitter lupine"
(Sarrantomo, 1992) The benefits credited to green manure crops include increases in
organic matter content and available plant nutrients and improvement in the
microbiological and physical properties of the soil (Singh et al, 1992) Of these the role
of green manures in supplymg plant nutnents, particularly N is most prominent
Studies evaluating the fate of I5N from legume residues decomposing under field
conditions led to the conclusions that i) < 30% of legume N was recovered by a
subsequent crop (i e non-legume), n) large amounts of legume N were retained in soil,
mostly in organic forms, m) total recovery ot legume N in crops and soil after 1 year
averaged 70 to 90 %, and iv) < 5% ot legume N from the original application was
recovered by a second crop (Harris et al, 1994, Ladd et al, 1983, Muller and Sundman,
1988)
Green manuring in vegetable production
Green manuring has mostly been studied for staple crops such as nee (e g Singh et
al, 1991), coin (Reeves et al, 1986, Worsham, 1986) wheat (Ladd and Amato, 1986),
sorghum (Hargrove, 1986), or cash crops such as cotton (Brown et al 1985) Research
has been extended to vegetable crops (Stivers and Shennan, 1991, Shennan, 1992, Rayns
and Lennartsson, 1994, Abdul-Baki and Teasdale, 1993) in temperate zones only in
recent years, and to date only little data are available
3
Feasibility of meeting N needs of vegetables with legume green manures
In the present study soybean (Glycine max Merr L) and indigo (Indigofera
ttnctona L) were chosen as two test crops for use as green manures in Southeast Asia
Soybean is widely used as a food crop and seed is readily available The latter is one of
the key factors for green manure production Indigo was chosen as an indigenous green
manure crop used mainly in the northern provinces of the Philippines preceding nee, but
also in India preceding maize, cotton or sugarcane (Mann, 1990)
The tomato crop was chosen as a test vegetable crop in this study, for it is one of
the most cited among vegetables given first pnonty for research While summer (rainy
season) vegetables in Southeast Asia are mostly indigenous (bnnjal, cucurbits, basela,
amaranths), winter (dry season) vegetables are of European / Amencan/ North Asian
origin (cabbage, cauliflower, tomato, radish, beans) Winter vegetables contributed more
than 70 % of the total vegetable production in 1987-88 in Bangladesh (Hossain, 1992b)
Tomato is also one of the most important vegetables in the Philippines In 1986 the total
area planted was 17,790 ha with a production of 143,88 t It is commonly grown in the
lowland rice paddies during the dry season The crop is grown throughout the country
but the bulk of production is concentrated in the northern regions (Ilocos provinces)
Supply and prices fluctuate widely from the dry to the wet season and can vary by a
factor of 10 Average yields in Southeast Asia are relatively low and range between 7 - 60
t ha-1 The average tomato yield in the Philippines was about 91 ha *in 1986 (Sonano et
al, 1989) Major causes of low productivity are shortage of improved seeds, poor
growing practices, and a lack of well-trained and adequately supported extension workers
(Villareal, 1980)
Tomato response to N
Tomato yields usually increase with moderate applications of N both under glass
(Winsor and Long, 1967) and in the field (Palevitch et al, 1965) In some instances, no
response was found to N applied before planting (e g Wilcox, 1964), and there was little
response of field crops to supplementary dressings of N (Reeve et al, 1962) Heavy
applications of N may depress the yield (Adams et al, 1978) The response to applied
nitrogen depends not only on the initial N content of the soil but also dunng cropping on
immobilization and on mineralization of N, or denitnfication
Tomato yields following winter legume cover crops were found to be comparable
with (Stivers and Shennan, 1991) or outyielding those with normal fertilizer doses
(Abdul-Baki and Teasdale, 1993)
The current research initiated in Taiwan from 1992-1994 This study was part ot
the research tocus (AVRDC, 1992b ) of the Production Systems Program of the Asian
4
Vegetable Research and Development Center (AVRDC) in Taiwan AVRDC was
established in 1971 to promote production, marketing and consumption of vegetables in
Asia, and since 1990 the mandate has been extended to activities in Africa and Latin
America Crop and soil management research at AVRDC is directed toward promoting
the sustainable use of both natural resources and inputs One approach is the study on a
system basis, of the agronomic/ physiological and economic interrelations of crop
rotation, intercropping, the application of green manures and erosion control practices in
order to preserve and improve the natural pioduction environment Many studies initiated
in Taiwan are extended to other production areas
In a collaborative research project with the International Rice Research Institute
(IRRI), further field experiments were performed within this study in two of the mam
vegetable production areas in the Philippines from 1994-1995 There were planned to
venfy the applicability of results obtained in Taiwan
Outline and goals of this thesis
The goal of this research work was to test the feasibility of meeting N-needs of
tomato crops with legume green manures The determination of the optimal seeding
density and growth duration for legumes to reach maximum biomass and N accumulation
in a short growth duration were the first steps undertaken Legumes were tested as N
fertilizer substitutes in field tomato production when applied as surface mulch or
mcorpoiated into the soil Legume decomposition and N-release patterns were studied in
the soil after green manure application in ordei to understand ongoing processes in the
field and under controlled conditions, and to improve the synchronization of N-ielease
with N-uptake Legume N was traced in tomato and soil organic matter in an 15N
experiment in the field
REFERENCES
Abdul-Baki, A A,and J R Teasdale 1993 A no-tillage tomato production system
using hairy vetch and subterranean clover mulches HortScience 28(2) 106-108
Adams, P,C J Craves, and G W Wmsor 1978 Tomato yields in relation to the
nitrogen, potassium and magnesium status of the plants and of the peat substrate J
HortScience 49 137-149
AVRDC 1992a Asian Vegetable Research and Development Center Information leaflet
AVRDC, Shanhua, Tainan, Taiwan
AVRDC 1992b Translating strategy into action An action plan for 1993 - 1997
AVRDC, Taipei, Taiwan
5
Brown, S M,J T Whitwell, J T Touchton, and C H Burmester 1985 Conservation
tillage systems for cotton production Soil Sci Soc Am J 49 1256-1260
Hargrove, W L 1986 Winter legumes as a nitrogen source for no-till grain sorghum
Agron J 78 70-74
Harns, G H ,O B Hesterman, E A Paul, S E Peters, and R R Janke 1994 Fate of
legume and fertilizer nitrogen-15 in a long-term cropping systems experiment
Agron J 86 910-915
Hossam, S M M 1992a Status, constraints and strategies of vegetable research p 31
41 In AVRDC (ed ) Vegetable production and marketing Proc of a national
review and planning workshop AVRDC, Shanhua, Tainan, Taiwan
Hossain, MAE 1992b Risk of off-season vegetable cultivation in Bangladesh p
139-146 In AVRDC (ed ) Vegetable production and marketing, Proc of a national
review and planning workshop AVRDC, Shanhua, Tainan, Taiwan
Ladd, J N,M Amato, R B Jackson and J H Butler 1993 Utilization by wheat crops
of nitrogen from legume residues decomposing in soils in the field Soil Biol
Biochem 15 231-238
Ladd, J N, and M Amato 1986 The fate of nitrogen from legume and fertilizer sources
in soils successively cropped with wheat under field conditions Soil Biol
Biochem 18 417-425
Lanyon, LE 1995 Does nitrogen cycle9 Changes in the spatial dynamics of nitrogen
with industrial nitrogen fixation J Prod Agnc 8(1) 70-78
Mann, R A 1990 The sustainability of wheat-nce cropping systems Use of Indigofera
tmctona L intercropped with wheat as a green manure for the nee crop Ph D
diss Umv of the Philippines, Los Banos
MullerMM and V Sundman 1988 The fate of nitrogen (15N) released from different
plant materials during decomposition under field conditions Plant Soil 105 133-
139
Mulongoy, K ,and I O Akobundu 1985 Nitrogen uptake of maize in live mulch p
285-290 In B T Kang, and J Van der Heide (eds ) Nitrogen management in
farming systems in humid and subhumid tropics Ibadan, Nigeria
Palevitch, D,N Kedar, H Koyumdjisky, and J Hagm 1965 The effect of manure
and fertilizer treatments on the yields of winter tomatoes in the Western Negev
Israel J Agnc Res 15(2) 65-72
Pieters, A J 1927 Green manunng Wiley, New York
Rayns, F W,and E K M Lennartsson 1994 The effects of gieen manures on nitrate
leaching in organic horticultural systems In press Biol Agnc Hort
6
Reeve, E,W A Robbins, W S Taylor, and J F Kelly 1962 Cultural and nitrogen
fertilization practices in relation to tomato fruit set and yield p 129-147 In Pioc
PI Sci Symp Camden, New Jersey Campbell Soup Company
Reeves, DW.CB Rickerl, C B Elkins, and J T Touchton 1986 No-tillage update
report - Alabama p In RE Philipps (ed ), Southern Region No-Till Conference
Lexington, KY
Ruthenberg, H 1980 Farming systems in the tropics 3d ed Clarendon Press, Oxford,
UK
Sarrantomo, M 1992 Opportunities and challenges for the inclusion of soil-improving
crops m vegetable production systems HortScience 27(7) 754 758
Shennan, C 1992 Cover crops, nitrogen cycling and soil properties in semi-irrigated
vegetable production systems HortScience 27(7) 749-754
Singh Y ,C S Khind and B Singh 1991 Efficient management of leguminous green
manures in wetland rice Adv Agron 45 135-189
Singh Y ,B Singh, and C S Khind 1992 Nutrient transformations in soils amended
with green manures Adv Agron 20 237 309
Soriano, J M,R L Villareal, and V P Roxas 1989 Tomato and pepper production in
the Philippines In AVRDC (ed ) Tomato and pepper production in the tropics
Pioc Int Sym on integrated management practices Shanhua, Tainan 21-26
March 1988 AVRDC, Taiwan
Stivers, L J,and C Shennan 1991 Meeting the nitrogen needs of processing tomatoes
through winter cover cropping J Prod Agnc 4(3) 330 335
Villareal, RL 1980 Tomato production in the tropics IADS development oriented
literature series
Wilcox, G E 1964 Effect of potassium on tomato growth and production Proc Am
Soc Hort Sci 85 484-489
Winsor, GW, and MIE Long 1967 The effects of nitrogen, phosphorus,
potassium, magnesium, and lime in factorial combination on ripening disorders of
glasshouse tomatoes J Hort Sci 42 391-402
Worsham, AD 1986 No-tillage research update - North Carolina In R E Philipps
(ed ), Southern Region No Till Conference Southern Region Series Bulletin 319
Lexington, KY
7
2
Green Manure N Accumulation and Effects of Age
and Composition on Net Mineralization.
ABSTRACT
Nitrogen contribution of leguminous green manures to succeeding crops depends
on their ability to accumulate high amounts of biomass and N in a short time, and their
ability to decompose at a rate which matches the N needs of the subsequent crop Factors
affecting legume biomass and N accumulation, such as seeding density, growth duration,
and season were evaluated in a field experiment for Medicago sativa L,Desmodium
intortum (Mill) Urb, Indigofera tinctona L and Glycine max (L) Merr 60, 75, and
90 days after sowing (d) The influence of these factors on selected chemical components
of the plants and on N release into the soil were estimated by determining N, C, lignin,
polyphenol, and tannin concentrations in the plant material Nitrogen release in the soil
was investigated in an aerobic incubation experiment with tropical soils (a silty loamy,
mixed, hyperthermic Fluvaquentic Entochrept, a clayey, kaolmitic, isohypertheimic
Ultisol, and a clayey, mixed, isohypertheimic Fluvaquentic Ustropept) from Taiwan and
the Philippines
Legume species, seeding density, growth duration, and season were key factois
affecting biomass production Highest biomass in both seasons (wet and dry) was
achieved by soybean (4 2 to 9 9 t dry matter ha-1) with a total N uptake of 144 to 314 kg
N ha"1 at 75 d, and by indigofera (3 9 to 5 11 dry matter ha J) with 101 to 160 kg N ha"1
at 90 d Nitrogen lelease was faster with plant material harvested at 60 d than at 90 d
Initial N concentration and C/N were the two major factors driving net N mineralization
of plant material in two soils Soil chemical properties such as high pH, low P
concentration, and high clay content may have slowed down N transformation processes
in the clayey, mixed, isohyperthermic Fluvaquentic Ustropept soil Rapid N
mineralization after the addition of 60-d-old plant material in two soils indicate the
possibility of substituting N fertilizer with soybean and indigofera legume green
manures
8
INTRODUCTION
Recent attempts to evaluate the usefulness of legume green manures (GM) in the
context of agricultural sustainability have been hindered by a lack of information on
nutrient release patterns (Singh et al, 1992) The ability of legumes to accumulate large
amounts of N in short duration is desirable due to shortage of land and time in intensively
cropped systems Nitrogen accumulation of GM crops varies with legume species,
environmental conditions, soil fertility, and management practices One key decision in
green manuring is choosing the growth stage of plants at which a GM crop is
incorporated into the soil to obtain a synchronized pattern of N release and N uptake by
the subsequent crop Estimates of biomass production and N accumulation in different
symbiotic systems range between 2 9 - 8 9 t ha'1 dry matter and 60 - 225 kg N ha"1 for
tropical legumes grown for 50 to 60 d, and 4 - 241 biomass ha"1 and 40 - 240 kg N ha-1
for food legumes (Singh et al, 1992) The range of these estimates is very broad and
site-specific, and studies on the specific influence of seeding density and plant age on
biomass accumulation and plant chemical properties of legumes for green manure use are
few
Plant chemical properties such as the C/N, N, lignm, polyphenol, and tannin
concentrations have been found to influence N mineralization dynamics in the soil (Fox et
al, 1990, Palm and Sanchez, 1991, Oglesby and Fownes, 1992, Becker et al, 1994,
Clement et al, 1995) Rates of N mineralization differ among species having varying leaf
chemistry (Palm and Sanchez, 1991) and among tissue types (e g leaves, stems, roots)
within a species (Frankenberger and Abdelmagid, 1985) No consensus exists on the
relative influence ot specific plant properties or plant property combinations on N
mineralization Cumulative net N mineralization was found to be negatively correlated
with initial soluble polyphenol concentration in early decomposition phases, and with
initial hgnin concentration in later phases (Oglesby and Fownes, 1992) Factors such as
incubation time on initial plant properties determining N mineralization have only been
studied marginally
Decomposition and N release from organic matenals are subject to the influence of
various soil properties, such as clay content (Sorensen, 1975, Burns, 1978), soil pH
(Alexander, 1977) There seems to be no single factor controlling the late of N release
from green manure in different soils Very few studies relate the influence of chemical
plant properties on N release to more than one soil which makes conclusions difficult To
implement GM practices to sustain soil fertility and/ or to substitute N fertilizei, it is
essential to understand factors that govern N transformation processes in difteient soil
types
9
The objectives of this study were 1) to determine the influence of seeding density,
age, and growing season on biomass and N accumulation and chemical composition of
four legume species in southern Taiwan, and n) to assess the influence of GM chemical
composition on NH4 and NO3 mineralization in three tropical soils from Taiwan and the
Philippines in early mineralization stages
MATERIALS AND METHODS
Field experiments
Biomass and N accumulation at 60, 75, 90 days after sowing (d) of four legume
species - alfalfa (Medicago sativa L), desmodium {Desmodium intoitum (Mill) Urb ),
mdigofera (Indigofera tmctona L ) and soybean {Glycine max (L ) Merr) - were
compared in a field experiment in 1993 at the Asian Vegetable Research and Development
Center in Taiwan Each species was planted at two seeding rates (normal and double)
0 75 and 1 5 g nr2 for desmodium, 0 5 and 1 g nr2 for alfalfa, 0 66 and 1 32 g nr2 foi
indigofera, and 40 and 80 seeds nr2 for soybean Seeds were inoculated with a
rhizobium strain mixture that was specific for each legume species, provided by the Soil
Science Department of the Chung Hsing University in Taichung, Taiwan Inorganic P
and K fertilizers, 35 and 90 kg ha"1, respectively, were broadcast and incorporated before
sowing The field experiment was earned out in two seasons in the wet season (19
March 1993 to 15 June 1993) rainfall totaled 863 mm and mean temperature averaged
24 9°C, in the dry season (7 September 1993 to 9 December 1993) rainfall totaled 18 mm
and mean temperature averaged 24°C Mean air temperature increased from 21 to 28"C
during the wet season, and decreased from 29 to 19°C dunng the dry season The eight
treatments (four legume species at two seeding densities) were arranged as a two-factorial
experiment in a randomized complete block design with four replications The plots were
1 5 m by 4 m At 60, 75, and 90 d a 1 m2 area per plot was harvested to determine
legume above ground biomass Samples were oven-dned at 60°C tor 48 h
Plant material was analyzed for total N (Kjeldahl), total carbon with a Carlo-ERBA-
Gas analyzer (Carlo ERBA Strumentazione, Nitrogen analyzer 1500, Cable Erbadas,
Milan, Italy), and ligmn concentration with the acid detergent fiber method (Goenng and
Van Soest, 1970) Polyphenols were extracted with 1% HC1 in methanol solution and
determined using the Folin & Ciocalteu reagent with tannic acid as a standard (Singleton
and Rossi, 1965) Condensed tannins were analyzed in the extract (methanol HC1=10 1)
using the vanillin assay method (Broadhurst and Jones, 1978) with catechin as a
standard
10
Dry matter and N data were subjected to analysis of variance (ANOVA) using the
JMP (SAS for Macintosh) program (SAS Institute Inc, 1989) The interaction between
treatments and time was tested using the repeated measure analysis (Rowell and Walters,
1976), where the growth responses over tune are compared using orthogonal polynomial
contrasts
Incubation study
Soil andplant material
Soil was collected from the top 10 cm on the experimental farms of AVRDC, the
Mariano Maicos State University (MMSU, Ilocos Norte, Philippines) and Bukidnon
Resources Corporation Inc (BRCI, Mindanao, Philippines)
Soil properties are listed in Table 1 AVRDC soil was a silty loamy, mixed,
hyperthermic Fluvaquentic Entochiept, BRCI soil was a clayey, koahmtic,
isohyperthermic Ultisol, and MMSU soil a clayey, mixed, isohypeithermic Fluvaquentic
Ustropept Soil was sieved (4 mm) and homogenized Fresh legume shoot samples (stem
with leaves) of indigofera and soybean were collected at 60 and 90 d from the double
density treatments of the field experiment during the 1993 wet season These were oven-
dried at 60°C for 48 h and ground (<1 5 mm)
Table 1. Properties ot soils from experimental sites of the Asian Vegetable Research
and Development Center (AVRDC) in Taiwan, the Bukidnon Resources Corporation Inc
(BRCI) and the Mariano Marcos State University (MMSU) in the Philippines
Soil
Soil property AVRDC BRCI MMSU
Clay kg kg1 0 16 0 59 0 53
Silt kg kg"1 0 52 0 25 0 33
Sand kg kg"1 0 32 0 16 0 14
pH (H20, 1 1) 76 6 1 8 1
EC (1 1) dSml 1 6 02 03
C (Walkley-Black) gkg1 70 19 5 59
Kjeldahl N gkg"1 09 2 1 07
Olsen P mg kg"1 24 0 37 0 2 1
K cmol kg-1 0 26 0 50 0 54
exch Ca cmol kg-1 46 37 34 6
Cation exchange cmol kg-1 7 1 12 9 45 2
t Exchange with 0 5 M NH4OAC at pH 7
11
Incubation procedure
Soils were incubated at the Soil Microbiology Laboratory, IRRI, Philippines
(Keeney and Bremner, 1966) For the incubation air-dried soil samples from AVRDC
BRCI, and MMSU were weighed into a 125 ml Erlenmeyer flask at a rate of 40 g per
flask (oven-dry basis) Dned plant material (0 1 g per sample) of 60 and 90 d soybean
and indigofera was mixed into the soil This plant-to-soil ratio approximated a mulch or
incorporation rate of 3 5 t plant dry matter ha-1 The five treatments were one control (no
plant material added) and four residue additions with 60 and 90 d soybean and
indigofera
The flasks were sealed with cotton stoppers and incubated at 25°C Based on
gravimetric determinations moisture content of incubated soil - residue mixture was
maintained between -0 01 and -0 03 MPa by adding deiomzed water every second week
Three replicates of each treatment were randomly selected for extraction of nitrate and
ammonium after 1, 2, 4, 6, 8, and 10 weeks of incubation (wk) Sixty ml of 1M KC1
solution was added into each sampling flask The mixture was shaken for 1 hour and
thereafter filtered with Whatman No 42 filter paper Ammomum-N and NO3-N were
determined with an ammonia gas sensing electrode ORION 95-12 (Siegel, 1980) The
incubations were repeated after three months to confirm results from the fust incubation
The first and second incubations were designated A and B, respectively
Data analysis
The NO3-N and NH4-N contents per sampling date and treatment were analyzed
using ANOVA Net N mineralization and immobilization of N by legumes were
determined by subtracting extractable mineral N (NO3-N or NH4-N) in the control soil
from that amended with legume Stepwise regressions between net mineralized/
immobilized inorganic N and initial legume chemical composition were performed on
untransformed data
12
RESULTS
Field experiments
Biomass and N accumulations of alfalfa, desmodium, and soybean in the wet
season were almost double those in the dry season at comparable growth durations (Table
2) In contrast growth conditions in the dry season were more favorable for indigofera
Soybean produced about 5 t dry matter ha-1 and 100 kg N ha-1 less in the dry season
compared to the wet season
Both legume species and seeding density had significant effects on harvested
biomass and nitrogen (Table 2) Soybean accumulated most biomass and N, followed by
indigofera, desmodium, and altalfa In both seasons, 10 to 30% more biomass was
produced when legumes were planted at double density Nitrogen accumulation in
soybean and indigofera was greater by 20 to 40 kg N ha * with double density,
compared with 5 to 20 kg N ha"' increase for alfalfa and desmodium in both seasons
Interactions between the linear and quadratic effects of time and legume species on
biomass and N accumulation were significant and differed among species and seasons
(Table 3) The growth pattern of soybean was different from those ot the other legumes
Nitrogen accumulation from 75 to 90 d declined in all four legume species in the wet
season compared to the predominantly linear increases in the dry season
Table 3. Repeated measure analysis for the linear and quadratic effect of time, legumespecies (alfalfa, desmodium, indigofera and soybean) and seeding density on legumebiomass and N accumulation when grown for 60, 75, and 90 days after sowing in the
wet (WS) and dry (DS) seasons in Taiwan, 1993
df
Dry matter Nitrogen
Variables WS DS WS DS
Linear (L) 1 *#* *** *ns
L * Species (S) 3 *** *** ** ***
L * Density (D) 1 ns ns ns ns
L*S*D 3 ns *ns ns
Quadratic (Q) 1 *ns *
ns
Q * Species (S) 3 *** * * *
Q * Density (D) 1 ns ns ns ns
Q*S*D 3 ns ns ns ns
*, ** ***, significant at the 0 05, 0 01, and 0 001 levels, respectively, ns, not
significant at the 0 05 level
level.
0.05
the
at
sign
ific
ant
not
ns,
respecuvely;
leve
ls,
0.001
and
0.01,
0.05,
the
at
significant
,
ns
**
***
4108.5
4311.2
190904
313
*D
S
(D)
Density
(S)
Species
148.2
159.7
75.7
38.8
263.1
101.3
58.3
409
double
DS
155.6
124.0
69.8
342
256.2
83.2
44.5
33.7
normal
90
143.9
131.7
45.1
922
314.1
106.7
73.3
67.1
double
ns
746.9
3*D
S121.7
105.6
940
29.1
276.4
93.0
55.1
616
normal
75
*4409.3
1(D)
Density
142.2
90.1
423
16.8
209.3
56.0
247
57.0
double
***
656638
3(S
)Species
97.4
28.2
18.7
16.4
187.7
50.8
39.5
50.0
normal
60
ns
***
3.4
5.4
3
1
WS
S*D
(D)
Density
3.1
5.1
2.9
-1)
1.2
ha
(kg
N
8.3
3.9
3.1
81
double
***
198
3(S
)Species
3.0
4.0
2.4
1.3
7.7
3.0
2.3
1.5
normal
90
DS
4.2
3.7
1.8
0.9
9.9
2.8
3.0
2.3
double
3.3
2.8
1.5
1.0
8.2
2.4
2.3
2.0
normal
75
ns
1.6
3*D
S36
2.2
0.9
0.7
6.1
1.3
1.4
1.7
double
***
8.9
1(D)
Density
2.8
1.3
0.7
0.6
5.3
1.1
1.2
1.4
normal
60
***
5414
3(S
)Species
WS
ha"'
)(t
matter
dry
Significance
squaresof
Sum
df
variationof
Source
Soybean
Indi
gofe
raDesmodium
Alfalfa
Soybean
Indi
gofe
raDesmodium
Alfalfa
(day
s)
1993)
December
-
(October
season
Dry
1993)
June
(Apr
ilseason
Wet
variance
of
Analysis
species
Legume
density
Seed
ing
duration
Growth
1993.
Taiwan,
in
seasons
(DS)
dry
and
(WS)
wet
the
in
densities
seeding
two
at
grown
sowing
after
days
90
and
75,
60,
soybean
and
indi
gofe
ra,
desmodium,
alfa
lfa,
of
ha-1
)(k
gN
and
ha-1)
tmatter,
(dry
biomass
ground
Above
2.Table
2(14)
13
1(15)
17
9(15)
19
8(14)
14
6)
2(0
12
4)
6(0
10
9)
4(0
17
9)
5(0
16
4(10)
16
211(15)
2)6(2
27
6(10)
23
4)
1(0
15
7)4(0
12
7(17)
15
6)
2(0
15
90
60
ratio
C/N
02)
1(0
02)
11(0
13(15)
03)
2(0
0
01)
3(0
01)
19(0
2)18(0
03)
3(0
0
02)
(03
02)
(05
04)
(07
12(10)
0
01)
2(0
08)
16(0
3)1(0
205)
2(0
0
90
60
kg-1)
(gTannin
1)15(0
6)1(0
49)
0(0
51)
18(0
1)13(0
3)2(0
31(10)
71)
14(0
90
(gkg-l)
1)1(0
21)
1(0
42)
9(0
61)
0(0
21)
17(0
3)7(0
38)
3(0
72)
16(0
60
Poly
phenol
6)(0
25
3)(0
84
3)(0
86
3)(0
35
5)(0
88
3)(0
47
3)(0
56
4)(0
46
8)(0
17
9)(0
57
3)(0
97
1)(0
64
5)(0
49
6)(0
96
8)(0
99
6)(0
08
90
60
J)Lignin(gkg
9)
(08
4
2)(0
93
3)(0
13
3)(0
24
2)(0
62
5)(0
62
3)(0
23
3)(0
62
1)(0
23
3)(0
43
2)(0
62
4)(0
24
1)(0
91
3)(0
33
2)(0
22
2)(0
43
90
60
N(gkg-l)
Soybean
Indi
gofe
raDesmodium
Alfalfa
Soybean
Indigo
fera
Desmodium
Alfalfa
(days)
property
Harvest
Chemical
1993)
December
-
(October
season
Dry
1993)
June
-
(April
season
Wet
deviation
standard
indicate
pare
nthe
sis
within
Values
repl
icat
esthree
of
means
are
shown
Values
1993
Taiwan,
AVRDC,
at
seasons
two
in
field
the
in
grown
soybean
and
indi
gofera,
desmodium,
alfalfa,
of
properties
Chemical
4.
Table
15
Legume age, species, and density interactions were significant only for biomass m
the dry season (Table 3) which can be explained as follows the rate of biomass
accumulation was increased with double density compared to normal density for
desmodium and indigofera Seeding density did not affect biomass accumulation of
alfalfa over time The rate of biomass accumulation for soybean slightly increased from
60 to 75 d with double density, whereas it decreased strongly from 75 to 90 d It can be
concluded that to obtain the highest biomass and N accumulation, 75 d is suitable for
soybean whereas for indigofera longer duration (90 d) is required
Lignin concentration and the C/N of plant tissue generally increased between 60
and 90 d, while polyphenol and tannin concentrations decreased in both seasons (Table
4) Nitrogen concentrations decreased between 60 to 90 d in the wet season for all four
legumes, while in the dry season it increased between 60 and 90 d for alfalfa and
soybean, was constant for desmodium, and decreased for indigofera
Lignin concentration was strongly negatively correlated with N and tannm
concentrations and to a lower degree with polyphenol concentrations (Table 5) Tannin
was positively correlated with N and polyphenol
Table 5. Correlation matrix of chemical characteristics of plant matenal (60 and 90 d
soybean and indigo grown in the wet season) used for stepwise regression (n=8)
Variable Nitrogen (%) Lignin (%) Polyphenol (%) Tannin (%)
Lignin (%) 0918 **
Polyphenol (%) 0 277 -0 542
Tannin (%) 0 742 * -0 929 *** 0 811 *
C/Niaao -0 969*** 0 804* -0 038 -0 056
*, ** ***, significant at the 0 05, 0 01, and 0 001 levels, respectively
Based on biomass and N accumulation results of these field studies only 60 and 90
d soybean and indigofera were tested in the incubation experiment
Incubation experiments
Nitrate was the dominant form of extractable inorganic N in all three soils in the
first incubation Effects of GM on net NO3-N and NH4-N contents relative to control soil
N are presented in Figures 1 and 2 The effects of GM on net NO3-N in the second
incubation were broadly similar to those observed in the corresponding treatments of the
first incubation (Fig 1) With the exception of an increase of 20 mg NH4-N kg-1 soil one
05)
(P<0
difference
significant
least
indicate
bars
Error
replicates
three
of
means
are
shown
Values
incubation
the
of
Band
Arun
in
material
plan
td
90
and
60
with
amended
soils
MMSU
and
BRCI,
AVRDC,
in
cont
rol)
the
to
(rel
auve
release
NO3-N
Net
1.Fig.
10
8
(weeks)
period
Incubation
42
010
86
d90
Indi
gofe
ra—•—
d60
Indi
gofe
ra-
O-
—
d90
Soybean
A
d60
Soybean
A--
—
.i
.i
ii
*
MMSU
a)T*
»—ft-
ns.n
zoBRCI
J05
0LSD
I•^
en
i
Zto,c^t^A
st g/kg
AVRDC
,
V
II
•
-60
-40
-200'
20
40
-60
-40
-20
(0
20
40
-60
-40
-200i
20
40
BA
05)
(P<0
difference
significant
least
indicate
bars
Error
reph
cate
sthree
of
means
are
shown
Values
incubation
the
of
Band
Arun
in
matenal
plant
indi
gofera
and
soybean
d90
and
60
with
amended
soils
MMSU
and
BRCI,
AVRDC,
in
control)
the
to
(rel
aUve
release
NH4-N
Net
2.Fi
g.
10
(weeks)
peri
odIncubation
0246
10
02468
MMSU
05
0LSD
I
AVRDC
i.
...
i....
ii
i
AVRDC
d60
Indi
gofe
ra—
o—
d90
Soybean
—*•—
d60
Soybean
&-
—
•
20
s*'So
100
10
20
BA
18
week after the incorporation of 60 d mdigotera into the BRCI soil, NH4-N contents were
comparable to those of the control (Fig 2)
Legume species, age, and soil properties strongly influenced die N release pattern
Significantly more N release occurred tor each species with young plant material (60 d)
A lag period ot one to four wk before active N release commenced occurred in AVRDC
and BRCI soils Nitrogen from plant material hai vested at 90 d was released only aftei
eight wk with the exception of 90 d soybean in AVRDC soil Legume addition resulted in
a net N-immobihzation in MMSU soil, with the exception of 60 d indigofera (Fig 1 and
Fig 2) In all three soils most NO3 was leleased with indigofera harvested 60 d,
followed by 60 and 90 d soybean Of the initial plant N, 5% more N was released with
60 d compaied to 90 d soybean m all three soils Sixty-day-old indigofera released 22%
more of its initial N content than did 90 d indigofera in AVRDC and BRCI soils, whereas
the difference was only 4% in MMSU soil
Net N release correlated best with initial N, and C/N during most time periods in
AVRDC and BRCI soils (Table 6) Weaker hneai correlations of N release were obtained
with hgnin, whereas polyphenols showed no effect on N release The only plant chemical
pioperties correlated with net N release in the MMSU soil were tannin and lignm at two
wk of incubation Multiple 1 egression analysis revealed the relative strength of C/N and
N concentrations for predicting N release depending on soil type and sampling date in
AVRDC and BRCI soils (Table 7) Plant tannin and polyphenol concentrations partly
governed N-release dynamics in MMSU soil, although coefficients ot determination were
low
19
Table 6. Correlation coefficients from line
and 90 d soybean and indigofera grown n
change, 1 to 10 weeks after residue addition
Soils Time Chemical property
(weeks) Nitrogen Lignin
% %
1 0 77 ** -0 65
2 0 82* -0 75
4 0 60 -0 34
6 0 95 *** -0 91
8 0 87 ** -0 83
10 0 83 * -0 71
1 0 58 -0 4
2 0 87 ** -0 82
4 0 85 ** -0 72
6 0 80* -0 64
8 0 96 *** -0 85
10 0 83 ** -0 82
1 0 18 -0 15
2 0 64 -0 75
4 0 26 -0 46
6 0 40 -0 50
8 0 27 -0 48
10 -0 19 0 22
:ar regressions of chemical properties of 60
i the wet season with net soil inorganic N
to the soil (n=8)
Polyphenol Tannin C/N ratio
% %
0 14 0 39 -0 83 *
•0 03 051 -0 84 **
0 35 0 06 -0 72*
018 0 71 * -0 93 ***
0 10 0 62 -0 86 **
0 09 0 52 -0 84 **
0 06 0 27 -0 62
0 32 0 70 0 82*
0 20 0 56 -0 85 **
0 34 0 56 -0 77 *
031 071 -0 92 **
0 46 0 75 * -0 75 *
0 06 0 12 -0 18
061 0 78 * -0 51
0 63 0 60 -0 10
0 45 0 54 -0 30
0 38 0 53 0 15
001 -0 15 0 19
*, ** ***, significant at the 0 05, 0 01, and 0 001 levels, respectively
***P<0.001.
**P<0.01;
P<0.05;
*
ro
0.30
0.54
PP
T
4.0
54'
7-24
-17.5
0.63
N18.8*
-46.5
0.66
C/N
**
-2.9
56.8
10
**
0.90
N***
15.2
-39.6
*0.72
N**
13.7
2-40
8
0.58
N20.4*
-60.2
**
0.89
N***
o19
8-58
6
0.68
N**
25.1
-73.7
0.43
C/N
*-3.1
453
4
*0.71
N**
19.5
-60.0
0.65
C/N
**
-4,4
867
2
0.63
C/N
-4.0*
44.6
1
abxR2
xR2
ab
bxR2
a
MMSU
BRCI
AVRDC
(weeks)
Soils
Time
(PP).
polyphenols
or
(T);
%tannin
or
(N);
%nitrogen
or
C/N;
(x=
bx
+a
=N
Net
-immobihzed
released/
Nnet
influe
ncin
gproperties
chemical
init
ial
legume
of
determmation
of
coefficients
and
regression
hnear
multiple
stepwise
of
coefficients
Best
7.
Table
21
DISCUSSION
Biomass and N Accumulation
Biomass and N accumulation of soybean compared favorably with estimates of
different symbiotic systems ranging between 4 - 241 biomass ha l and 40 - 240 kg N ha
1 for food legumes (Singh et al, 1992) Biomass and nitrogen accumulation of
indigofera and soybean in our study were comparable with those obtained for the same
species grown for 60 d in Thailand (Meelu and Morns, 1988) Relatively low biomass
yields were obtamed with alfalfa and desmodium because of slow establishment and poor
early growth Based on biomass and N accumulation, alfalfa and desmodium were not
well adapted (at least compared to soybean and indigofera) and therefore the discussion
emphasizes the latter two species
If legumes are to be incorporated at 60 d, doubling the seeding density increases
biomass over normal seeding density by 15 and 28% for soybean, and 18 and 70% for
indigofera, in wet and dry season respectively The benefit of higher biomass due to the
higher seeding density for soybean was reduced with plant age as there was only a small
increase (3 -7%) at the higher seeding density by 90 d in either season This agrees with
the increase in soybean biomass of about 8% when grown to maturity at double density
(60 plants m 2) compared to normal density under temperate conditions (Kahnt et al,
1986)
Increasing temperatures and longer photo period in the wet season increased
soybean growth, whereas indigofera appears to be better suited for the dry season Even
though indigofera is better adapted to the diy season, N yields are inferior to those
obtained with soybean Soybean biomass declined after 75 d in the wet season due to
plant senescence and leaf fall
Chemical Properties of Plant Tissues
After maximum biomass was reached, N concentrations decreased in most
legumes, whereas hgnin and consequently C/N increased Similar trends were found in
Japan for milky vetch which nearly doubled its biomass from bloom commencement to
pod formation stage (Ishikawa, 1988) In the same study, protein concentration
decreased with legume crop growth after bloom commencement while soluble
carbohydrates, cellulose, and C/N increased, resulting in less decomposable material In
contrast an increase in N and a decrease in C/N in 49 d compared with 35 d cowpea was
reported by Franzluebbers et al (1994), as N accumulation of field grown cowpeas is
greater during the pod-fill stage compared with the vegetative stage Different component
parts of a green manure vary in N and hgnin concentration as well as C/N Foliage
22
contains higher N and lower lignin concentration and C/N than do stem and roots
(Watanabe, 1984, Frankenberger and Abdelmagid, 1985, Palm et al, 1988)
Net N Mineralization or Immobilization
Nitrogen release in this study conforms to earlier reports that N mineralization is
stiongly influenced by the properties of crop residues, soil properties, and the
mineralization time (Smith and Sharpley, 1990, Becker et al, 1994) The extra growth
time of 60 vs 90 d material changed the N mineralization pattern of harvested materials
diastically from significant N release to negligible mineralization or net N immobilization
during 10 wk The critical harvest time of legumes for GM use is gauged by weighing the
advantages of a higher biomass and N accumulation, and the disadvantages of a reduced
N mineralization rate, with increasing plant age Higher N mineralization rates from only
2 wk younger cowpeas (John et al, 1989), crimson clover and hairy vetch (Wagger,
1989) were attributed to lower C/N (John et al, 1989), cellulose, hemicellulose and
lignin concentrations (Wagger, 1989) and reduced N concentrations These reports
strengthen the importance of legume age as a key factor controlling N mineralization in
soil The older matenal is likely to have a smaller contnbution of foliage than stem and
roots to total biomass which is likely to affect the decomposition pattern
The initial lag penod of about 2 wk before active N mineralization started for 60 d
legumes agrees with findings of Oglesby and Fownes (1992) and Smith and Sharpley
(1990) The organic N addition is considered to be directly assimilated into the microbial
biomass during the lag period and mineialized dunng the subsequent mineralization phase
(Aoyama and Nozawa, 1993) A field study at MMSU showed that 30% of the initial
soybean-15N was recovered in the soil 16 wk after green manure application (Thonmssen
et al, 1996, unpublished) The companson of N released after legume addition in
incubation and field studies may lead to the conclusion that a considerable part of legume-
N was immobilized in MMSU soil in the first months after residue addition
In two of the three soils tested net N mineralization/ immobilization patterns
descnbed were strongly correlated with initial N concentration and C/N of the plant
matenal, in agreement with Frankenberger and Abdelmagid (1985), Tian et al (1992)
and Constantinides and Fownes (1994) In contrast to these reports, the influence of
plant properties on N release in this study depended on sampling time and soil type This
suggests that incubation time should be considered when evaluating the effects of tissue
properties on N mineralization
Greater N mineralization in AVRDC and BRCI soils of 60 d indigofera vs 60 and
90 d soybean could be explained by a higher initial N concentration and a lower C/N in
indigofera Initial tannin and polyphenol concentrations were the only two factors
influencing N release in MMSU soil Relatively low coefficients of determination of the
23
stepwise regression however may indicate that factors other than initial chemical plant
composition were driving net N mineralization/ immobilization behavior in this soil
Thus, the importance of lignin and polyphenols in retarding N release, cited by vanous
authors (Fox et al, 1990, Palm and Sanchez, 1990, Olegsby and Fownes, 1992), could
only be partly confirmed The role of lignin in controlling N release seems to be moie
important in lowland (flooded) than upland soils (Becker et al, 1994) Clement et al
(1995) reported the inhibitory effect of high concentrations of tannins (2 1 %) with
addition of Cassia velosa on N mineralization under waterlogged conditions Relatively
high concentrations of tannin (1 6%) of 60 d indigofera in our experiment did not retard
N release
Relationships between soil characteristics and mineralization rates established in
other studies may explain different N mineralization patterns in our three soils These
relationships were however not specifically tested in our study Soil pH seems to play an
important role in organic matter turnover and N mineralization The decay of plant
residues and soil organic matter on acid soils was accelerated by liming (Singh and
Beauchamp, 1986) Salinity and alkalinity of soils depressed the mineralization of N
during decomposition of Sesbania aculeata and Mehlotus alba and N-mmeralization of
green manure was enhanced after addition of P to saline-alkali soils (Singh and Rai,
1975) Organic residues decompose more slowly in soils with higher clay content,
especially with clays having higher exchange capacities (Lynch and Cotnoir, 1956,
Sorensen, 1975) Comparing properties of AVRDC, BRCI, and MMSU soils the
following hypotheses could explain differences in N release patterns The neutral pH and
the low clay content ot AVRDC soil may have contributed to a relatively fast N
mineralization of incorporated residues compared to MMSU soil BRCI soil was limed at
a rate of 5 t ha-1 previous to incubation, raising its pH from 4 to 6 1 The higher pH may
have favored a more rapid mineralization of the organic N pool in this soil as well as from
added plant material The high clay content of BRCI soil may not have strongly reduced
N mineralization as the exchange capacity was relatively low compared to MMSU soil
Greater N release in clay compared to sandy soil under flooded conditions was attributed
to lower organic matter content and possibly to lower available soil P in the sandy soil
(Becker et al, 1994) Alkalinity, low P content and the high clay content combined with
a high exchange capacity could have strongly contributed to a slow N mineralization m
MMSU soil As nitrification is inhibited at high pH, NH3 volatilization losses (Janzen
and McGinn, 1991) may have occurred, although they should have been reduced to a
minimum since plant residues were well mixed in the soil Considering the high clay
content in MMSU soil some potential for occurrence of denitnfication in microzones with
high water saturation cannot be excluded
24
CONCLUSIONS
Highest biomass and N yields were consistently reached with soybean at 75 d and
indigofera at 90 d Our results show that legume species, seeding density, and age (60,
75, and 90 d) are key factors affecting biomass and N accumulation and chemical
composition in both wet and dry seasons
Nitrogen released from legumes accumulated mostly as NO3-N in the soil The
extra growth time of 60 vs 90 d plant material changed the N mineralization pattern
diastically from significant N release to negligible mineralization or net N immobilization
Initial N concentration and C/N were the two major factors determining net N
mineralization in AVRDC and BRCI soils
There may be a potential of N fertilizer substitution for subsequent crops in
AVRDC and BRCI soils if the time course of N release due to the addition of 60 d
legume GM matches the time course of plant N uptake of the subsequent crops It will be
essential to test this hypothesis in the field to synchronize the release of GM-N with N
uptake by specific crops Nitrogen transformation processes following green manuring in
MMSU soil may be too slow to substitute N fertilizer to a subsequent crop, but may
contribute to the soil 01game matter build-up
REFERENCES
Alexander, D 1977 Mineralization and immobilization of nitrogen In Alexander, D
(ed) Introduction to soil microbiology pp 225-250 Wiley, New York
Aoyama, M ,and T Nozawa 1993 Miciobial biomass nitrogen and mineralization -
immobilization processes of nitrogen in soils incubated with various organic
matenals Soil Sci Plant Nutr 39 23-32
Becker, M ,J K Ladha, IC Simpson, and J C G Ottow 1994 Parameters affecting
residue mineralization m flooded soils Soil Sci Soc Am J 58 1666-1671
Broadhurst, R B,and W T Jones 1978 Analysis of condensed tannins using acidified
vanillin J Sci Food Agnc 29 788-794
Burns, RG 1978 Enzyme activity in soil Some theoretical and practical consideration
In Burns, R G (ed ) Soil Enzymes pp 295-340 Academic Press, New York
Clement, A, J K Ladha, andFP Chalifour 1995 Effects of chemical composition of
residues on N mineiahzation, microbial biomass and nee yield in submerged soil
Soil Sci Soc in press
25
Constantinides, M ,and J H Fownes 1994 Nitrogen mineralization from leaves and
litter of tropical plants relationship to nitrogen, hgnin and soluble polyphenol
concentrations Soil Biol Biochem 26 49-55
Fox, R H,R J K Myers, and I Vallis 1990 The nitrogen mineralization rate of
legume residues m soil as influenced by their polyphenol, hgnin and nitrogen
content Plant Soil 129 251-259
Frankenberger, W T,and H M Abdelmagid 1985 Kinetic parameters of nitrogen
mineralization rates of leguminous crops incorporated into soil Plant Soil 87 257-
271
Franzluebbers, K ,R W Weaver, and A S R Juo 1994 Mineralization of labeled N
from cowpea [Vigna unguiculata L ] plant parts at two growth stages in sandy soil
Plant Soil 160 259-266
Goenng, H K,and V Soest 1970 Forage fiber analysis (apparatus, reagents,
procedures and some applications) USDA Agric Handbook 379 U S Gov
Print Office, Washington, D C
Ishikawa, M 1988 Green manure in rice the Japan experience p 45-61 In Gieen
manure in rice farming International Rice Research Institute, Los Banos,
Philippines
Janzen, H H ,and S M McGinn 1991 Volatile loss of nitrogen during decomposition
of legume green manure Soil Biol Biochem 23 291-297
John, P S,R K Pandey, R J Buresh, and R Prasad 1989 Lowland rice response to
urea following three cowpea cropping systems Agron J 81 853 857
Kahnt, G ,L A Hijazi, and M Rao 1986 Effect of field bean and soybean cultivation
on soil compaction amelioration and its influence on wheat and barley as
subsequent crops J Agron & Crop Sci 156 57-66
Keeney, D R,and J M Bremner 1966 Comparison and evaluation of laboratory
methods of obtaining an index of soil nitrogen availability Agron J 58 498-503
Lynch, D L, and L J Cotnoir Jr 1956 The influence of clay minerals on the
breakdown of certain organic substrates Soil Sci Soc Am Proc 20 367-370
Meelu, O P,and R A Morns 1988 Green manure management in rice-based cropping
systems p 209-222 In Green manure in rice farming International Rice
Research Institute, Los Banos, Philippines
Oglesby, K A ,and J H Fownes 1992 Effects of chemical composition on nitiogen
mineralization from green manures of seven tropical leguminous trees Plant Soil
142 127-132
Palm, O,W L Weerakoon, M A P De Silva, and R Thomas 1988 Nitrogen
mineralization of Sesbama sesban used as green manure for lowland rice in Sn
Lanka Plant Soil 108 201-209
26
Palm, C A ,and P A Sanchez 1990 Decomposition and nutrient release patterns of the
leaves of three tropical legumes Biotropica 22(4) 330-338
Palm, C A ,and P A Sanchez 1991 Nitrogen release from the leaves of some tropical
legumes as affected by their hgnin and polyphenohc contents Soil Biol Biochem
23 83-88
Rowell, J G,and D E Walters 1976 Analyzing data with repeated observation on each
experimental unit J Agnc Sci 87 423-432
SAS Institute Inc 1989 JMP user's guide Version 2 SAS Institute Inc, Cary, N C
Siegel, R S 1980 Determination of nitiate and exchangeable ammonium m soil extiacts
by an ammonia electrode Soil Sci Soc Am J 44 943-947
Singleton, VL, and J A Rossi Jr 1965 Colonmetry of total phenolics with
phosphomolybdic-phosphotungstic acid reagents Am J Enol Vitic 16 144-158
Singh, S,and R N Rai 1975 Effect of salinity, alkalinity, phosphate and age of plants
on mineralization ot N from Sesbama aculeata and Melwtus alba after then
incorporation in soil J Indian Soc Soil Sci 23 122
Singh Y ,and E G Beauchamp 1986 Nitrogen mineralization and nitnfier activity m
limed and urea treated soils Commun Soil Sci Plant Analysis 7 1369-1381
Singh Y ,B Singh, and C S Khind 1992 Nutrient transformations in soils amended
with green manures Adv Agron 20 237-309
Smith, S J,and A N Sharpley 1990 Soil nitrogen mineralization in the presence of
surface and incorporated crop lesidues Agron J 82 112 116
Sorensen, LH 1975 The influence of clay on the rate of decay of amino acid
metabolites synthesized in soil during the decomposition of cellulose Soil Biol
Biochem 7 171-177
Tian, G ,B T Kang, and L Brussaard 1992 Biological effects of plant residues with
contrasting chemical composition under humid tropical conditions - decomposition
and nutrient-release Soil Biol Biochem 24 1051-1060
Wagger, M G 1989 Time of desiccation ettects on plant composition and subsequent
nitrogen release from several winter annual cover crops Agron J 81 236-241
Watanabe, I 1984 Use ot green manures in Northeast Asia p 229-234 In Organic
matter and nee International Rice Research Institute, Los Banos, Philippines
27
3
Legume Decomposition and N Release when
Applied as Green Manure to Tropical Vegetable
Systems. I. In the Wet Season in Taiwan
ABSTRACT
There is increasing concern and interest in alternative fertilizing methods but studies
are few and have not involved the use of green manures for horticultural crops in diverse
farming systems in the tropics This study is the first of three monitoring decomposition
and N-release dynamics of green manures, and testing the feasibility of meeting N needs
of tomatoes under contrasting environments in Taiwan and the Philippines Biomass and
N-accumulation of two legumes species (Glycine max L Merr and Indigofera ttnctona
L) grown for 60 days after sowing (d) were investigated in a field experiment at AVRDC
in Taiwan on two different bed systems (i) high beds (45 cm high, 2 m wide with 2 m
furrows between the beds sown with nee and permanently flooded), (n) low beds (20 cm
high, 2 m wide with 50 cm wide irrigation furrows between beds) during the hot and
humid season (June to August) The legumes were either used as mulch or incorporated
into the soil Legume decomposition was investigated in a litter bag study
Five t ha-1 of legume dry matter were accumulated with soybean, 2 t ha-1 with
indigo, containing 183 kg N ha * and 58 kg N ha-', respectively Legume biomass
breakdown was significantly faster when incorporated compared to mulched Nitrogen
release following soybean incorporation was rapid, reaching a maximum of 75 kg NO3-
N ha-1 content in the soil 3 weeks after incorporation When mulched, N release was
more gradual, averaging 45 kg NO3-N ha-1 over a period of 10 wk Following indigo
incorporation N was released very quickly but in smaller amounts than for soybean A
chlonde leaching experiment showed that up to 50% of the released NO3-N might be
leached out from the rooting layer in green manure and fertilizer treatments within the first
month after application Tomato (Lycopersicum escukntum Mill) yields with green
manure were comparable with fertilizer treatments Green manunng enhanced the short
term N availability of the soil Subsequent to tomato, significantly more N was taken up
by maize (30 d) grown on green manured plots compared with N fertilizer treatments
28
Legume green manures applied in the hot and humid season in Taiwan have considerable
potential to substitute for N fertilizer requirements for tomato fully or partly
INTRODUCTION
Vegetable production systems in the tropics and elsewhere are very intensive,
because vegetables are high value cash crops Commonly high fertilizer rates are applied
to maximize yields There is an urgent need for the implementation of alternative methods
to reduce excessive use of mineral fertilizers and to improve soil fertility and vegetable
quality
The old practices of green manunng, applying of compost, crop rotations, and inter- and
relay-cropping, which were used in various soil fertility programs for developing
countnes up to the early 1960's, have declined in extent as the use ot mineral fertilizer
progressively increased (Singh, 1975) A major benefit of legume green manures is the
contnbution of N to the soil via N fixation by legumes This benefit includes both the
short term enhancement of N fertility, the increase of humic and fulvic fractions of the
soil organic matter, the increase ot soil microbial biomass (Azam et al, 1985), and the
maintenance of the soil organic matter content which can affect soil structure, buffering
capacity, cation exchange capacity, water holding capacity, infiltration, miciobial
diversity and soil porosity (Frankenberger and Abdelmagid, 1985) Because the release
ot N from organic sources, such as green manures, is so closely tied to complex
microbial cycling of C and N, the availability and effects of legume-N are more difficult
to predict than tor chemical feitilizer N (Groffman et al, 1987) Most recent research on
green manures has focused on staple crop production systems, especially with rice
(Ladha et al, 1989) No published data could be found for tiopical vegetable production
systems, although green manuring is still a common practice in some vegetable farms in
India and Nepal (pers communication ot Dr Babha Tnpathi)
This study was the first of a series investigating the fate of N in green manure
applied to horticultural crops Two studies were performed in Taiwan the first wet season
(WS) and the second in the dry season Two further field expenments were performed in
the two major tomato growing aieas in the Philippines The main purpose of this study
was to test the feasibility of meeting nitrogen needs of the tomato crop in rotation with
leguminous green manure crops in field experiments by quantifying legume biomass and
nitrogen accumulaUon as well as legume decomposition and N release in the soil
29
MATERIALS AND METHODS
Field trial 1993
This study was conducted during summer 1993 (April - September) on the
experimental farm of the Asian Vegetable Research and Development Center (AVRDC) in
Taiwan The field used had vegetables grown in rotation with flooded nee for the last
twenty years The soil is of Take series (loamy, mixed, hyperthermic, Fluvaquentic
Entochrept (Soil Survey Staff, 1992), pH (H2O) 8 2, total Kjeldahl N 0 7 g kg •
(Bremner, 1965), total C 6 4 g kg * (Walkley-Black Method) previously cropped with
corn for 1 month to obtain a homogeneous soil mineral N distribution Corn stubble was
removed before the trial started
Experimental design
Experiments were conducted using low or raised bed systems The raised beds
were 45 cm high, 2 m wide with 2 m furrows between the beds The furrows were
transplanted with nee (Oriza sativa) and permanently flooded The low bed was 20 cm
high and 2 m wide and with 50 cm wide irrigation furrows between beds Both
expenments were adjacent such that the soil type, cropping history and meteorological
conditions were the same
The raised bed field plot was 27m by 54m and the low bed plot was 27m by 26m
Experimental design for each bed system was a randomized complete block Treatment
plots were 2m by 6m with four replicates
Table 1. Treatments in raised and low beds (AVRDC Taiwan, 1993)
treatments (abbrev) legume green manuring/
fertilization
vegetable
1 (Ssi) soybean incorporation tomato
2 (Ssm) soybean mulch tomato
3 (Isi) indigo incorporation tomato
4 (Ism) indigo mulch t tomato
5 (Ck 0 = control) weedfree fallow OkgN/ha tomato
6 (Ck30) weedfree fallow 30 kg N /ha tomato
7 (Ck 60) weedfree fallow 60 kg N/ha tomato
8 (Ck 120) weedfree fallow 120 kg N/ha tomato
t Indigofera (living mulch) regrew after the first cut for mulch (62 d) and was trimmed back another 3
times 9 July (84 d) 26 July (101 d) and 5 August (111 d) The cuttings were added as surface mulch to
the tomatoes
30
The 8 treatments were as follows (Table 1) two legume species, two green manuie
systems (mulch and incorporation), and four controls having weedfree fallow while
legumes were grown m the legume treatments with 0, 30, 60 or 120 kg N ha"1 applied to
the tomato crop
Green manure and tomato crop
To obtain high legume biomass in a short time, the legumes were sown in four
double rows at double the normal late for the legumes, as suggested by Yamoah and
Mayfield (1990) and indicated by data in Chapter 2 One double row was sown on each
edge of the bed and the other 65 cm from the edge row The rows in each double row
were 10 cm apart
The expenment commenced on 12 April 1993 (Fig 1) and the legumes were hand-
sown at 200 seeds nr2 tor soybean (Glycine max, 2-3 seeds/10 cm row) and 1 32 g m"2
lor indigo (Indigofera tinctona) Seeds were inoculated with a rhizobium strain mixture
that was specific for each legume species, provided by the Soil Science Department of the
Chung Hsing University in Taichung, Taiwan
/ Legumes 11 Tomato ll Maize /
April May June July August September
Fig. 1. Cropping pattein including legumes, tomato and maize
Phosphorus at 35 kg P ha_1 as super phosphate and potassium at 83 kg K ha _1as
potassium chloride (KC1) was broadcast in all beds Legumes were sampled for biomass
accumulation on 14 May (33 d) On 15 June (68 d) all the legumes were cut at soil level,
chopped into 10 cm pieces and either incorporated by rototilhng to 15 cm depth, or left as
mulch on the soil surface as required for treatment On 18 June, 30-day-old tomato
(Lycopersicon esculentum Mill, determinate bushy type, shoit duration, AVRDC line
5915-93-1-0-3) seedlings were transplanted in two rows per bed spaced 40 cm within
and 100 cm between rows Super phosphate (35 kg P ha 1) and K as KC1 (50 kg K ha-')
were applied at transplanting A further 50 kg K ha-1 was applied on 2 July, and on 23
July For the N fertilizei treatments 30 kg N ha-1 (as ammonium sulphate) was applied at
transplanting to Ck30, Ck60 and Ckl20 The first side dressing of 30 kg N ha"1 was
applied to the crop on 2 July in treatments Ck60 and Ckl20 A further 60 kg N ha-1 was
applied to Ckl20 as second side dressing on 23 July Red tomatoes were harvested on 17
31
and 24 August, with a final harvest on 1 September harvesting red and green tomatoes
After harvest, maize was sown on 3 September in 6 rows per bed (30 seeds m 2) and
harvested 30 days later on 4 5 October
Environmental monitoring
Weather data were collected throughout the trial period at the AVRDC weather
station To follow soil moisture as affected by green manure treatments, tensiometers
were placed in treatments Ssi, Ssm, Isi, Ism and CkO at 15, 30 and 45 cm depth in the
raised beds, and at 15 and 30 cm depth in the low beds, at tomato transplanting
Plant analysis
Legumes were sampled at 33 and 68 d The plants from 0 5 m2 area of each of the
four replicates, which was afterwards excluded from further sampling, were carefully
dug out to a depth of 15-20 cm and the soil then separated from the roots Root nodules
per plant were counted and samples of the nodules were cut open to assess their
effectiveness by the presence or absence of the pink colour produced by hemoglobin
(Vincent, 1970) Shoots, roots, and nodules were dried at 60°C for 48 hours and
weighed Nitrogen content in shoots and roots including nodules were determined by the
Kjeldahl distillation method (Bremner, 1965) for legumes 68 d only
At harvest, marketable fruit fresh weight, fresh and dry weights and nitrogen
content of tomato plants were determined Fruit nitrogen content was not determined but
calculated based on the assumption of 2 5% N in tomato fruit dry weight, and a 0 05 dry/
fresh weight ratio (Thoennissen, 1994, unpublished) The apparent N recovery (%) of
the applied N via fertilizer or legume N m tomato plants and fruit was calculated as
follows
Apparent N recovery (%) = CN in tomato with N application) -(N in tomato control)
N applied
Thirty days after sowing maize, plants (including roots) were pulled out from the
soil and biomass and total N were determined as a rough measure of the inorganic N
available in the soil after tomato harvest
The N-balance at the end of the experiment was calculated following the methods ot
Myres and Wood (1987)
Decomposition experiment
Nylon bags (mesh size 1 mm) containing 15 g fresh plant material (4 7-5 5 g dry
weight) were used to determine biomass breakdown ot 68-day-old incorporated or
mulched soybean and indigofera Mulch treatments contained shoot material only On 15
June all bags were either buned at 10 cm soil depth for incoiporation treatments or left at
32
the surface as mulch treatment Decomposition bags were sampled at the same dates as
the soil sampling for inorganic N, namely 0, 2, 5, 8, 14, 29, 42, 62, 75 days after
incorporation (DAI) Two randomly chosen bags per treatment weie retrieved, oven
dried at 60CC for 48 hours and weighed Samples were ashed by dry combustion in a
muffle furnace (500°C) for 8 hours to determine original ash-free dry weight remaining
(Aber et al, 1990)
Decomposition data analysis
Decomposition rates of two species can be compared in one site by fitting them to a
mathematical model to estimate constants describing the loss ot mass over time Two
decomposition models were compared The equation for the single exponential decay
function (Jenny et al, 1949, Olson, 1963) is
Nt = N0(l-e-kt),where Nt is the biomass lemaining No is the original biomass, k is the relative
decomposition rate ot each green manure treatment, t is the time in days The relative
decomposition rate k characterizes the loss of mass over time The assumption underlying
the single exponential model can be expressed in two ways, either the absolute
decomposition rate decreases linearly as the amount of substrate remaining declines, or
the relative decomposition rate remains constant (Wieder and Lang, 1982) For statistical
analyses the single exponential model was linearized (log transformed) Statistical
comparisons of slopes, intercepts and residual vanances among series of individual
regressions were made using analysis of covariance technique (Snedecor and Cochian,
1978)
The double exponential model (Hunt, 1977) assumes that litter can be partitioned
into two components, a relatively easily decomposed or labile traction (A), and a more
lecalcitiant ft action (1-A)
N, = Ae-kt+(l-A)e-ht,where k is the rate constant for the labile component, and h is the rate constant for the
tesistant component The A-value tor each legume species had to be estimated by the
model
Inorganic N
The effect of the legume species and the green manure application method on the
release of inorganic N in the soil was monitored in treatments CkO, Ssi, Ssm, Isi, Ism,
which weie sampled 0, 2, 5, 8, 14, 21, 29, 42, 62, 75 DAI Soil samples were collected
with a 5-cm-diameter auger from 5 treatments in blocks I, II, and III At each sampling
date three soil samples at 0 - 30 cm depth were taken from each treatment Each sample
was a mixed composite collected from 4 locations in each plot Soil samples were passed
33
through a 10 mm sieve, extracted with 1 N KCl (1 1 5 soil/water) and inorganic nitrogen
(ammonium and nitrate) was determined with an ammonia gas sensing electrode (Siegel,
1980)
Chloride analysis
To estimate whether the potential leaching of nitrate is affected by green manure or
nitrogen fertilizer application 50 g sodium chloride was broadcast on 1 m2 in the plots
(Cameron and Wild, 1982) grown with tomato following Ssi, Ssm, CkO and Ckl20 in
four replications in both bed systems Sodium chlonde was applied on the respective
plots after soybean incorporation and mulch on 23 June 1993 Soil samples in 10 cm soil
layer subsamples were taken 21 June, 23 July and 30 August, from 0-50 cm in the laised
beds and from 0-30 cm in the low beds Soil samples were air dried and extracted (1 2
soil/water) Chlonde in the water extracts was determined with a chlonde analyzer
(Chlonde Analyzer 926, Coramed AG, Dietlikon, Switzerland)
Statistical analysis
Data were analyzed by ANOVA procedure using JMP Version 2 (SAS Institute,
Inc 1989) and SAS version 6 03 (SAS Institute, Inc 1991)
RESULTS
Environmental monitoring
Rainfall patterns dunng the expenmental penod Apnl - October 1993 are shown in
Figs 2 and 3 Mean air temperature was 27 6°C, soil temperature, 28 3°C, and total
rainfall frorn April to September was 1349 mm
Soil matnc potential did not differ greatly between the control, and the legume
species and the incorporation and mulch treatments at the various soil depths in both bed
systems Therefore means of five treatments are shown in Fig 4 for the raised beds only
Soil matnc potential at 15 and 30 cm depth decreased slowly after tomato transplanting,
fluctuating between -0 01 to -0 05 MPa for the first 6 wk The rainfall event 7 wk after
tomato transplanting increased soil moisture levels to -0 01 MPa, aftei they had dropped
between -0 06 to -0 08 MPa shortly before in both bed systems
34
200
PrecipitationEvaporation
20
15 Sc
_o
10?o
&5 >
W
0
120 150 180
Days after beginning the experiment
Fig. 2. Precipitation and evaporation in mm, 12 Apnl to 5 October 1993. Days of
sowing legumes, transplanting tomatoes and sowing maize are shown with arrows 1, 2,
3, respectively (AVRDC, Taiwan, 1993)
35
U
4>
3 30
t-i
Oh
2 25
20
iti
t f
Air temperatureSoil temperature
i i i
35
Uo
30 53
0)
6o
25^o00
30 60 90 120 150
20
180
Days after beginning the experiment
Fig. 3. Mean air temperature (°C) and mean soil temperature (°C) at 10 cm soil depth, 12
April to 5 October 1993 Days of sowing legumes, transplanting tomatoes and sowing
maize are shown with arrows 1, 2, 3, respectively (AVRDC, Taiwan, 1993).
35
•g o.ooon
-0.02
-0.04
-0.06
-0.08
-0.10
0 7 14 21 28 35 42 49 56 63 70 77
Days after green manure application
Fig. 4. Soil matric potential (MPa) in tomato beds after green manure application on
raised beds at 15, 30, 45 cm soil depth. Values shown are means of five treatments:
control, indigo and soybean, mulch and incorporation. Standard deviation (SD) of
treatment means are shown (AVRDC, Taiwan, 1993).
- 45 cm depth
36
Legume biomass and N accumulation
The growth pattern of the two legume species differed greatly Soybean 33 d
accumulated 1 8 t ha • dry matter in the low beds, and 2 3 t ha 1in the raised beds
compared to indigofera with 0 11 ha 1in low beds and 0 3 t ha * raised beds Rhizobium
inoculation was successful as 70- 90% of the sampled nodules were active on soybean
and indigofera at 68 d Shoot and root biomass of soybean was significantly greater than
that ot indigofera at the final harvest (Table 2) Indigofera shoot weight in the low bed
system was reduced to neaily one quartei of the biomass obtained in the raised bed
Table 2 Dry matter yield, nitrogen content and nitrogen accumulation of soybean and
Indigofera tinctona L,68 d Values shown are means of four replicates (AVRDC,
Taiwan, 1993)
Soybean Indigo LSD (P<0 05)
bed system
raised low raised low raised low
Dry matter shoot 5159 5465 1683 420 462 322
(kg/ha) root 773 346 279 143 254 124
N content shoot 3 25 3 26 3 16 2 93 ns ns
(%) loot 1 95 2 38 171 1 68 ns 0 36
N shoot 167 2 177 2 52 9 12 3 15 4 24 0
(kg/ha) root 15 1 8 1 47 24 25 29
total 182 3 185 3 57 6 14 7 14 5 23 5
Soybean shoot weight was not affected by the bed system, but root biomass was
reduced by one half in the low bed system Soybean accumulated a total of 182 -185 kg
N ha ! of planted area in either bed system, while Indigo accumulated 58 kg N ha 1in the
raised beds and 15 kg N ha 1in the low beds
Decomposition
Biomass loss was approximately 20% greater in incorporation treatments than
mulched An exponential loss occurred for all treatments during the first 30 days, when
30 80 % of the dry weight was decomposed, after which the rate of subsequent weight
loss declined Similar patterns occurred in both bed systems (Fig 5)
37
Soybean incorporation
Soybean mulch
Indigo incorporation
Indigo mulch
-« = = ;!
7 14 21 28 35 42 49 56 63 70 77
Days after green manure application
Fig. 5. Decomposition of soybean and indigofera residues when used as mulch oi
incorporated into the soil in the low and raised bed system dunng the hot and humid
season at AVRDC, Taiwan, 1993
Biomass breakdown data oi soybean and indigofera were fitted to the single and
double exponential models described for litter decomposition by Wieder and Lang
(1982) The double exponential model fitted data very well However the model was
rejected for the estimation of the labile fraction (A-value) since one legume species
(soybean) when mulched gave highly different values than when incorporated The single
exponential model fitted the data well (Table 3) The higher the k-value, the faster the
organic matter decomposed Significantly higher k-values weie found for soybean and
indigofera incorporation compared to mulch Similar decomposition rates were obtained
for soybean and indigofera incorporation, whereas soybean mulch decomposed almost
38
twice as fast as indigofera mulch Bed system decomposition rates within the same green
manure treatment were not statistically different
Table 3. Decomposition rates, k, for a period of 77 days, when soybean and Indigoferatinctona (60 d) used as green manure were incorporated or left as surface mulch Values
were calculated using the single exponential model for decomposition (Wieder & Lang,1982) (AVRDC, Taiwan, 1993)
Gieen manure treatments bed system kla) r2(b)
Soybean incorporation raised 0 0272 A 0 89 ***
low 0 0361 A 0 88 ***
Soybean mulch raised 0 0175 B 0 9) ***
low 0 0144 B q 90 ***
Indigo incorporation raised 00266 C 0 88 ***
low 0 0262 C 0 93 ***
Indigo mulch raised 0 0098 D 0 75 **
low 00111 D 0 9j ***
(a) K- values within the bed system were tested using a pairwise t-test comparison for slopes K values
with different letters are sig different (P<0 05)
(b) Regressions are significant at **< 0 01 ***<0 001
Inorganic N
Inorganic N was determined as a measure ot the availability of applied legume N to
tomato plants (Fig 6 and 7) Ammonium levels in both bed systems were as low as 5 kg
NHa N ha-1 throughout the tomato growing season with the exception of minor peaks of
10-15 kg NH4-N ha-1 7 and 21 days alter green manure application (GM application)
Soil nitrate contents in both bed systems were much highei than ammonium
contents Differences in NO3 contents between bed systems could mainly be attributed to
diffeient amounts of applied legume biomass, especially with indigofera NO3-N
availability was highest two to four wk after GM application In the raised beds
comparable amounts of NO3 were mineralized following indigofera or soybean up to two
wk after legume incorpoiation Latei NO3 mineralization rates increased furthei for
soybean compared to indigofera, where the NO3 content in the soil declined rapidly from
50 kg NO3-N ha-1 2 to 4 wk after incorporation to 20 kg NO3-N ha ! 6 wk after
incorpoiation
Most NO3 was released in the soybean incorporation treatment which differed
significantly from all other treatments Nitrate levels following indigofera incorporation
were lower than following soybean incorporation and differed from the control in the
39
raised bed system. Nitrate contents in indigofera mulch treatments were similar to the
control
control
- - • - - Indigofera incorporated
.
—°— Indigofera mulch- -A - •
Soybean incorporated
I
r 1 —A— Soybean mulch
/
./
/
\
*< AAJLSD 0.05
j
I 'ik/
\\A>
- '/y*k / \ T^
TnL \ yf ^ ^v. y^
Iff* ^^»-^ Na,'
? iii
A0 7 14 21 28 35 42 49 56 63 70 77
Days after mulch or incorporation
Fig. 6. NO3-N and NH4-N (kg ha _1) contents in soil (0-30 cm) after indigofera and
soybean green manure mulch or incorporation for tomato in raised beds Values shown
are means of three replicates Error bars indicate least significant difference (AVRDC,Taiwan, 1993)
40
7 14 21 28 35 42 49 56 63 70 77
Days after green manure application
Fig. 7. NO3-N and NH4-N (kg ha ') contents in soil (0-30 cm) after indigotera and
soybean green manure mulch or incorporation for tomato in low beds Values shown are
means of three replicates Erroi bars indicate least significant difference (AVRDC,Taiwan, 1993)
The release of inorganic N was very fast reaching its maximum release 2-4 wk after
application Nevertheless incorporating or mulching soybean maintained soil NC>3-levels
10 wk after GM APPLICATION, above that at tomato transplanting Soil nitrate in those
41
treatments was significantly higher than in the control and indigofera incorporation and
mulch treatments at final tomato harvest
Leaching experiment
Chloride leaching followed similar patterns m control, 120 kg N ha l, soybean
mulch and incorporation treatments Therefore Cl-loss (%) data of these four treatments
were averaged for each sampling dates and bed system (Table 4) From CI values 21
June it can be concluded that background concentration in 10 - 50 cm soil depth was
rather low, the same is likely for the first 10 cm Of the applied CI 42-50% had been lost
by 23 July 1993, one month after application The greatest loss occurred at 0-10 cm soil
depth, whereas Cl-accumulation occurred at soil depths of 10-20 cm and 20-30 cm
Chloride did not accumulate at depths >30 cm
Table 4. Percent remaining chloride at different soil depths are presented for three
sampling dates in raised and low beds Values shown are means of four treatments
control, Ck 120 kg N ha-1, soybean incorporation and mulch and standard deviation
between treatment means Treatments means were means of four replicates (AVRDC,Taiwan, 1993)
soil depth (cm)
0-10 10-20 20-30 30-40 40-50 Sum
Raised beds
% CI t
21 June 78 2±4 1 6 1 ±23 42±07 6 8+11 47±09 100
23 July 23 5 ± 5 8 118 + 08 6 2 ± 2 1 45± 12 40±09 50
30 August 142 ± 1 8 9 1 ± 19 54±14 33±09 29± 10 34 9
Low beds
21 June 860±3 2 76 + 26 64±08 100
23 July 283+100 187 ±43 106 ±22 57 6
30 August 23 3± 1 2 155 ±35 11 1 ±23 49 9
t % CI was calculated by setting the Cl-contents (g Cl/m') from 0- 50 cm (raised beds) and 0 30 cm (low
beds) to 100% on 21 June 1993
Tomato yield
Tomato fruit yields and plant dry matter are shown m Table 5 On raised beds
highest yields were obtained in the soybean mulch treatment The incorporation of
indigofera also increased tomato yield significantly This yield compared favourably with
that obtained using 30 or 60 kg N ha 1 fertilizer Diffeiences between yields of indigofera
42
mulch, soybean incorporation and 120 kg N ha-1 treatments and the control were not
significant
In the low beds tomato yields differed significantly from the control following 120
kg N ha-1 fertilizer application only Soybean mulch was the only legume green manure
treatment increasing the tomato yield over the control, but differences were not
significant In both bed systems tomato biomass followed the same trends as for tomato
yields (Table 5)
Table 5. Tomato fruit and biomass (yield per planted area in t ha*1) in raised and low
beds following legume green manure or fertilizer N treatments Values shown are means
of four replicates (AVRDC, Taiwan, 1993)
Tomato fruit yield(tha-1)
Tomato plant dry matter
(tha-1)raised beds low beds raised beds low beds
Ammonium sulfate
(kg N ha-1)0 (control) 2 52 2 85 1 11 0 88
30 5 83 3 68 1 56 0 81
60 5 92 3 65 2 11 0 86
120 4 50 451 1 80 101
Legumes
Soybean incorporation 4 68 241 1 67 0 64
Soybean mulch 6 31 4 34 2 05 0 88
Indigo incorporation 5 11 2 57 198 0 74
Indigo mulch 4 73 1 18 1 32 0 36
LSD (P<0 05) 2 38 1 65 0 74 0 41
The maximal N uptake at 48 kg ha-1 in the tomato plant and 7 9 kg N ha-1 in the
fruit was recorded in soybean mulch in the raised beds (Table 6) Increasing rates of
inorganic N fertilizer lesulted in an increase of N uptake In untreated control in the raised
beds uptake was 24 kg N ha 1 and in the low beds 20 kg N ha"1, these can be attributed to
the N mineralized from the soil organic N pool
Ot the legume tieatments, highest apparent N recovery (41%) in tomato was
obtained in the indigofera incorporation which compared favourably with 60 kg N ha-1
The apparent N recovery from soybean incorporation and indigo mulch were low,
although these treatments provided the highest and lowest soil N03-levels throughout the
WS This contradictory lesult may indicate that this method of estimating N recovery
43
does not relate to yield potential of tomatoes Recovery of N following soybean mulch
was a little higher than that after soybean incorporation Apparent N recovery for the low
beds was very low in all treatments (data not presented)
Table 6. Tomato plant and fruit nitrogen uptake (kg N ha ') and apparent N recovery(%) of the added N with legume or chemical N fertilizer in raised and low beds Valuesshown are means of four replicates (AVRDC, Taiwan, 1993)
N in tomato Apparent N
recovery
Raised beds Low beds Raised beds
plant fruitt plant fruitt
Ammonium sulfate
(kgNha-1)(kg hal) %
0 (control) 20 8 32 16 7 36
30 27 9 73 16 0 46 37 3
60 417 74 167 46 41 8
120 45 8 56 21 5 57 22 8
Legumes
Soybean incorporation 37 3 59 13 5 30 10 7
Soybean mulch 48 3 79 19 3 54 17 3
Indigofera incorporation 417 64 16 6 32 414
Indigofera mulch 26 7 59 7 1 1 5 15 1
LSD (P<0 05) 180 ns
t tomato fruit N uptake (kg/ha) was calculated assuming N content in tomato truit dry matter ot 2 5%
Dry matter was considered as 5% ot the fruit fresh matter
Maize biomass and N-uptake, N-balance
Maize harvested 30 d was used to estimate N availability and N treatment effects for
a subsequent crop following tomato (Table 7) Maize yields in all green manure
treatments in the raised beds, and soybean treatments in the low beds were significantly
greater than the control, and tended to be higher yielding than Ckl20
Nitrogen accumulation in maize in the raised beds was greater in all treatments (but
for Ck 30) than the CkO control whilst in the low beds maize N uptake only differed trom
CkO following Ism, Ssi, and Ssm Highest N uptake in maize occurred in soybean mulch
and incorporation treatments
44
Table 7. Residual effect of tomato N fertilization (0, 30,60,120 kg N ha"1 as
(NH4)2S04 or soybean or indigofera green manure mulched or incorporated) on the drymatter yields, the N content and the N uptake of maize 33 d in the raised and low bedsValues shown are means of four replicates (AVRDC, Taiwan, 1993)
Maize
dry matter
(tha-1)N% N
(kg ha- ')tR *L R L R L
Ammonium sulfate
(kg N ha-1)0 (control) 124 105 211 1 87 26 1 19 6
30 1 35 104 2 10 1 94 29 3 20 5
60 1 59 1 00 2 51 194 40 4 196
120 1 86 1 20 2 64 2 13 49 3 26 6
Green manure
Soybean incorpoiation 2 46 1 73 2 39 2 43 59 3 42 4
Soybean mulch 2 43 1 73 2 33 2 13 56 6 37 0
Indigo incorporation 2 03 1 20 2 00 1 93 40 6 23 3
Indigo mulch 2 13 1 31 2 23 2 04 47 8 27 1
LSD (P<0 05) 0 35 0 39 0 45 0 25 12 7 56
t R raised beds,
t L low beds
The N-balances presented in Table 8 show N inputs by fertilizer or green manure as
well as N outputs by tomato (plant and fruit) and maize plants The remaining N atter
subtracting the outputs from the inputs is a rough estimation of the N cycling m this
pai titular cropping system as N leaching as well as volatilization, demtnfication,
immobilization of N were not measured
N balances were positive in both bed systems in Ckl20 and soybean green manure
treatments, the same treatments in which maize N uptake differed significantly from CkO
With indigo green manure or 30 oi 60 kg N ha-1 fertilizer application soil N was depleted
by 20 - 30 kg N ha-1
45
Table 8. Nitrogen balance after N inputs of 0, 30, 60 and 120 kg N ha"1 (NH4)2S04 or
indigo or soybean green manure to tomato and N-outputs by tomato (plant and fruit) and
maize (30 d) on raised and low beds (AVRDC, Taiwan, 1993)
Raised beds Low beds
input output balance input output balanc
Tomato Maize Tomato Maize
N kg ha- l
Ammonium sulfate
(kgNha1)0 (control) 0 24 0 26 1 -50 1 0 20 3 19 6 -39 9
30 30 35 2 29 3 -34 5 30 20 6 20 5 11 1
60 60 49 1 40 4 -29 5 60 21 3 19 6 - 19 1
120 120 514 49 3 19 3 120 27 2 26 6 66 2
Legumes
Soybean 178 7 43 2 59 3 76 2 191 7 165 42 4 132 8
incorporationSoybean mulch 185 9 56 2 56 6 69 5 178 9 24 7 37 0 1172
Indigo 58 2 48 1 40 6 -30 5 12 5 19 8 23 3 -30 6
incorporation
Indigo mulch 56 9 32 6 47 8 -23 5 16 9 86 27 1 - 18 8
DISCUSSION
Legume biomass and N accumulation
Continuous plowing and puddling of the soil previously grown with flooded nee
resulted in a hard-pan at 50 - 60 cm soil depth in the raised beds and 20 - 35 cm in the
low beds Consequently the soil volume which can be penetrated by roots was
considerably greater m the raised compared to the low beds, reducing soybean root
biomass strongly in the low beds Nitrogen accumulation m 60 d soybean and indigofera
obtained in this experiment was greater than that recorded by Meelu and Moms (1988)
for soybean 60 d (138 kg N ha*1) but lower than recorded for indigofera (84 kg N ha-1)
Decomposition
The strong effect of incorporation of legume green manure compared to mulch on
the speed of decomposition has been shown for maize stover and Leucaena in field
experiments by Wilson et al (1986) Fast initial decomposition of soybean matches with
the findings of Broderand Wagner (1988) where incorporated soybean residue lost 68%
46
of its total organic matter over the course of 32 days Reinertsen et al (1984) associate
the more rapid decay immediately after the bunal of the residue with the decomposition of
water soluble organic constituents Hunt (1977) describes differences in decomposition
patterns and rates among substrates as a function of the amount of the labile or rapidly
decomposing fractions (sugars, starches, proteins) and the recalcitrant or slowly
decomposing fraction (cellulose, ligmn, fats, tannins, waxes) The influence of the
source quality, the micro climatic conditions as well as the soil depth on decomposition
rates as well as on decomposer biomass dynamics have been stressed by various authors
(Hunt, 1977, Swift et al, 1979)
Soil moisture and N release
Soil temperature (25 - 30 °C) and soil matnc potential (-0 01 to -0 05 MPa) were
optimal to ensure strong N mineralization in the soil, as outlined by Cassman and Munns
(1980)
The NO3 release followed the same initial exponential reaction as biomass loss,
showing well the causal linkage between these two processes Slow plant decomposition
is followed by lower amounts of NO3 released, and vice versa Lowest NO3 contents in
the soil with mdigofera mulch was likely due to the competition for NO3 by tomato and
the regrowth of mdigofera The greater N released after incorporation as compared to
mulch was attributed by Wilson et al (1986) to a more rapid decomposition rate by
incorporation, results which were confirmed in this experiment (Figs 6 and 7)
N release patterns described here correspond with those reported for incubation
studies at 25- 30°C (Weeraratna, 1979, Palm and Sanchez, 1991) The high soil
temperature and moisture conditions may have mainly contnbuted to the fast release of
NO3 after GM application Decline of soil inorganic N over time indicate the period of
greatest N uptake by the tomato plants or/and further losses due to leaching,
demtnfication or biological N immobilization (Mary and Recous, 1994) Numerous
authors, e g Janzen and McGinn (1991), have stressed the importance of volatilization
losses when legume green manure is applied as surface mulch since drying and
decomposing conditions are reported to enhance volatilization The volatile loss of labile
N from decomposing green manure mulch may appreciably diminish its fertility benefit,
whereas NH3 losses from incorporated green manure elsewhere have been reported to be
negligible (Janzen and McGinn, 1991) If, however, lower mineralization rates are
actually the cause of the reduced inorganic N accumulations under non tillage, then such a
system could be more conservative of organic N in the long term (Sarrantomo and Scott,
1988) Slightly increasing NO3 contents in the soil 10 wk after GM application (Fig 6
and 7) may indicate remineralization of N which was immobilized earlier (6 to 9 weeks
47
after GM apphcation), although this is descnbed as a relatively slow process in temperate
soils (Mary and Recous 1994)
Chloride leaching
The great loss of CI and thus nitrate within the first month of fertilizer application is
probably due to the two ramfall events soon thereafter (Fig 2) The soil >30 cm depth in
the raised beds was permanently submerged (due to the standmg water in the nee beds of
this system) so that CI may have been leached with the rice bed lrngation An improved
infiltration rate of the soil on the raised beds may also have increased the leaching
(Shennan, 1992) Shuford et al (1977) found increased leaching losses in non-tilled as
compared to tilled treatments due to increased soil moisture and preferential flow through
cracks and continuous macropores This tendency could not be confirmed in our
companson of legume mulch (non tilled) and incorporation (tilled) treatments
Tomato yield
Major production constraints for tomatoes grown durmg the rainy tropical summer
in the lowlands are high day and night temperatures affecting the fruit setting, tropical
storms flooding the fields temporarily, and fungal (black leaf mold caused by
Pseudocercosporafuhgena ) and viral (tobacco mosaic virus) diseases For these reasons
yields m all treatments in our expenment were variable and low especially in low beds
The economic implications of growing tomatoes under these adverse conditions are
particularly important since fresh tomatoes command high market prices
It is likely that under adverse growth conditions for tomato, soybean incorporation
resulting in strong N release may have lead to an over fertilization of tomato similar to the
120 kg N ha ' fertilizer treatment, as descnbed by Coltman (1989), and Lorenz and Bartz
(1968) Excess N-uptake has been shown to induce strong vegetative growth detrimental
to the fruit yield, when tomatoes were intercropped with cowpea (Olasantan, 1991)
Storage of N in vegetative parts of the plant may have induced a delay in fruit setting and
ripening The tomato plant might have been unable to take up the available N at a time
when large amounts of NO3 were present in the soil, leading to losses of NO3 through
leaching
The second side dressing applied in the 120 kg N ha 1 treatment may have helped
tomato plants to overcome the strong environmental stress in the low beds Lower tomato
yields in the low beds following indigo and soybean incorporation as compared to
soybean mulch could not be explained by soil mineral N conditions
Low NO3 contents m the soil in indigofera mulch treatments, and competition of
indigo living mulch with tomato for N and water, affected the tomato yield Short term
yield loss due to net immobilization of soil nutrients by some cover crops (e g vetch,
48
Yamoah and Mayfield, 1990) may also occur, and this can make legumes unacceptable to
farmers After the application of indigofera green manure, N was released very quickly
Part of the released N may have been taken up by the regrowing indigofera living mulch,
which was probably a competitor with tomato, and some NO3 may have been leached
Low fertilizer N lecovery and little response to applied N makes tomatoes relatively
inefficient users of fertilizer N (Hills et al, 1983) Substantial amounts of the applied
fertilizer N may not be absorbed as tomato roots do not appear to proliferate in soil with
higher mineral N content (Jackson and Bloom, 1990) Nevertheless legume-N was
sufficiently available to a succeeding tomato crop, and tomato yields following green
manuie were comparable to those in fertilizer treated plots (Stivers and Shennan, 1991)
Yields of field giown tomatoes responded strongest with the application of 50 - 60 kg N
ha"1 mineral N compared to control, whereas fuither increases ot N up to 280 kg N ha 1
did not increase tomato yields greatly over those obtained with 50 - 60 kg N ha"1
(Garnson et al, 1967, Stivers and Shennan, 1991) In our experiments high tomato
yields were obtained undei soybean mulch, which may have acted as slow release N
fertilizer The slower release ot N and the high yield response in soybean mulch
treatments strengthen conclusion of Wilson et al (1986) and, Hochmuth (1992) that slow
lelease-N sources have highest utility for longer-term vegetable crops such as tomatoes
Nitrogen recovery from incorporated indigo (58 kg N ha-1) was 41% which was as high
as 60 kg N ha ' N fertilizer N from indigo incorporation was released quickly and was
effectively absorbed by the subsequent tomato crop
Maize biomass and N-uptake, N-balance
Maize grown tor one month accumulated more N than tomato plants during 10
weeks The strong lesponse of maize to green manure can be due to a remmeralization of
N, partly immobilized 6 weeks after green manure application The nitrogen supplying
potential of legumes for succeeding non-legume crops estimated from the accumulation of
inorganic N in baie fallow soil (Bowen et al, 1988) may differ strongly depending on
the succeeding crop That N recovery by crops is often higher with incorporated than
mulched green manure, as described by Vacro et al (1989), was confirmed with indigo
for tomato and for maize in this experiment
The high amounts ot N which remained after 30 d maize in soybean green manuie
treatments would probably be taken up by maize if grown to grain maturity The final N
balance (Table 8) shows that the pioductivity in the low beds was significantly lower than
in the raised beds
49
CONCLUSIONS
Legume green manures in these experiments in the hot and humid season were
shown to have considerable potential to enhance the short term N fertility of the soil as
well as to substitute N fertilizer requirements for tomato fully or partially depending on
the way they are applied Further research on decomposition and N-release of legume
green manures, specific N-uptake patterns by different crops in contrasting environments
are needed to develop practicable recommendations for farmers on the optimal way of
application and insertion of legume green manures into specific cropping patterns
REFERENCES
Aber, J D,J M MeliUo, C A McCIaugherty 1990 Predicting long-term patterns of
mass loss, nitrogen dynamics, and soil organic matter formation from initial fine
litter chemistry in temperate forest ecosystems Can J Bot 68 2201-2208
Azam, F,K A Malik, and MI Sajjad 1985 Transformation in soil and availability to
plants of 15N applied as inorganic fertilizer and legume residues Plant Soil 86 3-
13
Bowen, W T, J O Quintana, J Pereira, D R Bouldin, W S Reid, and D J Lathwell
1988 Screening legume green manures as nitrogen sources to succeeding non-
legume crops I The fallow soil method Plant Soil 111 75-80
Bremner, J M 1965 Total nitrogen In Black, C A (ed ) Methods of Soil Analysis
Part 2 Agronomy 9 1149-1178 American Society of Agronomy, Madison
Broder, M W and G H Wagner 1988 Microbial colonization and decomposition of
corn, wheat and soybean residue Soil Sci Soc Am J 52 112-117
Cameron, K C and A Wild 1982 Comparative rates of leaching of chlonde, nitrate
and tntiated water under field conditions J Soil Sci 33 649-657
Cassman, K G, Munns, D N 1980 Nitrogen mineralization as affected by soil
moisture, temperature and depth Soil Sci Soc Am J 44 1233-1237
Coltman, R R 1989 Managing nitrogen fertilization of tomatoes using nitrate quick
tests p 619 In Asian Vegetable Research and Development Center (Ed ) Tomato
and pepper production in the tropics Proc Symposium on Integrated Management
Practices, Shanhua, Tainan, Taiwan (ROC)
Frankenberger, W T,and H M Abdelmagid 1985 Kinetic parameters of nitrogen
mineralization rates of leguminous corps incorporated into soil Plant Soil 87 257
271
50
Garrison, S A,G A Taylor and W O Dnnkwater 1967 The influence ot nitrogen
nutrition on flowering, fruit set and yield of processing tomatoes Am Soc Hort
Sci 91 534-543
Groffman, P M,D A Hendnx, and D A Crossley 1987 Nitrogen dynamics in
conventional and no-tillage agroecosystems with inorganic or legume nitrogen
inputs Plant Soil 97 315-332
Hills, F J,F E Broadbent, and O A Lorenz 1983 Fertilizer nitrogen utilization by
com, tomato, and sugar beet Agron J 75 423-426
Hochmuth, G J 1992 Concepts and practices for improving nitrogen management for
vegetables HorfTechnology 2(1) 121-125
Hunt H W 1977 A simulation model for decomposition in grasslands Ecology 58
469 484
Jackson, LE, and A J Bloom 1990 Root distribution in relation to soil nitrogen
availability in field-grown tomatoes Plant Soil 128 115-126
Janzen, HH and S M McGinn 1991 Volatile loss of nitrogen during decomposition
of legume green manure Soil Biol Biochem 23(3) 291-297
Jenny, H ,S P Gessel and F T Bingham 1949 Comparative study of decomposition
ot organic matter in tempeiate and tropical regions Soil Sci 68 419 432
Ladha, J K,S Miyan and M Garcia 1989 Sesbania rostrata as a green manure for
lowland rice Growth, N2 fixation, Rhizobium sp inoculation and effect of
succeeding crop yields and nitrogen balance Biol Fert Soil 7 191-197
Lorenz OA,andJF Bartz 1968 Fertilization lor high yields and quality of vegetable
crops p 327 352 In Nelson, LB (ed) Changing patterns in fertilizer use Soil
Sci Soc Amer,Madison, Wisconsin
Maiy, B,and Recous, S 1994 Measurement of nitrogen mineralization and
immobilization fluxes in soil as a means of predicting net mineralization Eur J
Agron 3 (4) 291-300
Meelu O P,and R A Morns 1988 Green manure management in rice-based cropping
systems p 209-222 In IRRI (ed) Green manure in rice farming IRRI, Los
Banos, Philippines
Myres, R J K and IM Wood 1987 Food legumes in the nitrogen cycle of farming
systems p 46-52 In Wallis, E S,and D Byth (eds ) Food legume improvement
for Asian farming systems ACIAR, Canberra, Australia
Olasantan, F O 1991 Response of tomato and okra to mtrogen fertilizer in sole cropping
and intercropping with cowpea J Hort Sci 66(2) 191-199
Olson, J S 1963 Energy storage and the balance of producers and decomposers in
ecological systems Ecology 44 322-331
51
Palm, C A, and P A Sanchez 1991 Nitrogen release from the leaves of some tropical
legumes as affected by their hgnin and polyphenols contents Soil Biol Biochem
23(1) 83-88
Reinertsen, SA, LF Elliot, VL Cochran, and GS Campbell 1984 Role of
available carbon and nitrogen in determining the rate of wheat straw decomposition
Soil Biol Biochem 16 459-464
Sarrantomo, M ,and T W Scott 1988 Tillage effects on availability of nitrogen to corn
following a winter green manure crop Soil Sci Soc Am J 52 1661-1668
SAS Institute, Inc 1989 JMP user's guide Version 2 SAS Institute Inc, Cary, N C
SAS Institute, Inc 1991 SAS user's guide SAS Institute, Sparks Press, Raleigh, North
Carolina, USA
Shennan, C 1992 Cover crops, nitrogen cycling and soil properties in semi irrigated
vegetable production systems HortScience 27(7) 749-754
Shuford, J W,D D Fntton, and D E Baker 1977 Nitrate-nitrogen and chlonde
movement through undisturbed field soil J Environ Qual, 6 255-259
Siegel, R S 1980 Determination of nitrate and exchangeable ammonium m soil extracts
by an ammonia electrode Soil Sci Soc Am J 44 943-947
Singh, A 1975 Use of organic materials and green manuring as feitilizers in developing
countries FAO Soils Bulletin 27 19-30
Snedecor, G W ,and Cochran, W G 1978 Statistical methods Sixth edition Iowa
State University Press, Ames, Iowa, USA
Soil Survey Staff (ed ) 1992 Keys to soil taxonomy 5th ed U S Department of
Agriculture
Stivers, L J,and C Shennan 1991 Meeting the Nitrogen needs of processing tomatoes
through winter cover cropping J Prod Agnc 4(3) 330 -335
Swift, M J,O W Heal, and JM Anderson 1979 Decomposition in terrestrial
ecosystems Blackwell Scientific, Oxford, Great Britain
Varco, J J,W W Frye, M S Smith, and C T Mackown 1989 Tillage effects on
nitrogen recovery by com from a nitrogen-15 labeled legume cover crop Soil Sci
Soc Am J 53 822-827
Vincent, J M 1970 A manual for the practical of root-nodule bacteria Blackwell
Scientific, Oxford, Great Britain
Weeraratna, C S 1979 Pattern of nitrogen release during decomposition of some green
manures in a tropical alluvial soil Plant Soil 53 287-294
Wieder, R K,and G E Lang 1982 A critique of the analytical methods used in
examining decomposition data obtained from litter bags Ecology 63(6) 1636-
1642
52
Wilson, G F,B T Kang, and K Mulongoy 1986 Alley cropping Trees as sources of
green-manure and mulch in the tropics Biological Agriculture and Horticulture 3
251-267
Yamoah, C F,and M Mayfield 1990 Herbaceous Legumes as Nutrient Souices and
Cover Crops in the Rwandan Highlands Biol Agnc Hort 7 1-15
53
4
Legume Decomposition and N Release when
Applied as Green Manure to Tropical Vegetable
Systems. II. In the Dry Season in Taiwan
ABSTRACT
The feasibility of meeting N needs of vegetables with legume green manures (GM)
was tested in a 6 months experimental cropping pattern in a field experiment in the dry
season (DS) at the Asian Vegetable Research and Development Center (AVRDC) in
Southern Taiwan Two legume species (soybean -Glycine max L Merr, and indigofera
-Indigofera tmctona L ) were grown for 60 days (d) and then both used as mulch or
incorporated into the soil A tomato (Lycopersicum esculentum Mill) or Chinese
cabbage (Brassica pekinensts ) crop was transplanted nght after GM application and
grown up to harvest (60 - 120 d) Green manure amended vegetable yields were
compared to vegetable yields with 30, 60, and 120 kg inorganic N fertilizer ha-1 The
residual effect of the fertilizing method on 30 d maize following the vegetable crop was
determined based upon biomass and N uptake Legume, vegetable and maize biomass,
yields, and N uptake were studied on two different bed systems (raised versus low beds)
simultaneously Legume decomposition was investigated in a litter bag study, and N
release in soil was followed with frequent soil sampling
Soybean accumulated 2 5 - 4 0 t ha *, and indigofera 0 4 - 0 9 t ha 1 biomass,
containing 95-143 kg N ha-1 and 5 -38 kg N ha-1, respectively Incorporated legume
decomposed significantly faster than mulched After the initial exponential weight loss
within the first 3 - 5 wk of 20 - 60 %, very little decomposition took place Before
legume incorporation (60 d) greater amounts of nitrate had accumulated in fallow than
legume plots Legume incorporation resulted m an exponential N release in soil, which
was significantly greater than with mulched legumes Nitrate contents decreased abruptly
six wk after legume application Cabbage and tomato yields were not increased above
those obtained m control by any of the GM treatments Although far more N was applied
with soybean compared with indigofera, N uptake in either succeeding vegetable crop
was comparable However maize following cabbage accumulated significantly more N
54
than if following tomato A residual effect of increased N availability, reflected in N
uptake values in maize was obtained in tieatments with soybean mcorpoiation
Two months fallow was more beneficial than was GM for cabbage and tomato
production in the dry season, since N accumulated in legumes was not released fast
enough to meet vegetable N needs
INTRODUCTION
Theie is common agreement among organic farmeis and agricultural teseaichers
that organic approaches to pest and nutrient management in vegetable production have not
been sufficiently studied (Grubinger, 1992) Rotation with legumes to supply N for
succeeding crops is an age-old practice, but is seldom seen on vegetable farms today
(Kelly, 1990) Studies on green manunng have reported many site and year-specific
results on dry matter accumulation and amount of N fixed (Hoyt and Hargrove, 1986,
Smith et al, 1987) Legumes intercropped with corn gave large increases in yield for the
tollowing crop of nee In a wet yeai legume green manures slightly mcreased the yield of
corn, in a dry year the corn yield decreased (Van de Goor, 1954) Stivers and Shennan
(1991), Abdul-Baki and Teasdale (1993) and our wet season results (Chapter 3), report
tomato yields following legume GM and mulch, comparable to those amended with
synthetic fertihzeis, while Lennartsson (1990) showed that vegetable yields following
green manures did not outyield those grown after fallow Further investigations are
needed to evaluate the potential role of legume green manures in horticulture and to
estimate the nsks involved before promoting it as a widespread practice for farmers
In a first study evaluating the use of legume green manures in tomato production in
the wet season in Taiwan, legumes proved to be effective nitrate catch crops Nitrogen
needs of tomatoes were partly or fully met by legume green manures depending on their
field management (Chapter 3) In this second study the same expenment was repeated in
the dry season in older to i) quantify legume biomass and nitrogen accumulation, and
legume decomposition and N release in the soil, n) establish a comparison to our wet
season expenments (Chapter 3) in as far as meeting N needs of tomato crops amended
with leguminous GM, in) compaie the response of another vegetable crop (Chinese
cabbage) to GM application, iv) evaluate the effect of actively growing plants on N
release, v) evaluate carry over benefits to a further succeeding crop (maize)
55
MATERIALS AND METHODS
Field experiment
A field experiment was conducted in the dry season 1993/94 (October 93 - April
94) on the expenmental farm of the Asian Vegetable Research and Development Center
(AVRDC) in Taiwan The soil is of Take series (loamy, mixed, hyperthermic,
Fluvaquentic Entochrept (Soil Survey Staff, 1992)), pH (H20) 8 2, total Kjeldahl N 0 7
g kg"1, total C 6 4 g kg"1 (Walkley-Black Method)
Experimental design
Experiments were conducted using raised and low bed systems The raised beds
were 45 cm high, 2 m wide with 2 m furrows between the beds The furrows were sown
with nee (Oriza sativa ) and permanently flooded The low beds were 20 cm high and
2m wide with 50 cm wide irrigation furrows between beds Both experiments were
adjacent such that the soil type, the cropping history and meteorological conditions were
the same
The raised bed field area was 27m by 54m and the low bed area was 27m by 26 m
The expenmental design for each bed system was a randomized complete block
Treatment plots were 2m by 6 m with four replicates
Table 1. Treatments of the expenmental cropping pattern m which legumes (60 d) are
grown in rotation with vegetables in raised and low beds (AVRDC, Taiwan, 1993/94)Vegetable yields amended with legume green manures (GM) are compared with those
amended with mineral N fertilizer N mineralization in the soil is studied in planted and
unplatted vegetable plots
treatments treatment legume species/ GM vegetable (planted)/
abbreviation fallow application/
fertilization
unplanted
1 Si soybean incorporation tomato/ cabbage/ unplanted
2 Sm soybean mulch tomato/ cabbage
3 Ii indigo incorporation tomato/ cabbage/ unplanted
4 Im indigo mulch t tomato/ cabbage
5 CkO (control) weedfree fallow OkgNha"1 tomato/ cabbage/ unplanted
6 Ck30 weedfree fallow 30 kg N ha"1 tomato/ cabbage
7 Ck60 weedfree fallow eOkgNha1 tomato/cabbage
8 Ckl20 weedfree fallow 120 kg N ha"1 tomato/ cabbaget Indigofera (living mulch) regrew after the
another two times Regrowth was poor (<100 kg dryto the tomatoes/ cabbage in respective ptots
tirst cut for mulch (60 d) and was trimmed back
matter ha '), cuttings were added as surface mulch
56
The eight treatments were as follows (Table 1) two legume species and, two GM
systems (mulch and incorporation) in all four combinations, and four treatments having
weedfree fallow (while legumes were grown in the legume treatments) with 0, 30, 60 or
120 kg N ha-1 applied to the vegetable crop The sub treatments were Chinese cabbage
and tomato each grown on one half (in lm by 6m) of the mam treatment plot
Green manure and vegetable crop
To obtain high legume biomass in a short time, the legumes were sown in four
double rows at double the noimal late for the legumes, as suggested by Yamoah and
Mayfield (1990) One double row was sown on each edge of the bed and the third and
foith double rows 65 cm distance trom the edge rows The rows in each double row
were 10 cm apart
The expenment commenced on 8 October 1993 and legumes were hand-sown at 80
seeds nr2 tor soybean (Glycine max (L ) Men ,2-3 seeds/10 cm row) and 1 32 g m"2
tor mdigofera (Indigofera ttnctona L) Local varieties were used Seeds were inoculated
with a rhizobium strain mixture that was specific foi each legume species, provided by
the Soil Science Department of the Chung Hsing University in Taichung, Taiwan
Phosphorus at 35 kg P ha ', as super phosphate and potassium at 83 kg K ha-1, as
potassium chlonde was broadcast in all beds On 6 December (60 d) all legumes were cut
at soil level, chopped into pieces of 10 cm and either incorporated by rototilhng to 15 cm
depth, or left as mulch on the soil surface in accord with the treatment On 8 December 26
d Chinese cabbage seedlings (Biassica pekinensis) and 30 d tomato (Lycopersiton
esculentum Mill, determinate bushy type, shoit duration, AVRDC line 5915-93-1 0-3)
seedlings were transplanted each as one row pel bed and spaced 40 cm within and 100
cm between rows Inorganic P and K fertilizers at 35 and 50 kg ha-1, respectively, were
applied to all plots A further 50 kg K ha-1 was applied on 24 December, and again on 14
January Foi the N fertilizer treatments 30 kg N ha-1 (as ammonium sulfate) was applied
at tiansplanting to the 30, 60 and 120 kg N ha-1 treatments The first side dressing of 30
kg N ha-1 was applied to the crop on 24 December in treatments providing 60 and 120 kg
N ha-1 A fuithei 60 kg N ha *was applied to the 120 kg N ha-1 treatment as second side
diessing on 14 January Chinese cabbage was harvested on 4 Febiuary Red tomatoes
were harvested on 1 March, with a final haivest on 15 - 16 Match Maize was sown after
tomato harvest on 18 March in 6 lows per bed (30 seeds nr2) and sampled 30 days later
on 18 April
Environmental monitoring
Weathei data were collected throughout the experimental period at the AVRDC
meteorological station Soil moisture was measured in the GM treatments, with
57
tensiometers placed in tomato subplots at tomato transplanting in treatments Ssi, Ssm,
Isi, Ism and CkO (Table 1 for abbreviations) at 15, 30 and 45 cm depth in the raised
beds, and at 15 and 30 cm depth in the low beds
Plant analysis
Legumes were sampled at 32 and 60 d Plants from 0 5 m2 area of each of the four
replicates, which was afterwards excluded from further sampling, were carefully dug out
to a depth of 15-20 cm and the soil then separated from the roots Root nodules per plant
were counted and samples of the nodules were cut open to assess their effectiveness by
the presence or absence of the pmk colour produced by hemoglobin (Vincent, 1970)
Shoots, roots, and nodules were dned at 60°C for 48 hours and weighed Nitrogen
content in shoots and roots including nodules were determined by the Kjeldahl distillation
method (Bremner, 1965) for 60 d legumes only
Fresh and dry weight of the total and marketable yield and total nitrogen content of
Chinese cabbage were measured At tomato harvest, marketable fruit fresh weight, fresh
and dry weights and nitrogen content of tomato fruit and plant were determined
Maize plants (including roots, 30 d) were pulled out from the soil and biomass and
total N were determined as a rough measure of the inorganic N available m the soil after
vegetable harvest The N balance at the end of the experiment was calculated following
the methods of Myres and Wood (1987)
Decomposition experiment
Nylon bags (mesh size 1 mm) containing 15 g fresh plant material (4 7-5 5 g dry
weight) were used to determine biomass breakdown of 60 d incorporated or mulched
soybean and indigofera The bags were filled with root and shoot material in the fresh
weight ratio Mulch treatments contained shoot material only On 6 December all bags
were either buried at 10 cm soil depth for incorporation treatments or left on the surface
as mulch treatment Decomposition bags were sampled at the same dates as the soil
sampling for inorganic N, namely 0, 2, 5, 8, 14, 29, 42, 62, 75 days after incorporation
(DAI) Two randomly chosen bags per treatment were retrieved, oven dned at 60°C for
48 hours and weighed Samples were ashed by dry combustion in a muffle furnace
(500°C) for 8 hours to determine original ash-free dry weight remaining (Aber et al,
1990)
Decomposition data analysis
Decomposition rates of two species can be compared in one site by fitting them to a
mathematical model to estimate constants describing the loss of mass over time The
58
equation for the single exponential decay function (Jenny et al, 1949, Olson, 1963)
seemed the most appropriate,
Nt=N0(l-e-kt),
where Nt is the biomass remaining, No is the ongmal biomass, k is the relative
decomposition rate of each GM treatment, t is the time in days The relative
decomposition rate k characterizes the loss of mass over time The assumption underlying
the single exponential model can be expressed in two ways, either the absolute
decomposition rate decieases linearly as the amount of substrate remaining declines, or
the relative decomposition rate remains constant (Wieder and Lang, 1982) For statistical
analyses the single exponential model was linearized (log transformed) Statistical
comparisons of slopes, intercepts and residual variances among series of individual
regressions were made using analysis of covanance technique (Snedecor and Cochran,
1978)
Inorganic N
Inorganic N content's in the soil under tomato
The effect of the legume species and method of GM application on the release of
inorganic N in the soil was monitored in treatments CkO, Ssi, Ssm, Isi, Ism, sampled on
the same days as decomposition bags Soil samples were collected with a 5-cm-diameter
auger from 5 treatments in blocks I, II and III At each sampling date three soil samples
at 0 - 30 cm depth were taken from each treatment Each sample was a mixed composite
collected from 4 locations in each plot Soil samples were passed through a 10 mm sieve,
extracted with 1 N KC1 (115 soil/water) and inorganic nitrogen (ammonium and nitrate)
was determined with an ammonia gas sensing electrode (Siegel, 1980)
Comparison ofinorganic N contents in soil underfallow, cabbage and tomato
The influence of the presence or absence of a crop on N mineralization was tested
Treatment plots CkO, Ssi, and Isi in blocks I, II, and III were split into three parts the
subplots grown to tomato or cabbage were shortened by 1 5 m at one end of the plot, and
this part (2m by 1 5m) was kept as weedtree tallow Soil samples from tomato, cabbage
and fallow subplots were taken as mentioned above
Statistical analysis
Data were analyzed by ANOVA proceduie using JMP Version 2 (SAS Institute,
Inc 1989) and SAS version 6 03 (SAS Institute, Inc 1991)
59
RESULTS
Environmental monitoring
Mean monthly air and soil temperature decreased from October 1993 to January
1994, and increased again from February to April (Table 2) The average rainfall during
the experimental period was low with the exception of the strong rainfall event in
February Soil matnc potentials were not strongly affected by GM treatments (data not
presented) and ranged between -0 02 and -0 06 MPa during the tomato growing season
At tomato transplanting mean values were -0 02 MPa, then soil moisture decreased
slowly to -0 06 MPa up to 8 wk after GM application The relatively high rainfall in
February 1994 provoked a sudden increase in soil moisture up to -0 005 MPa in all soil
depths After that it decreased again slowly to -0 04 MPa in the last four weeks At the
same soil depths soil moisture trends were comparable across bed systems
Table 2. Monthly weather data during the dry season (October to April), AVRDC,Taiwan, 1993/94
1993 1994
Oct Nov Dec Jan Feb Mar Apr
Mean air temp, °C 25 5 23 0 18 7 18 5 19 7 19 3 26 4
Mean soil temp, °C 26 7 24 5 20 5 20 3 20 8 20 9 27 6
Rainfall/month, mm 0 0 30 5 5 0 6 5 78 5 19 0 12 0
Mean evaporation, mm/ day 54 32 33 33 36 41 64
Legumes
Biomass and N accumulation of soybean was 3 - 4 times greater than that of
indigofera (Table 3) Indigofera shoot N concentration tended to be greater than in
soybean, but its root N concentration was lower Soybean biomass was only slightly
affected by the bed system in 1993/94 (10% less biomass in the low beds) whereas
indigofera shoot biomass was greater by 50% in the raised beds When grown in
1992/93, soybean biomass on raised beds exceeded that on low beds by 1 t ha-1 (data not
shown), and indigofera biomass was very low on both bed systems and was only half of
that obtained in 1993/94
60
Table 3. Biomass yield (dry matter), nitrogen content and nitrogen accumulation of
soybean and mdigofera (60 d), grown in two bed systems Values shown are means offour replicates (AVRDC, Taiwan, dry season, 1993) Least significant difference (LSD)between species within a bed system is shown at 0 05 level
1993
soybean indigo LSD (P<0 05)
tR *L R L R L
Dry matter (kg ha 1) shoot 2832 2610 851 560 516 254
root 343 275 92 85 98 51
N content (%) shoot 4 1 4 1 43 43 ns ns
root 23 20 1 8 1 8 05 02
N (kg ha !) shoot 116 107 36 27 23 11
root 8 22 2 1 3 1
total 124 129 38 28 25 6
t R raised beds
t L low beds
Decomposition
Incorporated GM decomposed faster than mulched in both bed systems (Fig 1)
After the initial exponential weight loss of 20-60 % within the first 3-5 weeks, very little
decomposition took place for the last 9 wk in incorporated treatments, with slightly more
decomposition in mulch tieatments Decay rates (k) of the same treatments were of the
same older of magnitude in both bed systems (Table 4) Decay rates of the 1992/93
experiment showed the same trends (data not shown) Incorporated mdigofera
decomposed fastest, followed by incorporated soybean Mulched soybean was more
resistant to decomposition than mulched mdigofera.
61
120
100
80
60
40
20
0
120
100
80
60
40
20
0
^Ov
Soybean incorporation
Soybean mulch
Indigo incorporation
Indigo mulch
^Raised beds
i I
\LSD 0.05
V±><
--^
~zzz^=--—^
Low beds
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14
Weeks after green manure application
Fig. 1. Decomposition of soybean and indigofera residues when used as mulch or
incorporated into the soil in the raised and low bed system during the dry season at
AVRDC, Taiwan, 1993/94. Error bars indicate least significant difference at 0.05 level
62
Table 4. Decomposition rates, k, for a period of 94 days, when soybean and indigotera(60 d) green manures were incorporated into the soil or left as surface mulch Values
weie calculated using the single exponential model for decomposition (AVRDC,Taiwan, 1993/94)
Green manure treatments bed system kW R2(b)
Soybean incorporation tR 0 0236 A 0 954 ***
*L 0 0251 A 0 798 *
Soybean mulch R 0 0094 B 0 891 **
L 0 0071 B 0 949 ***
Indigo incorporation R 0 0313 C 0 941 **
L 0 0350 C q 97j ***
Indigo mulch R 0 0162 D 0 979 ***
L 0 0147 D 0 945 **
t R raised beds $ L low beds
(a) K values within the bed system were tested using a pairwise t test comparison for slopes K values
with different letters are sig different at the 0 05 level (b) *, **, ***, significant at the 0 05, 0 01 and
0 001 level
Inorganic N
Inorganic N content in soil under tomato
Nitrate was the dominant form of inorganic N in the soil alter GM application, and
ammonium levels stayed low and ranged between 5 - 15 kg NHa-N ha ' (Fig 2, Fig 3)
Nitrogen mineralization in the soil during the 60 days prior to GM application (l e 0 days
after incorpoiation, DAI) resulted in an accumulation of 27 kg NO3-N ha lin fallow
(control and fertilizer N, kept weedfree fallow dunng this time) treatments in the raised
beds In legume plots in the raised beds only 7 kg NO3-N ha"1 were accumulated in the
soil, indicating that parts of the nitrate derived from soil N mineralization may have been
accumulated by legumes Soil nitrate content after soybean (5 kg NO3-N ha-1) was less
than after indigotera (16 kg NO3-N ha '), and fallow (28 kg NO3 N ha ') in the low
beds, but differences were not significant because of strong variability Nitrate contents
increased strongly within one wk after GM application in both bed systems Significant
differences in N-release (NO3) between tilled (fallow and incorporation treatments) and
untilled (mulch) treatments were found only in low beds Nitrate content in all GM
tieatments peaked at 3 to 5 weeks, and decieased markedly thereafter More NO3 was
released in both bed systems when legumes weie incorporated compared to mulched
Incorporated soybean did release more NO3 than incorporated indigofera in the raised
beds but diffeiences were not significant In the low beds the amount of NO3 released
1993/94)
season,
dry
Taiwan,
(AVRDC,
difference
signific
ant
least
indicate
bars
Error
reph
cates
three
of
means
are
shown
Values
beds
low
and
raised
on
prod
ucti
ontomato
for
soil
in
inco
rpor
ated
or
mulch
manure
green
as
applied
indi
gofe
raand
soybean
after
cm)
(0-3
0soil
in
contents
(NH4-N)
ammonium
and
(NO3-N)
Nitrate
2.
Fig.
appl
icat
ion
manure
green
after
Weeks
14
13
12
11
10
98
76
54
32
10
14
13
12
11
10
98
76
54
32
10
CO
(0.05)
|LSD
(0.1)
LSD
mulch
Soybean
—A—
incorporation
oybean
S—
+—
mulch
Indi
go~~°—
incorporation
Indi
go—
"•
—
Control
60
Zo80
en
z^100
~ob
120
^|
140
160
1993/94)
season,
dry
Taiwan,
(AVRDC,
difference
significant
least
indicate
bars
Error
reph
cate
sthree
of
means
are
shown
Values
beds
low
and
raised
in
tomato)
(cabbage,
planted
or
unplanted
when
plots
incorporation
or
mulch
manure
green
indigofera
and
soybean
control,
of
cm)
(0-30
soil
in
contents
)(NO3-N
Nitrate
3.Fig.
appl
icat
ion
manure
green
after
Weeks
14
1213
11
10
0123456789
14
1213
11
10
0123456789
J>
CT)
Iincorporation
Soybean
tomato
-*—
cabbage
-o—
unplanted
——
Iincorporation
Indigo
beds
Low
I
(control)
/ha
Nkg
0beds
Raised
0
20
40
60
80
100
1200
20
40
60
80
100
1200
20
40
60
80
100
120
zoCO
65
with soybean and indigofera was comparable When legumes were mulched released
amounts of NO3 were comparable across species
N-release attributable to legumes was relatively low compared to N-release m
fallow It is difficult to estimate the real N release due to the legume addition, as no
control treatment with a similar initial N-content as in legume plots was available A
rough approximation of N release in control starting off with the same nitrate content as
legume GM treatments could be made by a parallel shift of the actual control nitrate curve
starting at the same point (7 kg NO3-N ha !) With this assumption an initial N
immobilization (lag period) period of about two to foui wk after GM application would
be evident in all legume amended plots in either bed system A significant net N release
(compared to the approximated control) of 30 - 50 kg NO3-N ha 1, and 10 - 60 NO3 N
ha ] could be calculated during the peak nitrate release in the raised and low beds,
respectively
Comparison ofinorganic N contents in soil underfallow, cabbage and tomato
The effect of subplot treatments on soil nitrate contents was comparable across
main treatments of both bed systems (Fig 3) Over 14 wk most nitrate was found in
unplanted subplots, less in tomato and least m cabbage subplots Ten to 50 kg NO3-N ha
1 less nitrate was found m planted compared to unplanted plots between 3 and 8 wk
Mam treatments modified this general trend slightly in the first three wk only Peak nitrate
contents were found 3-5 wk after GM application, whereafter nitrate content consistently
declined to a minimum by wk 8 This decrease was more marked in planted than in
unplanted subplots N-uptake by tomato or cabbage, measured as the difference of NO3
in the soil in planted vs unplanted plots, started generally after 1 to 3 wk after
transplanting Soil nitrate contents increased by 4-10 and 20-26 kg NO3-N ha 1in low
and raised beds, respectively, during the 6 wk fallow penod after cabbage harvest in GM
amended plots and were as high as those of unplanted sub plots Cabbage apparently was
a stronger N sink than tomato as less NO3 was found under cabbage compared with
tomato
Soil ammonium contents weie incieased by 1 - 8 kg NH4-N ha 'in planted
compared to unplanted plots in GM treatments in the low beds, whereas in raised beds no
significant differences occurred (data not shown)
66
Cabbage yield
Cabbage yields tended to be greater in raised beds (Table 5) Soybean
incorporation, mulch, and mdigofera incorporation did not increase cabbage yields above
those in fallow (control) in either bed systems When mdigofera mulch was applied
cabbage yields were significantly reduced Cabbage total yields were significantly
increased with 60 to 120 kg N ha-1 in raised beds, and 120 kg N ha"l in low beds, while
cabbage head yields were already increased with 30 kg N ha-1 in raised beds, and 120 kg
N ha 1in low beds The cabbage heading index was slightly increased with mineral N
fertihzeis compaied to GM treatments When total dry matter yields were compared no
differences were found among treatments, with the exception for mdigofera mulch which
significantly leduced dry matter yield in the low beds
Table 5. Chinese cabbage yields (t ha-1) following legume green manure or fertilizer N
treatments Values shown are means of four replicates (AVRDC, Taiwan, 1993/94)
Chinese cabbage
Fiesh mattei Dry matter
Total yield
tR *L
Head yieldR L
Total yieldR L
Ammonium sulfate
(kg N ha"1)0 (control) 42 5 37 2 21 3 17 5 195 1 83
30 46 3 35 9 26 4 17 3 2 17 1 72
60 49 4 46 1 27 0 23 4 2 03 2 17
120 51 1 50 2 27 9 25 5 1 81 2 27
Legumes
Soybean incorporation 42 2 35 5 20 4 16 0 2 04 1 88
Soybean mulch 37 7 27 8 18 5 104 194 1 39
Indigo incorporation 40 2 35 2 20 4 14 5 195 1 76
Indigo mulch 33 7 25 6 15 6 79 1 60 1 29
LSD (P<0 05) 64 97 46 73 ns 0 48
t R raised beds
$ L low beds
67
Tomato yield
Tomato yields were comparable between bed systems (Table 6) The application of
GM did not raise tomato yields above those of the 0 or 30 kg N ha-1 treatments Heavier
N doses increased tomato yield and plant dry matter significantly in both bed systems
Tomato plant dry matter was significantly reduced by indigofera mulch if on low beds
There was a tendency towards lower tomato yields and plant dry matter in mulch than in
incorporation treatments
Table 6. Tomato fruit and biomass (yield per planted area, t ha-1) in raised and low beds
following legume green manure or fertilizer N treatments Values shown are means of
four replicates (AVRDC, Taiwan, dry season, 1994)
Tomato fruit yield
tR tL
Tomato plant dry matter
R L
Ammonium sulfate
(kgNha-1)0 (control) 40 7 46 0 1 87 1 82
30 46 2 47 0 1 85 190
60 56 2 59 0 2 10 2 29
120 67 3 73 3 2 62 2 83
Legumes
Soybean incorporation 42 6 39 8 187 1 83
Soybean mulch 33 4 29 0 142 1 56
Indigo incorporation 37 4 37 8 1 79 1 80
Indigo mulch 31 5 28 8 147 140
LSD (P<0 05) 127 10 7 0 58* 0 41
t R- raised beds
t L low beds
* significant at the 0 1 level
Maize yield
Dry matter was 30- 40 % greater in the raised compared to the low beds (Table 7)
Maize dry matter yields on raised beds were markedly enhanced by soybean
incorporation and equaled that of 120 kg N ha J Soybean mulch, indigofeia
incorporation or the application of 60 and 120 kg N ha-1 to the previous crop also
improved dry matter yields compared to the 0 kg N ha * control In the low beds greatest
dry matter was reached in soybean mulch and 120 kg N ha !, the only two treatments that
differed significantiy from the control Maize yields were greater following cabbage than
following tomato in both bed systems
68
Table 7. Residual effect of N fertilization (0, 30, 60, 120 kg N ha"1 or soybean or
indigofera green manure mulched or incorporated) to previous crops (cabbage and
tomato) on maize biomass (30 d) on raised and low beds Values shown are means of
four replicates (AVRDC, Taiwan, dry season, 1994)
Maize dry matter (30 d)
(tha1)raised beds low beds
Previous crop Cabbage Tomato tMPmeans
Cabbage Tomato tMPmeans
Ammonium sulfate
(kg N hd"1)0 (control)
30
1 4
1 7
1 3
1 2
1 4
1 0
1 2
1 5
09
1 0
1 1
1 3
60 2 1 1 6 19 1 6 10 1 3
120 23 25 24 1 9 1 4 1 7
Legumes
Soybean mcorporation
Soybean mulch
23
23
1 8
1 3
2 1
1 8
1 3
1 4
1 1
1 3
1 2
1 4
Indigo incorporation
Indigo mulch
20
1 6
1 6
1 1
1 8
1 4
1 3
1 2
09
09
1 1
1 1
20 1 6t SP means
LSD (P<0 05) ditt between MP means
LSD (P<0 05) diff between SP means
1 8
04
01
1 4 1 1 1 3
03
0 1
t MP main plotsi SP subplots
N-uptake and N-balance
Cabbage N-uptake increased with increasing N fertilizer dose in both bed systems
(Table 8a, Table 8b), but differences compared to the control were only significant with
60, 120 kg N ha-1 and indigofera mulch in low beds Slightly more N was taken up in
GM incorporation than mulch treatments
Increasing lates ot fertilizer N increased tomato N uptake Significantly more N
accumulation was evident only in 60 -120 kg N ha-1 treatments in both bed systems
Tomato N uptake in mulched GM treatments was less than in fallow (control) plots
Tomato N-uptake was increased by 20-30 kg N ha * with GM incorporation compared
with mulch Although tar more N (80 kg N ha"1) was added with soybean GM compared
with indigofera, comparable amounts of N were taken up by cabbage and tomato
following both legume species This was evident m both bed systems
03
77
333
ns
ns
LSD(P<005)
5-42
819
773
5-26
427
150
51
942
mulch
Indigo
5-70
629
985
1-52
438
758
45
738
incorporation
Indigo
435
924
773
622
648
862
134
6128
mulch
Soybean
incorporation
0-2
432
698
112
548
468
129
7122
Soybean
Legumes
1-64
850
3160
190
349
778
147
27
120
120
7-68
826
9128
8-38
241
684
87
27
60
60
3-64
919
1014
9-48
634
713
57
27
30
30
4-87
323
191
4-68
632
862
27
27
0(c
ontr
ol)
0ha-1)
N(kg
ha"1
)N(kg
sulfate
Ammonium
balance
balance
(NQj)
legume-N
N-
maize
tomato
re¬
maize
cabbage
total
Nsoil
fert
iliz
er/
Outputs
Outputs
DAI)
(0In
put
1993/94)
season
dry
Taiwan,
(AVRDC,
beds
raised
on
d)(30
maize
and
cabbage
tomato,
by
N-outputs
and
cabbage,
or
tomato
to
manure
green
soybean
or
indi
goor
ha"1,
Nkg
120
and
60,
30,
0,of
inputs
Nafter
balance
nitrogen
Apparent
8a.
Table
o
34
622
76
018
05)
(P<0
LSD
039
615
457
-196
119
534
34
10
24
mulch
Indigo
8-61
614
280
8-39
022
850
33
528
incorporation
Indi
go
936
922
259
947
527
643
119
4115
mulch
Soybean
incorporation
513
422
178
337
825
950
114
5109
Soybean
Legumes
9-38
923
0164
128
736
284
149
29
120
120
7-41
717
1130
-165
933
671
89
29
60
60
1-48
815
391
021
224
855
59
29
30
30
0-77
318
787
5-46
619
955
29
29
0(c
ontr
ol)
0(kgN
ha-1
)ha-1)
(kg
Nsulfate
Ammonium
balance
balance
(N03)
legume-N
N-
maize
tomato
Nmaize
cabbage
total
Nsoil
fertilizer/
Outp
uts
Outp
uts
DAI)
(0Input
1993/94)
season
dry
Taiwan,
(AVRDC,
beds
low
on
d)(30
maize
and
cabbage
tomato,
by
N-outputs
and
cabb
age,
or
tomato
to
manure
green
soybean
or
indi
goor
ha"1,
Nkg
120
and
60,
30,
0,of
inputs
Nafter
balance
nitrogen
Apparent
8b.
Table
71
Maize grown following cabbage accumulated more N than after tomato m both bed
systems The incorporation ofGM increased maize N-uptake by about 10 kg N ha-1 with
soybean and 20-25 kg N ha-1 with mdigofera m cabbage plots Differences in maize N
uptake following cabbage or tomato were significant between 0 kg N ha * (control) and
120 kg N ha-' and soybean GM applied to the vegetable crop
N-balance was positive in soybean GM treatments as it was with 120 kg N ha"1 m
both bed systems m cabbage plots, but only in soybean GM treatments in tomato plots
N-balance in cabbage plots was about 22 kg N ha-1 greater than in tomato plots in both
bed systems
DISCUSSION
Legume biomass and N accumulation
Legumes respond strongly to changes in photopenod and temperature Soybean
and mdigofera grown in the DS for 60 d accumulated half as much biomass as in the
same experiment in the wet season (WS) (Chapter 3) This corresponded to 60 kg N ha"1
less N accumulation in soybean in the DS compared with the WS N accumulation of
mdigofera was not reduced in the dry season (DS) These results are confirmed in a study
on the effect of season on legume biomass and N accumulation (Chapter 2) Soybean
grown for 60 d in Thailand accumulated comparable amounts of N (138 kg N ha-1 )
(Meelu and Morns, 1988), but about 50 kg N ha"1 more N was obtained with mdigofera
compared to our results Meelu and Morris (1988) stress the importance of the effect of
the environment on N accumulation by GM species which implies that GM species must
be adapted to the physical environment that they will expenence during growth
Decomposition
The exponential weight loss pattern suggests that residues contain labile and
recalcitrant fractions having different degrees of resistance to microbial degradation
Faster decomposition of mdigofera was caused by its smaller and more tender leaves and
less lignified stems as compared to those of soybean Decomposition rates of
incorporated soybean were similar in the DS (1993/94, 1992/93, data not shown) and
WS 1993 (Chaptei 3) experiments When mulched, decomposition rates of soybean in
the DS (1993/94) were about half of those m the WS (1993) and the DS (1992/ 93)
mdigofera decomposed slightly faster than soybean in the DS (1992/93, data not shown,
and 1993/94), whereas in the WS the reverse was true Decomposition rates of
incorporated GM differed less between seasons and years than if mulched because in the
former the residue is in a generally more favourable environment for microbial
72
decomposition e g close soil contact, adequate soil moisture etc,as has been
demonstrated by Wilson and Hargrove (1986) in the U S A
Decomposition processes can be predicted from initial litter chemistry (Aber et al,
1990, Mehllo et al, 1982, Neely et al, 1991) Seasonal effects on chemical composition
of legumes (i e C/N ratio, initial N-, lignm-, polyphenol- and tannin- content) have been
shown to exist in the same site (Chapter 2) However, when chemical composition of
soybean and indigofera (Chapter 2) were related to decomposition rates in DS and WS
(1993), surprisingly no common chemical components determined decomposition rates in
the same way in both seasons The relatively high polyphenol ( 3 7%) and tanmn (1 6%)
content of indigofera may have retaided decomposition compared with soybean
(polyphenol 1 7%, tannin 0 2%) in the wet season, whereas in the dry season
indigofera's lower C/N-ratio (10 6) and higher initial N content (4 2 %) may have
determined its taster decomposition compared with soybean (C/N ratio 12 2, N 3 9 %)
Results of that study confirm the complexity of decomposition processes where the
intei action of both resource quality and microclimate conditions influence the conditions
and the activity of decomposer communities and those in turn mediate processes of
decomposition and nutrient release (Neely et al, 1991)
Inorganic N
Significant amounts of inorganic N present in clean fallow plots prior to vegetable
crops may have been the result of an enhanced soil N mineralization during the transition
period from WS to DS (George et al, 1992) combined with only marginal losses through
leachmg in the DS The higher the soil N supply, the more legumes denve N to meet their
requuement from soil rather than from biological N fixation Low nitrate contents in
legume plots can be explained by the effectiveness of legumes to assimilate NO3 derived
from soil N mineralization (N catch crops) (George et al, 1994, Maidl et al 1991)
Nitrate release curves look like response curves to decomposition curves High
decomposition rates of indigofeia lead to N release in soil comparable to that of soybean,
although far less N (90 kg N ha-1) was incorporated with indigofera Minimal NO3 was
leleased in legume mulch treatments Reduced mineralization rates of surface applied
residues (mulch) weie attributed to poor soil/ lesidue contact and drastic temperature and
moistuie fluctuations at the soil surface (McCalla and Duley, 1943) The general N
release pattern was similar in DS and WS (Chapter 3), exponential in the first few weeks
and decreasing abruptly after 6 wk Griffith et al (1994) attributed the rapid proliferation
of microorganisms responsible for N mineralization to relationships between C N ratios
ot substrate and decomposers upon the addition of residues Low rates of N release at
later stages of decomposition may represent recalcitrant organic fractions that contribute
to the formation of soil humus (Wilson and Hargrove, 1986) Decline of soil inorganic N
73
6 wk after GM APPLICATION could be explained as a combined effect of increased
plant N-uptake, leaching losses due to strong rainfall in the middle February, and N
immobilization If microbial needs for mineral N in soil were large, this pool may have
been rapidly depleted and the decomposition rate of organic compounds would decline
(Mary and Recous, 1994), leading to N immobilization (6 wk ) and delayed N
lemineralization In experiments done by Broadbent and Tyler (1962) nitrate was
immobilized to a considerable extent when this was the only form available to soil
microorganisms, a process driven by the activity of nitnfiers in relation to that of the
heterotrophic flora Mary and Recous (1994) described immobilization - remineralization
of N as a function of the amount and nature of recently added oiganic residues and soil
mineral N, whereas basal mineralization is explained as a function of soil texture and
long-term C and N inputs
Vegetable yield
Nitrogen released by GM did not match demand by cabbage or tomato in order to
obtain yields comparable to those with high N fertilizer inputs The initial lag period
before N became available in the legume plots and the net decrease of N after 6 wk were
important factors leading to this mismatch Higher initial NO3 contents in soil followed
by a strong N mineralization in fallow (control) plus fertilizer compared to legume GM
treatments gave cabbage and tomato a better growth start Control and fertilizer treatments
starting at the same initial NO3 level as legume treatments were missing, leading to an
underestimate ot the N supplying capacity of legume GM for cabbage/ tomato production
in the DS High rainfall in the WS (Chapter 3) did not allow nitrate accumulation in
fallow plots Transplanting time may have been too soon after legume harvest, hence
tomato plants were stressed by an early N deficit in legume plots
Generally cabbage shows a strong response to N fertilization (Fieyman et al
1991) Smith and Hadley (1988, and 1992) reported lower yield responses to organic N
sources than to ammonium nitrate The inefficient N utilization from organic N material
was attributed to NH3 volatilization or a lower N release to the crop In a similar manner
to our data Lennartsson (1990) found that none of the GM treatments affected spring
cabbage yields significantly when compared with unmanured fallow
Tomatoes grown under adverse and sub-optimal conditions in the hot tropical WS
(Chapter 3) were able to use N released from green manures (soybean mulch and
indigofera incorporation) to produce yields comparable with those reached with N
fertilizer application Climatic conditions in the DS were far more favourable for tomato
production, for tomato yields were about 10 times greater Tomato plants grew better,
less flower-drop occurred and N uptake was about three to four times as great as in the
WS Interferences between N and other plant nutrients may have led to lower tomato
74
yield responses in legume amended plots The depletion of soil nutrients, particularly P,
the alteration of soil structure and the phytotoxicity from upland crops and their
immobilization of soil N once incorporated may contnbute to the disadvantages of legume
compared with fallow treatments (Hamid et al, 1984) In field experiments in California,
tomato yields for the legume cover-cropped plots were as high as those in the fertilizer-
treated plots, but the response to applied N was low (Stivers and Shennan, 1991) Wien
and Minotti (1987) reported that tomatoes forage efficiently for soil N, obtaining only 30
40% from fertilizer sources For the N treatments and cultivars used in their study, timing
of N application did not appeal to be critical, as long as there was residual soil N In
another study with tomatoes grown in rotation with alfalfa yield was superior to
continuously grown tomatoes and/ or when grown in rotation with soybean (Johnston et
al 1992)
Residual effect
Differences in maize yields and N uptake between GM and fertilizer plots were
small in the DS The residual effect of GM management on maize depended strongly on
the preceding vegetable crop grown Maize (30 d) accumulated 10- 27 kg N ha-1 more N
and 0 3 - 1 1 t ha l biomass in GM plots in the WS (Chapter 3) than in the DS Less
legume N applied, greater tomato N uptake and stronger N immobilization in the DS
compared to the WS may have been among the main reasons for lower maize
performances in the DS Maize N uptake was greater with incorporated GM m raised
beds, while maize yields were not affected by GM management in the low beds Only
small differences in maize N uptake due to GM management (no-tillage vs conventional
tillage) were reported by Vacro (1986) No-tillage was found to have a higher
contribution to soil organic N as N was released slower and less N was recovered by
succeeding crops (Vacro, 1986)
Badaruddin and Meyer (1990) reported that total N uptake m above-ground parts of
wheat following fallow was generally greater than that following GM crops, which was
confirmed herein by cabbage and tomato N-uptake In both bed systems N-balance was
positive in soybean amended plots
CONCLUSIONS
We can conclude that a 2 month weedfree fallow is more suitable than GM for
cabbage and tomato production in the DS in Taiwan, as N accumulated in legumes is
released neither consistently nor fast enough for vegetable crops In the WS, however,
legumes are excellent nitrate catch crops reducing leaching losses associated with tallow,
75
and have considerable potential to substitute partially or fully the N fertihzei requirement
for vegetable crops
Further studies are needed to understand N mineralization and N immobilization
patterns in the DS, in order to quantify the amount of N derived from GM and taken up
by vegetable crops, and to understand whether N from GM is biologically immobilized or
accumulates in soil organic matter fractions Further experiments are implicated which
start with the same NO3-N contents in the soil and permit comparison of cabbage/ tomato
yields in fallow, in N fertilizer and in GM treatments
REFERENCES
Abdul-Baki, A A ,and J R Teasdale 1993 A no-tillage tomato production system
using hairy vetch and subterranean clover mulches HortScience 28(2) 106-108
Aber, J D,J M Melillo and C A McGlauherty 1990 Predicting long-term patterns of
mass loss, nitrogen dynamics, and soil organic matter formation from initial fine
litter chemistry in temperate forest ecosystems Can J Bot 68 2201-2208
Badaruddin, M,and D W Meyer 1990 Green manure legume effects on soil nitrogen,
grain yield, and nitrogen nutrition of wheat Crop Science 30 819-825
Bremner, JM 1965 Total nitrogen p 1149-1178 In Black, C A (ed ) Methods m
soil analysis Part 2 Agronomy 9 American Society of Agronomy, Madison
Broadbent, F E,and K B Tyler 1962 Laboratory and greenhouse investigations of
nitrogen immobilization Soil Sci Soc Am J 26 459-462
Freyman, S,PM Toivonen, PW Pernn, WC Lin, and J W Hall 1991 Effect of
nitrogen fertilization on yield, storage losses and chemical composition of winter
cabbage Can J Plant Sci 71(3) 943 946
George, T ,J K Ladha, R J Buresh, and D P Garnty 1992 Managing native and
legume-fixed nitrogen in lowland rice-based cropping systems Plant Soil 141 69-
91
George, T,J K Ladha, D P Garnty, and R J Buresh 1994 Legumes as nitrate
catch' crops during the dry-to-wet transition in lowland nee cropping systems
Agron J 86(2) 267-273
Griffiths, B S,M MI Van Vuuren, and D Robinson 1994 Microbial grazei
population in a 15N labeled organic residue and the uptake of residue N by wheat
Eur J Agron 3(4) 321-325
Grubinger, V P 1992 Organic vegetable production and how it relates to LISA
HortScience 27(7) 759-760
76
Hamid, A ,G M Paulsen, and H G Zandstra 1984 Performance of rice grown after
upland crops and fallow in the humid tropics Trop Agnc (Trmidad) 61(4) 305-
310
Hoyt, G D,and W L Hargrove 1986 Legume cover crops for improving crop and
soil management in the southern United States HortScience 21 397-402
Jenny, H ,S P Gessel, andFT Bingham 1949 Comparative study of decomposition
rates of organic matter in temperate and tropical regions Soil Sci 68 419-432
Johnston, R,V Shattuck, and J Seliga 1992 The effects of crop rotations and
nitrogen rates on processing tomato yields HortScience 27(6) 134
Kelly, W C 1990 Minimal use ot synthetic fertilizers in vegetable production
HortScience 25(2) 168-171
Lennartsson, E K T 1990 The use of green manures in organic horticultural systems
Papei presented at the 8th International Conference of the International Federation
of Organic Agricultural Movements, Budapest, Hungary
Maidl, FX ,J Suckert, R Funk, and G Fischbeck 1991 Field studies on nitrogen
dynamics after cultivation ot grain legumes J Agron Crop Sci 167 259-268
Mary, B,and S Recous 1994 Measurement of nitrogen mineralization and
immobilization fluxes in soil as a means of predicting net mineralization Eur J
Agron 3(4) 291-300
McCalla, T M,and F L Duley 1943 Disintegration of ci op residues as influenced by
subtillage and plowing Agion J 35 306 315
Meelu, O P,and R A Moms 1988 Green manure management in rice-based cropping
systems p 209-222 In IRRI (ed) Green manure in rice farming IRRI, Los
Banos Philippines
Melillo, J M JD Aber, and J E Muratore 1982 Nitrogen and lignin control of
hardwood leaf litter decomposition dynamics Ecology 63(3) 621-626
Myers, R J K,and I M Wood 1987 Food legumes in the nitrogen cycle of farming
systems p 46-52 In Wallis, E S and D Byth (eds ) food legume impiovement
for Asian farming systems ACIAR, Canberra, Australia
Neely, CL,MH Beare, W L Haigiove, and D C Coleman 1991 Relationship
between fungal and bacterial substrate-induced respiration (SIR), biomass and plant
residue decomposition Soil Biol Biochem 23(10) 947-954
Olson, J S 1963 Energy storage and the balance of producers and decomposers in
ecological systems Ecology, 44 322-331
SAS Insititue, Inc 1991 SAS users guide SAS Institute, Sparks Press, Raleigh,
North Carolina USA
SAS Institute, Inc 1989 JMP users guide Version 2 SAS Institute Inc, Cary, N C
77
Siegel, R S 1980 Determination of nitrate and exchangeable ammonium in soil extracts
by an ammonia electrode Soil Sci Soc Am J 44 943-947
Smith, M S,W W Frye, and J J Varco 1987 Legume winter covei crops Adv Soil
Sci 7 95-139
Smith, S R,and P Hadley 1988 A comparison of the effects of oigamc and inorganic
fertilizers on the growth response of summer cabbage (Brasiica oleracea var
capitatacv HispiFl) J HortScience 63 615-620
Smith, S R,and P Hadley 1992 Nitrogen fertilizer value of activated sewage derived
protein effect of environment and nitrification inhibitor on NO3 release, soil
microbial activity and yield of summer cabbage Fertilizer Res 33(1) 47-57
Snedecor, G W ,and W G Cochran 1978 Statistical methods Sixth edition Iowa
State University Press Ames, Iowa, USA
Soil Survey Staff (ed ) 1992 Keys to soil taxonomy 5th ed U S Departmnt of
Agriculture
Stivers, L J,and C Shennan 1991 Meeting the nitrogen needs of processing tomatoes
through winter cover cropping J Prod Agnc 4(3) 330 -335
Vacro, J J 1986 Tillage effects on transformation of legume and fertilizer nitrogen and
crop recovery of residue nitrogen Ph D diss University of Kentucky, Lexington
Van de Goor, G A W 1954 The value of some leguminous plants as green manures in
companson with Crotolana juncea J Netherlands Agr Sci 2 37-43
Vincent, J M 1970 A manual for the practical study of root-nodule bacteria Blackwell
Scientific Oxford, GB
Wieder, R K,and G E Lang 1982 A critique of the analytical methods used in
examining decomposition data obtained from litter bags Ecology 63(6) 1636-
1642
Wien, H C,and P L Minotti 1987 Growth, yield and nutrient uptake of transplanted
fresh-market tomatoes as affected by plastic mulch and initial nitrogen rate J
Amer Soc Hort Sci 112(5) 759-763
Wilson, D O ,and W L Hargrove 1986 Release of nitrogen from cnmson clover
residue under two tillage systems Soil Sci Soc Am J 50 1251 1254
Winsor, GW, and MIE Long 1967 The effects of nitiogen, phosphorus
potassium, magnesium, and lime in factorial combination on ripening disorders of
glasshouse tomatoes J Hort Sci 42 391-402
Yamoah, C F,and M Mayfield 1990 Herbaceous legumes as nutnent sources and
cover crops in the Rwandan Highlands Biol Agnc Hort 7 1-15
78
5
Legume Decomposition and N Release when
Applied as Green Manure to Tropical Vegetable
Systems. III. Two On-Farm Studies in the
Philippines
ABSTRACT
The applicability of results on the use of legume green manures (GM) to substitute
N fertilizer tor tomato crops gained in Taiwan was tested in two on-farm field
experiments on two different sites (Mariano Marcos State University, MMSU, Batac, and
in collaboiation with the Bukidnon Resources Company Inc, BRCI, in San Juan) in the
Philippines Soybean, indigofera and mungbean were grown for 60-70 d and
incorporated or applied as surface mulch to the tomato crop Legume decomposition
(litter bag study), N release in the soil (inorganic N), and tomato N uptake (nitrate sap
sampling, and total N) weie monitored Tomato yields obtained with GM were compared
to those amended with local N fertilizer recommendations Legume N recovery in tomato
and soil organic matter (SOM) fractions (mobile humic acids, calcium humates) was
traced with 15N at MMSU
Soybean was the most promising legume species, accumulating 3 - 41 biomass ha-1
and 106 - 140 kg N ha-1 Nitrogen released to the soil reached 80 - 100 kg NO3 ha-1
within 2 - 8 weeks after GM application, depending on the site Tomato yields were
doubled with soybean GM at MMSU, while those at BRCI were not affected by any of
the GM or fertilizer treatments The tomato crop recovered 9 - 15% gieen manure N at
MMSU Thirty to 60 percent soybean N remained in the soil after tomato harvest At that
time GM-N had not accumulated m labile fractions but was mostly found in the humin
fraction of the SOM The results present evidence that N release dynamics with GM are
comparable across sites Tomato yield response to applied N (GM or fertilizer N)
however depends on the soil properties, particularly soil N mineralization
79
INTRODUCTION
Previous studies on the use of legume green manures (GM) in vegetable farming
systems in Taiwan (Chapter 3 and 4) have shown that the potential N fertilizer
substitution with GM for tomato production depends greatly on the growing season The
validity of these results was tested in two on-farm experiments in two vegetable
production areas of the Philippines the provinces of Ilocos Norte (Luzon) and Bukidnon
(Mindanao) Although these two areas differed strongly in climatic, environmental and
economic conditions a legume crop fitted well into the commonly-used cropping systems
Major reasons for the interest in introducing legumes for GM were, however, different
The rainfed lowlands of the province Ilocos Norte are characterized by intensive cioppmg
systems, although soil fertility and rainfall distribution appear unfavourable (Tnpathi,
1995) Rice is grown during the wet season (WS) and upland crops (legumes, maize,
vegetables) are grown in the dry season (DS) The drying of the soil subsequently to
flooded rice favors aerobic N transformations and nitrate accumulates (George et al,
1992, George et al, 1993) Soil is subjected to increased tillage and irrigation, and often
to high N inputs during the DS when vegetables are grown Upland crops do not
generally deplete soil NO3 which is prone to loss when fields are flooded for rice
production in the WS (George et al, 1993) In this study legumes were grown as a post-
rice crop in order to fix atmosphenc N via biological N fixation and accumulate soil
nitrate in their biomass (nitrate catch-crops), and to be used as GM for vegetable crops in
order to reduce N inputs for vegetable production The addition of GM to this poor soil
may enhance soil microbial activity (Fraser et al, 1988) and short term soil organic N
pools (Muller and Sundman, 1988)
A project including a tomato-paste plant (Bukidnon Resources Company Inc,
BRCI) was started m 1993 in the Bukidnon province, Mindanao It combines a contract
farming scheme and the farmers' cooperative approach to assist about 3,000 small-holder
farmers to grow tomatoes The strategy of using leguminous GM for tomato production
was tested in order to reduce external costs and to improve crop rotation Soils in
Bukidnon are rich in organic matter and of volcanic origin Rather extensive subsistence
farming is practiced in this area due to low income
The overall objective of this study was to test the feasibility of meeting N-needs of
tomatoes with legume GM at two locations by integrating the legumes into the specific
cropping patterns of those areas This cropping strategy was tested for site specificity by
quantifying legume biomass, N-fixation, N accumulation, legume biomass
decomposition, N release to the soil, and tomato yield and N-uptake Tomato N nutrition
was monitored by NC>3-sap samplings at BRCI In order to describe N flows in the soil
80
and tomato plants legumes were labeled with 15N in an additional experiment, and 15N
was traced in tomato plant and organic matter fractions, at MMSU in Ilocos Norte
MATERIALS AND METHODS
Field trials
The first experiment was conducted on the experimental farm of the Mariano
Marcos State University (MMSU) in Batac, Ilocos Norte (IRRI rainted lowland
consortium site) from October 1994 to Apnl 1995 Average air temperatures in these six
months were 27 3° C (max 33°, mm 20°C) After strong rainfalls m October (175 mm)
no more rainfall occurred MMSU soil is a Fluvaquentic Ustropept (fine silty, mixed,
isohyperthermic), pH (H20) 8 1, total Kjeldahl N 0 7 g kg"1 (Bremnei, 1965), total C
5 9 g kg-1 (Wdlkley-Black method) This soil was previously cropped to nee to obtain a
homogeneous soil mineral N distribution Rice straw was removed from the field before
the tnal started
The second experiment was conducted on the experimental farm of the tomato
processing company Bukidnon Resources Company Incorporation (BRCI) in San Juan,
Bukidnon, Mindanao from April to October 1995 Average air temperatures in these
months were 24 1° C (max 28 3°, min 19 8°C) Total rainfall in these months was
1471 mm, with an average daily rainfall of 25 mm BRCI soil is a clayey, koalimtic,
isohyperthermic Ultisol, pH (H2O) 5 7 (after liming with 5 t ha"1 CaC03), total KjeldahlN 2 1 g kg-1 (Bremner, 1965), total C 19 5 g kg"1 (Walkley-Black method) This soil
was previously cropped with corn foi 1 month Corn stubbles were removed before the
expenment started
Experimental design
MMSU experiment
A randomized complete block design with split-plots and four replicates was used
Mam tieatment plots were 6m by 6m, sub-treatment plots 2m by 6m The 8 main
tieatments were (Table 1) two legume species (soybean and indigofera), two green
manure systems (mulch and incorporation), and four controls having weedfree fallow
while legumes were grown in the legume treatments, with 0, 38, 75, or 150 kg N ha •
applied to the tomato crop The three sub treatments were fallow, early (transplanting
tomato plants right after GM application) and late transplanting (transplanting tomato
plants two weeks after GM application)
81
Table 1 Main treatments of field expenments performed at the Manano Marcos State
University (MMSU), Ilocos Norte, and in collaboration with the BRCI company at San
Juan, Mindanao, 1994/95, Philippines
MMSU
treatments (abbrev) first crop (legumes) second crop (vegetables)
fertilization /
green manure management
1 CkO
(control)
fallow OkgNha-1 tomato
2 Ck38 fallow 38 kg N ha-1 tomato
3 Ck75 fallow 75 kg N ha-1 tomato
4 Ckl50 fallow 150 kg N ha-1 tomato
5 Si soybean incorporation tomato
6 Sm soybean mulch tomato
7 Ii indigofera incorporation tomato
8 Im indigofera mulch tomato
BRCI
treatments (abbiev) first crop (legumes) second crop (vegetables)
fertilization /
green manure management
1 CkO
(control)
fallow 0 kg N ha-1 tomato
2 Ck30 fallow 30 kg N ha-1 tomato
3 Ck60 fallow 60 kg N ha-1 tomato
4 Ckl20 fallow 120 kg N ha-1 tomato
5 Si soybean incorporation tomato
6 Sm soybean mulch tomato
7 SPi soybean residue t incorporation tomato
8 SPm soybean residue t mulch tomato
9 Mi mungbean incorporation tomato
10 Mm mungbean mulch tomato
11 MPi mungbean residue t incorporation tomato
12 MPm mungbean residue t mulch tomato
t legume residue after pod harvest
82
BRCI experiment
The same experimental design was used but with three replicates Main treatment
plots were 4 8m by 4 5m, sub-treatment plots grown to tomatoes were 4 8m by m, sub-
treatments kept fallow were 4 8m by 1 5m The 12 mam treatments were (Table 1) two
legume species (soybean and mungbean), two application rates (whole plant or residues
only after pod harvest), two green manure systems (mulch and incorporation), and four
controls having weedfree tallow while legumes were grown in the legume treatments,
with 0, 30, 60, or 120 kg N ha ' applied to the tomato crop The two sub-treatments
weie fallow, and tomato transplanted right after legume application
Green manure and tomato crop
The experiment commenced on 6 October 1994 (MMSU) and on 17 April 1995
(BRCI) Legumes were inoculated with a Rhizobium strain mixture, specific for each
legume species and provided by the Soil Microbiology Unit at IRRI, to assuie uniform
inoculation with N-fixmg bacteria Legumes were sown at double normal seeding rate in
order to obtain a high legume biomass in a short time as suggested in Chapter 2
Legumes were hand sown in rows at 80 seeds nr2 for soybean (Glycine max L ) at
MMSU, and 55 seeds nr2 for soybean and mungbean (Vigna radiata (L) Wilczek) at
BRCI, and 1 32 g m2 for indigofera (Indigofera tinctona L )
Phosphorus at 35 kg P ha-1 as super phosphate, and potassium at 83 kg K ha-1 as
potassium chloride was bioadcast in all beds On 13 December 1994 (MMSU, 74 d) and
on 21 June 1995 (BRCI, 66 d) legumes were cut at ground level, chopped into pieces ot
5-10 cm and either incorporated by rototillmg to 15 cm depth, or left as mulch on the soil
surface as required for treatment At BRCI legumes were incorporated 27 June 1996
On 15 December 1994 (MMSU) and 28 June 1995 (BRCI), 24-day-old tomato
(Noithern Food Corporation (NFC-line) for MMSU and 14-day-old tomato (BRCI
variety 1403, selection 1584) seedlings were transplanted in two rows per bed spaced 40
cm within and 100 cm (MMSU) or 150 cm (BRCI) between rows Tomato fertilization
was according to regional recommendations which were 30 kg P ha-1, as super
phosphate, and 13 5 kg K ha-1 as potassium chloride as basal fertilizer application at
MMSU Nitrogen basal fertilizer application were 48, 24, and 12 kg N ha-1 for Ckl50
(normal), Ck75 (half) and Ck 38 (quarter), respectively at MMSU Eighteen kg K ha"1
were applied to all treatments foi the first side dressing 7 January 1995, and 60, 30, and
15 5 kg N ha"1 to Ckl50, Ck75 and Ck38, respectively Fomteen kg K ha J were
applied to all treatments for the second side diessing 22 January 1995, and 42, 21, and
10 5 kg N ha"1 to Ckl50, Ck75 and Ck38, respectively At BRCI 2000 kg chicken dung
ha-1 (containing 2% N, 1 3% P and 2 1% K) and 40 kg P ha l, as solo phosphate, was
applied right before transplanting tomatoes Fifty percent of the following K, Mg, Zn and
83
B fertilizer were applied to the tomato crop at each side dressing (5 and 19 July 1995)
320 kg K ha"1, as potassium chloride, 23 kg Mg ha"1 as Kiesent (MgS04), 3 kg Zn ha l
as zmk sulphate, and 3 kg N ha-1 as borax For N fertilizer treatments 30 kg N ha-1 was
applied to Ck30, Ck60 and Ckl20 for the first side dressing (5 July 1996), and 30 and
60 kg N ha-1 for Ck60 and Ckl20, respectively, for the second side dressing (19 July
1996) Tomatoes were harvested sequentially on 28 February, 7, 14, 22 March, and 4
April 1995 (MMSU) and from 9 September (first harvest) to 18 October 1996 (fifth
harvest) at BRCI
Plant analysis
Legumes were sampled at 74 d (MMSU) and 66 d (BRCI) The plants from 0 64
m2 (micro plot, MMSU, see 15N expenment) and 0 5 m2 (BRCI) area of each treatment
replicate, which was afterwards excluded from further sampling, were carefully dug out
to a depth of 15-20 cm and the soil separated from the roots Shoots, roots, and nodules
were dried at 60°C for 48 hours and weighed Nitrogen content in shoots and roots
including nodules were determined by the Kjeldahl distillation method (Bremner, 1965)
At each harvest, marketable fruit fresh weight, was recorded, and at final harvest
fresh and dry weights and nitrogen content of tomato plants were determined
Decomposition experiment (at MMSU only)
Nylon bags (mesh size 1 mm) containing 15 g fresh plant matenal (4 7-5 5 g dry
weight) were used to determine biomass breakdown of 74-day-old incorporated or
mulched soybean and indigofera The bags were filled with root and shoot matenal in the
fresh weight ratio Mulch treatments contained shoot material only On 15 December all
bags were either buried at 10 cm soil depth for incorporation treatments or left at the
surface as mulch treatment Decomposition bags were sampled at the same dates as the
soil sampling for inorganic N, namely 0, 5, 21, 36. 58, 77, 113 days after incorporation
(DAI) Two randomly chosen bags per treatment were retneved, oven dned at 60CC for
48 hours and weighed Samples were ashed by dry combustion in a muffle furnace
(500°C) for 8 hours to determine original ash-free dry weight remaining (Aber et al,
1990)
Decomposition data analysis
Decomposition rates of the two species can be calculated and can be compared
among treatments of one site by fitting them to a mathematical model to estimate constants
describing the loss of mass over time The equation for the single exponential decay
function (Jenny et al, 1949, Olson, 1963) seemed the most appropnate,
Nt = No(l-e-kt),
84
where Nt is the biomass remaining, No is the original biomass, k is the relative
decomposition rate of each green manure treatment, t is the time in days. The relative
decomposition rate k characterizes the loss of mass over time. The assumption underlying
the single exponential model can be expressed in two ways, either the absolute
decomposition rate decreases linearly as the amount of substrate remaining declines, or
the relative decomposition rate remains constant (Wieder and Lang, 1982). For statistical
analyses the single exponential model was linearized (log transformed). Statistical
comparisons of slopes, intercepts and residual variances among series of individual
regressions were made using analysis of covariance technique (Snedecor and Cochran,
1978).
Inorganic N
The effect of the legume species and the green manure application method on the
release of inorganic N in the soil was monitored in treatments (see Table 1 for
abbreviations) Ck 0, Si, Sm, Ii, Im, which were sampled -74, 0, 5, 21, 36, 58, 77, 113
DAI (MMSU) and -7, 0, 14, 28, 42, 56, 70, 84, 96, 110 DAI (BRCI). Soil samples
were collected with a 5-cm-diameter auger from the five treatments in all blocks. At each
sampling date three soil samples at 0 - 30 cm depth were taken from each treatment.
Additional soil samples were taken at -74 (start of the experiment), 0, and 113 DAI at 30 -
60 cm soil depth at MMSU. Each sample was a mixed composite collected from 4
locations in each plot. Soil samples were dried, passed through a 10 mm sieve, extracted
with 1 N KC1 (1:1.5 soil/water) and inorganic nitrogen (ammonium and nitrate) was
determined with an ammonia gas sensing electrode (Siegel, 1980).
Plant petiole sap nitrate analysis (at BRCI only)
Tomato plants were sampled weekly for plant petiole sap nitrate content (sap-N)
between 6 and 10 am, two days after the second weekly irrigation, form 42 DAI (9
August 1995) to 91 DAI (26 September 1995). The fifth leaf (counted from the top) of 5
randomly selected plants per treatment was collected, in order to sample the most recently
matured leaf (Drews and Fischer, 1989; Drews and Fischer, 1992). Petioles were
chopped into 1 cm pieces and squeezed with a garlic press. Petiole sap was diluted by 50
times with distilled water, and mixed thoroughly for 1 minute. One drop of this solution
was poured onto two reaction zones of the Reflectoquant Nitrate test strips, and sap-N
was determined by refractrometric reading on the RQ-flex instrument (RQflex, Merck).
85
N fixation and 15N experiment (MMSU only)
Biological nitrogen fixation
Legumes and reference plants were grown in a small expenment run in parallel in a
field adjacent to the main field expenment to determine legume BNF with the difference
method (Talbott, 1985) Seeding rates and harvest dates (74 d) were the same as those
for legumes in the main field expenment A non-nodulating soybean hne (provided by the
Niftal project at IRRI) was the reference plant for soybean and an upland nee vanety (IR
600 80 -46A) was the non N-fixing reference plant for indigofera Plants were grown on
8 m2 plots with two replicates
Production ofi5N labeled legume plant material
The use of 15N labeled plant material makes it possible to distinguish between
tomato N derived from l5N labeled legume residue and soil native N The 15N
enrichment of young soil organic matter fractions was determined to quantify whether
residue-l5N had accumulated in these fractions
To produce l5N labeled legume matenal the above ground biomass of legumes
grown in an adjacent plot to the main field experiment (labeling plot) was misted with a
0 5 % solution made of Urea (30 atom % 15N) at a total rate of 10 kg N ha"1 (Zebarth et
al, 1991) 15N fertilizer was split for progressive foliai applications at 21, 28, 35, 42,
and 48 d
Application of I5N labeled legume material as GM
One day before legume harvest metal frames 08 by 08m by 03m height (micro
plots) were pushed into the soil to a depth of 25 cm in Si, Sm, Ii, Im treatments of the
main field experiment Legumes within the metal frames were removed including roots at
74 d 15N-labeled legumes were carefully dug out at 74 d from the labeling plot to a depth
of 15-20 cm and the soil separated from the roots Plants were chopped into pieces of 5-
10 cm and applied (incorporated or as mulch) to micro plots of the corresponding
treatments in the main field expenment
Legume 1SN recovery in vegetables and soil
Two tomato seedlings were transplanted into each micro plot Tomato fruit yield
and plant biomass, N and 15N content were determined to calculate 15N recovery in the
tomato plant Soil mobile humic acids (MHA) and calcium humates (CaHA)) were
determined as carbon pools representing early and late stages of the humification process
(Oik et al, 1995) Soil sampling for organic matter extraction and soil 15N determination
m was done in control and soybean incorporation plots 0 and 113 DAI
86
Environmental monitoring
In the MMSU experiment soil moisture changes as affected by green manure
treatments, were monitored with tensiometers placed in treatments Si, Sm, Ii, Im, CkO at
15, 30 and 45 cm depth at tomato transplanting
Statistical analysis
Data were analyzed by ANOVA procedure using JMP Version 2 (SAS Institute,
Inc 1989) and SAS version 6 03 (SAS Institute, Inc 1991)
RESULTS
Legume biomass and N accumulation
Soybean accumulated comparable amounts of biomass and N at MMSU and BRCI
(Table 2) Slightly lower soybean yields obtained at BRCI compared to MMSU may
partly be due to lowei seeding densities at BRCI Mungbean biomass and N accumulated
were infenoi to those ot soybean Indigofera yields were extremely low
Table 2. Biomass and N accumulation ot soybean and indigofera grown tor 74 d at
MMSU and of soybean and mungbean grown for 66 d at BRCI Values within
parentheses indicate standard deviation
Soybean IndigoferaMMSU
MungbeanMMSU BRCI BRCI
biomass shoot t 4 1 (11) 19 (0 5) 0 15 (0 03) 0 8 (0 3)
(tha-l)root$podtotal
0 1 (0 01)
4.2 (1.1)
16 (0 2)3.4 (0.7)
0 02 (0 02)
0.17
(0.05)
0 3 (0 1)1.1 (0.4)
N% shoot troot £pod
3 4 (0 7)15 (0 3)
16 (0 3)
4 8 (0 2)
3 5 (0 3)14 (0 3)
17 (0 2)
3 5 (0 1)
Nkghal shoot t 139 4 (47 4) 30 7 (115) 4 8 (0 8) 14 1 (6 2)
-H-
_
2
a.2
16 (0 3)
140.1 (47.2)75 3 (10 7)106.0 (20.4)
0 3 (0 3)
5.0 (0.9)
119 (3 9)26.0 (7.8)
t shoot includes pods, if pods are not presented separately$ root biomass was not sampled at BRCI
87
Table 3. Chemical properties of soybean and indigofera (74 d) at MMSU, 1994,
Philippines Values shown are means of four replicates.
Chemical Plant parts Soybean Indigofera LSD
properties
C/N latio shoot 13.75 12 29 ns
root 28.16 31.85 ns
Ligmn % shoot 5 41 5.30 ns
root 14.52 9 96 ns
Polyphenol % shoot 1.29 2.54 0.48 *
Tannin % shoot 0.55 0.48 0.04 (0.1)*, 0.1, significant at 0.05, and 0.1 level respectively
There were no significant differences in N concentrations, C/N ratio and lignin
content between soybean and mdigofera grown at MMSU (Table 3). Indigofera had a
higher polyphenol concentration, but a lower tannin concentration than soybean. Pod N
concentration was higher in soybean than mungbean (BRCI), but shoot N concentration
was comparable between species.
Soybean grown at MMSU derived 84 4% of the N accumulated in its plant biomass
from BNF, compared to mdigofera with 71.8%.
Decomposition
The manner of GM application (incorporation vs mulch) strongly affected
indigofera decomposition in early decomposition stages (0-40 d), whereas soybean
biomass breakdown when mulched or incorporated differed m later decomposition
stages (i.e after 40 d) Incorporated indigofera decomposed fastest (Table 4), loosing
about 50 % of its biomass within the first 20 d. When mulched, an mdigofera biomass
loss of 30 % occurred between 20 and 60 d, after which no further biomass loss took
place. Soybean lost 30% of its biomass within the first 40 d. A further 30 % of
incorporated soybean biomass was lost in the last 80 d, but no further decomposition
occurred when soybean was mulched. Comparing decomposition rates (Table 4)
incorporated mdigofera decomposed two to four times as fast as did the other treatments
88
Table 4. Decomposition rates, k, for a penod of 113 days, when soybean and
mdigotera 74 d old material was used as green manure were incorporated or left as
surface mulch Values shown were calculated using the single exponential model for
decomposition (Wieder & Lang, 1982)
Green manure treatments k R2
Soybean incorporation 0 0099 0 96***
Soybean mulch 0 0065 0 64*
Indigofera incorporation 0 0259 0 67 *
Indigofeia mulch 0 0059 0 52*** *, significant at 0 001 and 0 05 level respectively
Soil moisture and N release
Frequent irrigation eliminated treatment effects (l e changes in soil moisture due to
organic matter addition) of GM management on soil moisture dunng the tomato growing
penod at MMSU An optimal water supply for tomato plants was maintained with soil
matnc potentials ranging between -0 02 to 0 06 MPa Daily rainfall lead to soil moisture
contents near field capacity dunng the tomato growing penod at the BRCI site
Nitrate was the dominant form of inorganic N in the soil aftei legume application at
MMSU and BRCI (Fig 1) Soil ammonium contents remained low (± 5 kg NH4-N ha l
at MMSU, and ± 20 kg NH4-N ha 1at BRCI) and were comparable to those of the
control (data not presented) A significant increase of 5 kg NH4-N ha ' at MMSU and 30
kg NH4-N ha * at BRCI compared to the control occurred nght alter GM application but
decreased again within the first 10 days after incorporation (DAI) No differences in
ammonium contents occurred between planted and unplanted plots at MMSU
A fast and significant increase of NO3 content compared to control occurred at a
soil depth of 0 to 30 cm with soybean GM 10 DAI at MMSU Higher decomposition
rates with incorporated GM resulted in slightly greater soil NO3 contents compared to that
in mulched GM Nitrate release peaks occuned between 5 to 8 wk after GM application
with indigofera and 8 wk with soybean Nitrate contents dropped in all GM treatments
alter 8 weeks The fast increase in soil nitrate in control treatments right aftei GM
application, although no GM was applied, was a carry over effect due to soil rototilhng
before tomato seedlings were transplanted
The heterogeneity of the soil at the BRCI site lead to strong vanabihty in inorganic
N in soil within treatments Nitrate contents in soil increased nght after GM application,
with an increase in NO3 contents from 30 to 100 kg NO3-N ha * within 14 DAI Nitrate
contents peaked between 14 and 42 DAI in GM treatments Nitrate contents in GM
89
amended plots were only slightly increased compared to control within the first 40 DAI
The application of 30, 60, and 120 kg N ha 1 to tomato increased soil nitrate by 20 to 120
kg NO3-N ha J compared to the control between 7 and 42 DAI (data not presented)
Thereafter nitrate contents decreased to levels comparable to those of control Nitrate
contents in GM and fertilized plots increased sigmficantly compared to control from 84 to
110 DAI Nitrate released with soybean GM was comparable to that released with
mungbean GM Nitrogen release with legumes including pods was comparable to that
with legume residues only (data not presented)
90
A3
oz
120BRCI A
r100 ^"""rA N
80 ^V"^ Hi60
i aq
40 ]
20 j
J// sA
0 , > ' ,
Jl,
0 20 40 60 80 100 120
Days after green manure application
Control
— -•— - Indigofera incorporated-°— Indigofera mulch-*- ~ Soybean incorporated-&— Soybean mulch
"-" Mungbean incorporated-d— Mungbean mulch
Fig. 1. Nitrate contents in soil (0-30 cm) after green manure (soybean, indigofera at
MMSU, and soybean, mungbean at BRCI) application, 1994/95, Philippmes Error bars
shown indicate least significant difference at 0 05 level, and * at 0 1 level
91
Nitrate contents in early transplanted tomato plots at MMSU mainly stayed lower
than in late transplanted plots, indicating stronger tomato N -uptake in early transplanted
tomatoes (data not presented) Differences in soil nitrate contents between planted and
unplanted plots occurred about 8 wk after GM application, indicating strong nitrate
uptake by tomato plants from then on In late transplanted tomato plots NO3 contents
stayed greatei than unplanted until 8 weeks, indicating active and strong N-uptake
especially after 90 DAI
No significant differences in soil NO3 content occurred between GM and control
plots at 30 - 60 cm soil depth at MMSU (data not shown) There was a strong tendency
for NO3 contents m GM treatments to be lower than in the control before GM application
and higher than in the control at the end of the experiment At tomato harvest significantly
lower soil NO3 contents were found in planted compared to unplanted plots Green
manure application did not affect NH4 contents at 30 - 60 cm soil depth
Tomato yield, N uptake, and recovery
The application of soybean GM doubled tomato fruit yields compared to those of
the unfertilized control at MMSU (Table 5) Tomato biomass yield obtained in the
soybean incorporation treatment compared favourably with that of fertilizer treatment of
38 kg N ha-1 Tomato fruit and biomass yields were not affected by indigofera GM GM
management (mulch vs incorporation) had no effect on tomato fruit and biomass yield
Soybean incorporation significantly increased tomato fruit and biomass N uptake above
those of the unfertilized control Early transplanting significantly increased tomato truit
and biomass yields and N uptake
At BRCI there were no differences between treatments foi tomato fruit yields, plant
biomass, and tomato N uptake (Table 5) Greater nitrate sap contents (1000 -1472 mg
NO31 * plant sap) were found in 30, 60, and 120 kg N ha l fertilizer treatments in early
stages (42 DAI) compared to ±600 ml NO3 H plant sap in control and legume treatments
(data not presented) Thereafter nitrate sap contents decreased gradually in all treatments
and reached ±200 mg NO3 H 63 DAI Nitrate sap contents dropped further in all
treatments ranging between 9 - 100 mg NO3 H plant sap from 70 DAI until final tomato
harvest
CO
15
29
38
02
29
05)
LSD(P<0
442
216
226
14
824
tran
planun
gsecond
654
021
633
71
935
transplanting
firs
t
Subtreatments
ns
116
ns
17
ns
911
ns
06
ns
910
05)
(P<0
LSD
261
312
948
08
443
mulch
residue
Mungbean
359
113
146
08
840
incorparuon
residue
Mungbean
766
012
754
01
249
mulch
Mungbean
954
113
841
01
442
incorporation
Mungbean
617
94
82
08
510
mulch
Indigo
719
29
710
09
123
inco
rpor
ation
Indigo
457
512
944
09
548
mulch
residue
Soybean
660
711
948
09
546
lncorpartion
residue
Soybean
564
733
013
015
451
420
01
31
546
424
mulch
Soybean
062
241
211
119
850
122
08
71
546
225
incorporation
Soybean
Legumes
271
0124
511
342
759
781
08
27
150
670
150/120
274
075
713
526
560
648
09
22
747
647
75/60
876
353
811
116
165
137
08
61
043
739
38/30
766
721
120
311
754
104
09
09
547
612
(control)
0
(kgNha-1)
MMSU/BRCISulfate
Ammonium
treatments:
Maui
BRCI
MMSU
BRCI
MMSU
BRCI
MMSU
BRCI
MMSU
BRCI
MMSU
total
plan
tfruit
(dry
)plant
(fre
sh)
fruit
!)ha
(kg
')(tha
upta
keN
yield
Tomato
rephcates
four
of
means
are
shown
Values
Phil
ippi
nes
1994/95,
BRCI,
and
MMSU
at
mungbean)
indigofera,
(sob
ean,
manure
green
legume
or
fertilizer
mineral
with
amended
when
N-uptake
and
yields
biomass
and
fruit
Tomato
5.
Table
93
Green manure J5N recovery in plants and soil
Thirty percent of the 15N applied was recovered in soybean and 0.8 % in indigofera
(Table 6). In soybean and indigofera most of the 15N applied was recovered in the above
ground parts.
Table 6.15N recovery in legumes of foliar applied 15N labeled urea at MMSU, 1994,Philippines. Values within parenthesis indicate standard deviation.
Plant part
Soybean
15N%
atom excess
15N
kg ha-1
%!5N
recovered
Indigofera
15N%
atom excess
15N
kg ha-1
% 15N
recovered
Shoot
Pods
Roots
1.289 (0.08)
0.663 (0.07)
0.372 (0.03)
0.262
0.636
0.011
8.7
21.2
0.4
0.313 (0.02)
0.281 (0.01)
0.021
0.001
0.744
0.034
Total 2.324 0.910 30.3 0.586 0.022 0.778
Table 7. 15N recovery in tomato fruit and biomass at MMSU, 1994/95, Philippines.
15N
(kg ha-1)
input output %
(legumes) (tomato) recovery
fruit plant total
Main treatments
Soybean incorporation 0.910 0.0541 0.0387 0.0817 8.9
Soybean mulch 0.910 0.0584 0.0291 0.0875 9.6
Indigofera incorporation 0.022 0.0013 0.0009 0.0022 10.0
Indigofera mulch 0.022 0.0023 0.0010 0.0033 15.0
LSD (P<0.05) 0.0143 0.0134 0.0263
Sub treatments
Early transplanting 0.466 0.0328 0.0206 0.0479 10.3
Late transplanting 0.466 0.0253 0.0142 0.0394 8.5
LSD (P<0.05) ns ns ns
94
Recovery of GM 15N in tomato was comparable between soybean and indigofera
treatments (Table 7), indicating that 8 5 to 15 % of legume N is taken up by the tomato
crop Fifty-nine to 70 % of the 15N taken up by tomato plants accumulated in tomato
fruits
Total soil C and N contents in the soybean treatment increased by about 5%
between the time of soybean incorporation and tomato harvest (Table 8) Although MHA-
C increased and CaHA-C and -N decieased from the first to the second sampling, the
effect of GM application on these parameters is not clear because these parameters
changed similarly in the control plot between samplings
Table 8. Organic C and N in total soil and in organic fractions (mobile humic acids
(MHA), calcium humates (CaHA)) immediately before (1) and 113 days (2) after greenmanure application in control and soybean incorporation plots Standard deviation of
laboratory replicates of oigamc C and N contents of total soil are given in brackets,1994/95, MMSU, Philippines
Total soil Organic matter tractions
MHA CaHA
C N C N C N
g kg 1 soil
Control 1 7 11 (0 01)0 665(0 01) 0 162 0 0163 0 448 0 0409
2 6 85 (0 06) 0 669 (0 02) 0 212 0 0204 0 228 0 0243
Soybean 1 7 04 (0 03) 0 707 (0 01) 0 218 0 0227 0 369 0 0346
2 7 39 (0 04) 0 750 (0 02) 0 240 0 0231 0 226 0 0248
The MHA and CaHA did not seem to be more active in short-term N cycling than
the bulk SOM, as they contained only 4 5% total of the total soil 15N in the soybean plot
at tomato harvest (Table 9) Most of the ,5N was recovered in the humin (unextracted
organic matter) Moreover, the ratios of 15N to total N were similar tor the MHA and
CaHA as for the bulk soil, further suggesting that preferential accumulation of recently
added 15N did not occur in the extracted MHA and CaHA The MHA and CaHA had
comparable amounts of 15N m the soybean plots at harvest
15Nitiogen in total soil was not fully recovered in MHA, CaHA and humm, which
may be due to losses of 15N dunng extraction as fulvic acids Estimations of N losses
95
after tomato harvest were greater when calculated with 15N than with total N (Table 10),
due to lower N recoveries of 15N in tomato and soil 15N values for whole soils, MHA,
and CaHA for all treatments except soybean at tomato harvest were too low to allow
accurate measurement
Table 9. 15N in total soil and in soil organic fractions (mobile humic acids (MHA),calcium humates (CaHA), and humin) immediately before (1) and 113 days (2) after
green manure application in control and soybean incorporation plots at MMSU, 1994/95,
Philippines
Total soil MHA CaHA Humin
(g kg_1 soil * 1,000,000)
Control 1 11 3 0 27 0 57 3 06
2 87 051 0 36 5 21
Soybean 1 92 0 47 0 48 3 53
2 196 5 471 4 17 118 20
At tomato harvest estimations of N losses were greater in the N balance calculated
with 15N than with total N (Table 10), due to lower N recoveries of 15N in tomato and
soil
Table 10. Comparison of total N and ,5N balance atter tomato harvest in soybeanincorporation plots at MMSU, 1995, Philippines
Total N 15N
kg N ha"1 % recovery kg N ha ! % recovery
Input soybean 1193 100 0 0 910 100 0
Output tomato 19 5 t 163 0 082 89
left soil 64 0 53 7 0315 34 6
lost (?) 35 8 30 0 0513 56 5
t calculated by subtracting tomato N in soybean incorporation minus tomato N in control
96
DISCUSSION
Legumes
Soybean biomass and nitrogen yields at MMSU and BRCI compared favourably
with yields obtained in Taiwan (Chapter 2), and in Texas where soybean was grown for
hay production at high seeding densities (Munoz et al, 1983) Mungbean yields were
inferior to those reported by Meelu and Morns (1998), who obtained mungbean yields
comparable to those of soybean in this study Indigofera yields however were about 20
times lower than reported yields (Chapter 2, Batilan et al, 1989) The heterogeneous
seed quality of the indigenous indigofera seed used and a strong rainfall one week after
sowing followed by soil crusting were among the mam reasons for the poor performance
ot indigofera, which is the main green manure crop used m this area of the Philippines
(Garnty and Flinn, 1988)
Compared to the same legumes grown in Taiwan polyphenol and N concentrations
of soybean and indigofera in the current trials were mostly lower, but C/N and lignin
tended to be higher Polyphenols substances in plant material are associated with low
amounts of N and P in the soil (Davies at al, 1964) which may explain the contrasting
results between sites if viewed in the light of N and P contents of MMSU soil and
AVRDC soil (chapter 3, 4) Soybean had a higher and Indigofera a lower tannm
concentration at MMSU than when grown in Taiwan
Estimates of the contribution of soybean N fixation to total N content in our study
at MMSU was consistent with those ranging between 66 to 97% for soybean grown in
the tropics (George et al 1988, George et al, 1992, Eaglesham et al, 1981, Boddey et
al, 1984)
Decomposition
Decomposition rates from the litter bag study in this experiment were lower than
those obtained in Taiwan (Chapter 3, 4) The slow decomposition ot incorporated
soybean compared to indigoteia can hardly be explained by differences in plant chemical
composition as they were almost the same across species The physical consistency of
soybean pods containing full size yellow beans and hardy pods and stems (R6 to R7,
Fehr et al, 1971), compared to the small and easily decomposable leaves of indigofera
may have been one of the main reasons for this strong difference in decomposition rates
In experiments in Taiwan (Chapter 4) sixty-days-old soybean (R5 to R6) decomposed at
tates comparable to those of indigofera in the current study Many investigators have
observed that organic residues decompose more slowly in soils with higher clay contents,
especially clays having higher exchange capacities (Lynch and Cotnoir, 1956, Sorensen,
1975) Microbial activity is controlled by soil physical conditions such as space,
97
temperature and oxygen, chemical conditions such as substrate availabihty and predatory
or antagonistic organisms (Grant et al, 1993) Reduced soil aeration/ oxygen in the
clayey MMSU soil compared to the loamy soil at AVRDC may have further contributed
to a slower legume residue decomposition rate
Inorganic N
Greater amounts of NO3 were released with soybean compared to indigofera GM,
although decomposition rates of soybean were relatively low at MMSU This could
mainly be due to greater biomass and N applied with soybean, as N-release was not
proportionally greater than that of indigofera Nitrogen release peaks of soybean and
indigofera occurred about a month later than in field experiments at BRCI and in Taiwan
(Chapter 3, 4) N release peaks reached 80 -100 kg NO3 ha-1 in all four experiments,
although amounts of N applied with soybean GM varied between 93 and 182 kg N ha-1
Results of the incubation study comparing N-release after addition of organic residues in
different soils (Chapter 2) suggested certain soil chemical and physical properties as
strong factors retarding N release in MMSU compared with BRCI and AVRDC soil N
was released faster in the field experiment compared to the incubation study which was
partly due to higher rates of GM applied and the application of fresh versus dry plant
material The decline of soil NO3 content after 8 weeks could mainly be attributed to
tomato N-uptake Reduction of microbial activity and microbial N-immobihzatton after
consumption of the labile fractions of the residue m early decomposition stages may have
occurred due to the recalcitrant consistency of the residue left in the latter stages of
decomposition The decrease of N and C content of the residue after 8 weeks matches
this assumption
The fast early N-ielease after GM application in BRCI soil matched findings in
incubation studies with a soil of comparable origin (Chapter 2) Nitrogen release patterns
of incorporated GM at BRCI were comparable to those found in wet and dry season
experiments at AVRDC (Chapter 3, 4), although soil types were quite diffeient
Increased N mineralization at the end of the cropping cycle was also found in the wet
season experiment at AVRDC (Chapter 3,4), suggesting a greater N reminerahzation of
immobilized legume N in the wet season It is probable that liming of the soil and the
addition of chicken dung lead to a strong soil N mineralization in BRCI soil Decay of
plant residues and SOM were accelerated by liming of acid soils (Alexander, 1977, Singh
and Beauchamp, 1986)
Tomato yield
Greatest tomato yields of 65 to 70 t ha-1 were reached with 150 kg N ha-1 at
MMSU and 120 kg N ha"1 AVRDC in the dry season (Chapter 4) The effect of lower
98
rates of N applied, as well as ot GM management, on tomato yield differed between
experiments at MMSU and AVRDC While the addition of 38 kg N ha-1 doubled tomato
yields compared to the control at MMSU, 30 kg N ha-1 increased yields only by 13% at
AVRDC (Chapter 4) Strong basal N mineralization in AVRDC soil resulted in a tomato
yield of 40 t ha-1 after two month of fallow, while only 12 6 t ha-1 were reached at
MMSU The efficiency of GM-N utilization, or fertihzer-N use efficiency, depends on
crop N demand (Appel, 1994), the ability of soils to supply N by mineralization of
organic N (Campbell et al, 1981), and the growth and climatic conditions for the
subsequent crop The enhanced soil N mineralization due to hmmg and the application of
chicken dung at BRCI must have met tomato N demand to the extent that tomato yields
did not respond to GM oi fertihzei amendments anymore Tomato yields obtained at
BRCI were above average yields of 301 ha * obtained in this area
The congruence of N-release kinetics from GM with the N-uptake dynamics of the
subsequent crop is one of the key topics ot GM management Comparable tomato N
uptake in all treatments at BRCI indicate that the tomato crop was not able to absorb
abundant soil nitrate in early stages in fertilizer amended plots At MMSU a greater
proportion of N mineralized from decomposing GM appears to coincide with tomato
plant N demand ot early transplanted tomatoes, as higher yields and N-uptake were
achieved The apparent N supplying capacity of the GM amendment declined aftei 8
weeks which we assume was detrimental to maximum tomato plant growth and yield
development at MMSU In order to achieve an optimal tomato plant nutrition using GM a
combined approach of organic plus mineral N fertilizer could be most promising at
MMSU Mineral N fertilizer (30 to 60 kg N ha-1) could be applied to tomato plants
starting 8 weeks after GM application
*SN recovery in plant
Low indigofera shoot biomass lead to low recovenes of 15N labeled urea applied
15Nitrogen recoveries in both legume species in our study were lower compared to
studies of Zebarth et al (1991) and Vasilas et al (1980), where 30% and 57% were
recovered by alfalfa and red clover, and 44 - 67% by soybean, respectively Higher
labeled urea and greater quantities of labeled N fertilizer applied in their studies may have
lead to higher 15N recoveries compared to the present study The distribution of 15N
enrichment in soybean was comparable with results described by Vasilas et al (1980),
where highest enrichment was found in the seed and negligible amounts in the roots
Higher 15N enrichment (50 to 71 %) in tomato fruit is in line with results of Ladd et
al (1981) who found higher enrichment in repioductive plant parts 15N lecovery
obtained in tomato plants is within the leported range of 15N recoveries (7 to 25%) by
crops grown subsequently to the application of 15N labeled legume residues (Valhs,
99
1993, Yaacob and Blair, 1980, Norman et al, 1990, Muller et al, 1988, Harris and
Hestermann, 1990) Recoveries of apphed 15N in the subsequent crop and soil were high
(Ladd et al 1981, Muller and Sundman, 1988), giving evidence that the ability of the soil
to retain plant-derived N is strong in comparison with the ability of the subsequent crops
and different loss mechanisms to remove it (Muller and Sundman, 1988) Hams et al
(1994) recovered 19% of the applied legume N in microbial biomass, 38% of legume N
applied in non-biomass organic fractions, and only small amounts (<5%) of legume N
were recovered in the inorganic fraction Seligman et al (1986) suggested that some of
the added organic 15N was incorporated into stable soil organic N pools to be mineralized
at a rate approaching that of the stable soil N fraction
1SN recovery in soil fractions
The similarities of the total 15N/ total N ratios for MHA and CaHA compared to
humin suggest that the two fractions were no more labile than the rest of the SOM m this
soil Our results are comparable to those of He et al (1988), in that a significant
proportion of recently added 15N in the soil was not extractable (humin) Humin can be
very young and much of it is composed of alkyl compounds and carbohydrates as
microbial byproducts (He et al, 1988) Domination of soil C and N by the humin may be
especially clear in a soil where conditions are favourable for rapid degradation The rapid
decomposition of the soybean residues and the small quantities of MHA and CaHA
extracted from the MMSU soil in relation to other rice soils of the Philippines (Oik et al,
1996) demonstrate the favourable conditions for degiadation in this soil Organic
molecules resulting from microbial degradation such as microbial tissues will be
preserved in the soil only if they are stabilized, thereby protected from further
degradation One such form of protection is chemical binding to the mineral surface ol
such strength that the organic material is not extractable and hence considered as humin
The extremely high Ca levels in MMSU soil may contribute to the humin constituting a
high proportion of total SOM
N balance
Higher mineralization rates of labeled compared to unlabeled organic residues
(Amato and Ladd, 1980, Chichester et al, 1975) may have contributed to lower rates of
N recovery in 15N compared to total N balances Apparent N recovery in tomato
calculated by the difference of tomato N in legume and control plots may oveiestimate
legume N contnbution, as N mineralization of native soil may have been enhanced with
the application of GM, which was not the case in control Higher 15N loss (57%) than
total N (30%) may reflect volatilization and denitnfication at the beginning of the crop
cycle, as well as the low plant uptake at that time Total N loss would be lower on a
100
percent basis during this time because of the low basal N mineralization of SOM-N As
soybean quickly decomposed, transformations of GM-15N would be disproportionately
determined by soil conditions early in the crop cycle 15N was then either taken up or it
moved into relatively labile SOM fractions which were then largely mineralized quickly
This scenano could be based on the low total soil C and N levels, small quantities of
extracted MHA and CaHA, high losses of 15N from the system, and the greater relative
loss 15N than total N Given its low total C and N contents, the MMSU soil may not
have a large capacity to store added N, whether the N is added in organic or inorganic
forms We could conclude that lack of synchronization between N supply and demand
caused the single application of organic GM fertilizer to be less successful than split
applications of inorganic N
CONCLUSIONS
We can conclude that only 10 -15% legume N can be taken up by the tomato crop,
and that one third ot the legume N remains in the soil in relatively stable organic matter
ft actions Nitrogen release occurs slightly faster in the clayey (BRCI) than the tine silty
soil (MMSU) The tomato yield response to applied N (from GM or fertilizer source)
varied strongly among sites, and depended on soil N mineralization (l e availability)
REFERENCES
Aber, JD,JM Mehllo, and C A McClauherty 1990 Predicting long-term patterns of
mass loss, nitrogen dynamics, and soil organic mattei formation from initial fine
litter chemistry in temperate foi est ecosystems Can J Bot 68 2201 2208
Alexander, D 1977 Mineralization and immobilization of nitrogen In Alexander, D
(ed ) Introduction to soil microbiology pp 225-250 Wiley, New York
Amato, M,and J N Ladd 1980 Studies of nitrogen immobilization and mineralization
in calcareous soils- V Foimation and distribution of isotope-labeled biomass
during decomposition ot 14C and 15N labeled plant material Soil Biol Biochem
12 405-411
Appel, T 1994 Relevance of soil N mineralization, total N demand of crops and
efficiency of applied N tor fertilizer recommendations for ceieals- Theory and
application Z Pflanzenernahr Bodenk 157 407-414
101
Batilan, R.T., C.C. Batilan, and D.P. Garrity. 1989. Indigofera tinctoria L. as a green
manure crop in rainfed rice-based cropping systems. IRRI Saturday seminar July 1,
1989.
Boddey, R.M., P.M. Chalk, R.L. Victoria, and E. Matsui. 1984. Nitrogen fixation by
nodulated soybean under tropical field conditions estimated by the 15N isotope
technique. Soil Biol. Biochem. 16(6): 583-588.
Bremner, J.M. 1965. Total Nitrogen. In Black, C. A. (ed.) Methods in soil analysis.
Part 2. agronomy 9:1149-1178. American Society of Agronomy, Madison.
Campbell, C.A., R. J.K. Myers, and K.L. Weier. 1981. Potentially mineralizable
nitrogen, decomposition rates and their relationship to temperature for five
Queensland soils. Aust. J. Soil Res. 19: 323-332.
Chichester, F.W., J.O. Legg, and G. Stanford. 1975. Relative mineralization rates of
indigenous and recently incorporated 15N-labeled nitrogen. Soil Science, 120: 455-
460.
Davies, R.L, C.B. Coulson, and D.A. Lewis. 1964. Polyphenols in plant, humus and
soil. IV Factors leading to increase in biosynthesis of polyphenol in leaves and their
relationship to mull and mor formation. J. Soil Sci. 15(2): 310-317.
Drews, M., and S. Fischer. 1989. Methode zur Kontrolle der N- und K-Versorgung von
Gewachshausgurke und Gewachshaustomate iiber die Pressaftanalyse der
Blattstiele. Gartenbau 36:42-44.
Drews, M, and S. Fischer. 1992. Richtwerte fur Nitrat und Kalium zur Pressaftanalyse
bei Gewachshaustomate und -gurke. Gartenbauwissenschaft 57(3): 145-150.
Eaglesham, A.R.J., Ayawabaa, V. Rangarao, D.L. Eskew. 1981. Improving the
nitrogen nutrition of maize by intercropping with cowpea. Soil Biol. Biochem. 13:
169-171.
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 Science 11:
929-931.
Fraser, D.G., J.W. Doran, J.W. Sahs, and G.W. Lesoing. 1988. Soil microbial
populations and activities under conventional and organic management. J. Environ.
Quality 17(4): 585-590.
Garrity D.P. and J.C Flinn. 1988. Farm-level management systems for green manure
crops in Asian rice environments. In IRRI (ed.) Sustainable Agriculture: Green
manure in rice farming, p 111-130. International Rice Research Institute, Los
Banos, Philippines.
George, T., P.W. Singleton, and B.B. Bohlool. 1988. Yield, soil nitrogen uptake and
nitrogen fixation by soybean from four maturity groups grown at three elevations.
Agron. J. 80: 563-567.
102
George, T, JK Ladha, R J Buresh, and D P Garnty 1992 Managing native and
legume-fixed nitrogen in lowland rice-based cropping systems Plant Soil 141 69-
91
George, T ,J K Ladha, R J Buresh, and D P Garnty 1993 Nitrate dynamics during
die aerobic soil phase in lowland rice-based cropping systems Soil Sci Soc Am
J 57(6) 1526-1532
Grant, R F,N G Juma, and W B McGill 1993 Simulation of carbon and nitrogen
transformations in soil mineralization Soil Biol Biochem 25(10) 1317-1329
Harris, G H ,and O B Hesterman 1990 Quantifying the nitrogen contribution from
alfalfa to soil and two succeeding crops Agron J 82 129-134
Harris, G H ,O B Hesterman, E A Paul, S E Peters, and R R Janke 1994 Fate of
legume and fertilizer nitrogen 15 in a long-term cropping systems experiment
Agron I 86 910-915
He, X T,F J Stevenson, R L Mulvaney, and K R Kelly 1988 Incorporation of
newly immobilized 15N into stable organic forms in soil Soil Biol Biochem
20(1) 75-81
Jenny, H ,S P Gessel, and F T Bingham 1949 Comparative study of decomposition
rates of organic matter in temperate and tropical regions Soil Sci 68 419-432
Ladd, JN ,JM Oades, and M Amato 1981 Microbial biomass formed from 14C,
15N-labeled plant material decomposing in soils in the field Soil Biol Biochem
13 119-126
Lynch, D L, and L J Cotnoir Jr 1956 The influence of clay minerals on the
breakdown of certain organic substrates Soil Biol Biochem 20 367-370
Meelu, O P,and R A Moms 1988 Green manure management in rice-based cropping
systems In IRRI (ed) Green manure in nee farming p 202-222 IRRI, Los
Baiios, Philippines
Muller, M M ,and V Sundman 1988 The fate of nitrogen (15N) released from different
plant materials dunng decomposition under field conditions Plant Soil 105 133-
139
Munoz, A E, E C Holt, and R W Weaver 1983 Yield and quality of soybean hay as
influenced by stage growth and plant density Agron J 75 147-149
Norman, R J,
J T Gilmour, and B R Wells 1990 Mineralization of nitrogen from
nitrogen-15 labeled crop residues and utilization by rice Soil Sci Soc Am J 54
1351-1356
Oik, D C,K G Cassman, and T W M Fan 1995 Characterization of two humic acid
fractions from a calcerous vermicuhtic soil implications for the humification
process Geoderma 65 195-208
103
Oik, D.C., K.G. Cassman, E.W. Randall, P. Kinchesh, L.J. Sanger, and J.M.
Anderson. 1996. Changes in chemical properties of soil organic matter with
intensified rice cropping in tropical lowland soils. Eur. J. Soil Sci: (in press).
Olson, J.S. 1963. Energy storage and the balance of producers and decomposers in
ecological systems. Ecology, 44: 322-331.
SAS Institute, I. 1991. SAS user's guide. SAS Institute. Sparks Press, Raleigh, North
Carolina, USA.
SAS Institute, Inc. 1989. JMP user's guide. Version 2. SAS Institute Inc., Cary, N.C.
Seligman, N.G., S. Feigenbaum, P. Feinerman, and R.W. Benjamin. 1986. Uptake of
nitrogen from a high C-to-N ration 15N-labeled organic residues by spring wheat
grown under semi-arid conditions. Soil Biol. Biochem. 18: 303-307.
Siegel, R.S. 1980. Determination of nitrate and exchangeable ammonium in soil extracts
by an ammonia electrode. Soil Sci. Soc. Am. J. 44: 943-947.
Singh Y., and E.G. Beauchamp. 1986. Nitrogen mineralization and nitrifier activity in
limed and urea treated soils. Commun. Soil Sci. Plant Analysis 7:1369-1381.
Snedecor, G.W., and W.G. Cochran. 1978. Statistical methods. Sixth edition. Iowa
State University Press. Ames, Iowa, USA.
Sorensen, L.H. 1975. The influence of clay on the rate of decay of amino acid
metabolites synthesized in soil during the decomposition of cellulose. Z.
Pflanzernahr. Bodenk. 7: 171-177.
Talbott, H.J., W.J. Kenworthy, J.O. Legg, and L.W. Douglas. 1985. Soil-nitrogen
accumulation in nodulated and non-nodulated soybeans: A verification of the
differences method by 15N technique. Field Crops Research 11: 55-67.
Tripathi, B.P. 1995. Dynamics of soil and nitrogen in selected rice-based cropping
systems in Northern Luzon, Philippines. Ph.D. diss. Univ. Philippines, Los
Bafios.
Vallis, I. 1983. Uptake by grass and transfer to soil ofnitrogen from 15N-labeled legume
materials applied to a rhodesgrass pasture. Aust. J. Agric. Res. 34: 367-376.
Vasilas, B.L., J.D. Legg, and D.C. Wolf. 1980. Foliar fertilization of soybeans:
Absorption and transformation of 15N-labeled urea. Agron. J. 72: 271-275.
Wieder, R.K., and G.E. Lang. 1982. A critique of the analytical methods used in
examining decomposition data obtained from litter bags. Ecology. 63(6): 1636-
1642.
Yaacob, O., and G.J. Blair. 1980. Mineralization of 15N-labeled legume residues in soils
with different nitrogen contents and its uptake by Rhodes grass. Plant Soil. 57:
237-248.
104
Zebarth, BJ., V. Alder, and R.W. Sheard. 1991. In situ labeling of legume residueswith a foliar application of 15N-enriched urea solution. Commun. Soil Sci. Plant
Anal. 22: 437-447.
105
6
General Discussion
Nitrogen fertilizer substitution with legume green manures
Case study in Taiwan
Increasing problems due to environmental pollution in Taiwan, and its impact on
decreasing soil fertility, have pushed the awareness of environmental protection to
political actions The Taiwanese government has started to promote the introduction of
legume green manures (GM) as rotational crops in intensive cropping systems to reduce
fertihzer use and as a measure to maintain soil fertility and diversify cropping patterns
Fertihzer overuse was caused by low fertihzer pnces and the high profitability of
horticultural production Benefits related to the use of GM such as N fertilizer
substitution, short term enhancement of soil N fertility, and the maintenance of the soil
organic matter have led to a renewed interest in the old practice of green manuring in
various cropping systems
Major production constraints for tomatoes grown during the rainy tropical summer
(wet season, WS) in the lowlands of Taiwan are high day and night temperatures
affecting fruit setting, temporary flooding of the fields due to tropical storms (typhoons),
and bacterial, fungal and viral diseases Depending on typhoon mcidence and planting
date, tomato yields in tropical lowlands range between 0 to 15 t ha 1 Although yields are
low, high market prices for tomatoes encourage off-season production (WS) Production
m this season is often undertaken in highland areas, where temperatures are more
favourable Because of lower temperatures tomato yields are greater in the highlands, but
transportation to markets becomes a major issue, as roads are also affected by storms
This explains major efforts in breeding for heat and flooding tolerance of tomato, but also
in crop management research in tropical lowlands For example, tomato yields were
strongly enhanced when planted on raised beds, which has shown to be an effective tool
to overcome seasonal stresses such as flooding
Field experiments at the AVRDC in Southern Taiwan showed that N fertilizer
substitution for field grown tomatoes with soybean GM was possible in the WS 1993
(Chapter 3) Strong N immobilization and asynchrony between GM-N release and crop N
demand led to lower tomato and cabbage yields in GM plots compared to those grown
after a two months fallow in the dry season (DS) 1993/94 (Chapter 4)
106
Case studies in the Philippines
Agricultural soils in the lowlands of the Docos provinces in Northern Philippines
are cropped to cash crops such as tobacco, cotton and vegetables (garlic, onion, tomato,
pepper) in the DS, while nee is grown as a subsistence crop in the WS Relatively poor
soils and the profitability of vegetable production have led to the intensification of the
cropping cycle The application of high fertilizer doses is threatening ground water
quality Many small-holder farmers in the highlands of Bukidnon in Southern Philippines
grow subsistence crops, scavenge tor firewood to sell, or work as field laborers on
pineapple, sugar cane, or tomato plantations of multinational companies (e g Del Monte)
Both areas belong to the major tomato production sites of the Philippines and two tomato
paste companies founded by private, national and development agencies have been
established recently in Ilocos (1984, Northern Food Corporation) and Bukidnon (1993,
Bukidnon Resources Company Inc ) These companies contract small-holder farmers to
grow tomatoes The farmers are organized into cooperatives that borrow capital from the
bank to finance facilities for the enhancement of tomato production Investments in
irrigation facilities, fertilizers, and pesticides to meet tomato fruit quality and quantity
expectations are very high compared to farmers income Therefore,m both areas, a great
interest exists to use legume GM to reduce fertilizer costs and maintain soil fertihty over
time
In our case study in Docos Norte tomato yields were doubled compared to control
with soybean GM in Ilocos (Chapter 5) In contrast no tomato yield increase was
obtained with GM-N or fertilizer N applied in combination with lime and chicken dung in
the Bukidnon case study (Chapter 5)
Conclusion
Tomato yields responded to GM-N when soil N mineralization was poor (Ilocos),
or when high leaching losses inhibited a previous accumulation of soil N (Taiwan, WS)
When soil N was high (Taiwan, DS) and was enhanced by liming of acid soils
(Bukidnon, WS) tomato yield response was low to GM-N Our results support findings
that the efficiency of GM-N utilization, or fertihzer-N use efficiency, depends on crop N
demand (Appel, 1994), the ability of soils to supply N by mineralization of organic N
(Campbell et al, 1981), and the growth and climatic conditions for the subsequent crop
Tomato N uptake
The feasibility of meeting N needs of a crop with GM-N or fertilizer N depends
strongly on the N uptake pattern and the N uptake efficiency of a crop Maize grown for
one month in the WS in Taiwan (Chapter 3) accumulated more N than the tomato crop in
107
2 months. The root system of field grown tomato plants was relatively small. Substantial
amounts of the applied N may not be absorbed as tomato roots do not appear to
proliferate in soil with higher mineral N content (Jackson and Bloom, 1990). Tomato
fruit and biomass response to applied N (GM or fertilizer N) was high on poor soil
(Chapter 5), or on soil where strong N leaching inhibited previous N accumulation
(Chapter 3). In other soils tomato N needs were partly met by soil N mineralization. The
application of further 10 - 50 kg N ha"1 fertilizer N was marginal, compared to 120 kg N
ha-1 (Chapter 4). Tomato fruit and plants responded similarly to applied N (Chapters 3,
4,5).
Green manure N versus fertilizer N
Unlike inorganic N fertilizer, GM must undergo decomposition before N becomes
available. Because the release of N from organic sources is so closely tied to complex
cycling of C and N, the availability and effects of legume N are more difficult to predict
than for chemical fertilizer (Groffmann et al., 1987). Incorporation ofGM as compared to
surface mulch enhanced decomposition and N release in all three soils tested (Chapter 3,
4, 5). Nitrogen mineralization in two of three soils amended with GM commenced after a
lag period (N immobilization) of 1 - 2 weeks in incubation and field studies (Chapter 2,
3, 4, 5). Soil chemical properties such as high pH, low P concentration and high clay
content may have been major factors delaying N mineralization in the clayey,
isohyperthermic Fluvaquentic Ustropept soil (Chapter 2,5).
While N fertilizer applications are often split into basal and side dressings, the GM
is applied all at once. Nitrogen release of GM should meet N uptake pattern of the crop in
order to achieve optimal plant nutrition. If this is not the case for the whole duration of a
crop, a combined approach of organic (GM) plus mineral N fertilizer could be more
promising.
Legumes for green manure use
Recent attempts to evaluate the usefulness of GM in the context of agricultural
sustainability have been hindered by a lack of information on nutrient release patterns
(Singh et al., 1992). The ability of legumes to accumulate large amounts of N in a short
duration is desirable due to shortage of land and time in intensively cropped systems. Of
four legume species tested (Chapter 2) soybean was found to be the most promising GM
species accumulating a minimum of 3 t biomass and 100 kg N ha-1 in 60 days at all sites
(Chapter 3,4,5). The critical harvest time of legumes for GM use is gauged by weighing
the advantages of a higher biomass and N accumulation, and the disadvantages of a
reduced N mineralization rate with increasing plant age. The extra growth of 90 vs. 60 d
legume material changed the N mineralization pattern of harvested materials drastically
108
from significant N release to no N mineralization during 10 wk incubation (Chapter 2)
The application of fresh GM in field studies compared to dry plant material in incubation
studies, explains net N release after GM application in the field compared to no N release
in the incubation experiment with 60 days-old plant material in the Ilocos
Initial plant N concentration and C/N ratio were the two major factors driving net N
mineralization of plant material m the soils from Taiwan and Bukidnon The relative
influence of these plant properties on N release changed with sampling time and soil type
(Chapter 2) No correlation between plant properties and N release was found in the
Docos soil (Chapter 2)
Residual N effect
Low fertilizer recovery and associated response to applied N makes tomatoes
relatively inefficient users of fertilizer N compared to maize The lower the tomato N
uptake (Chapter 3) the greater the residual effect on maize, grown after tomato (Chapter
4) The residual effect of the GM application on maize biomass and N accumulation was
comparable to that of fertilizer N application (Chapter 3,4) While 9 - 15% legume N was
recovered in the tomato crop m Ilocos (Chapter 5), 30 to 50 % legume N remamed in the
soil in relatively stable fractions of soil organic matter This suggests that although labile
fractions of the GM are decomposed quickly, and N release occurs soon after GM
application, fairly quick reminerahzation of N immobilized during the decomposition
process may increase short term N availability for a second crop as shown with maize
(Chapter 3, 4) under tropical conditions The recalcitrant fraction of the GM may
contribute to long term soil organic matter build up (Chapter 2)
Conclusions, perspectives and future research needs
In comparison with relevant literature two major results can be concluded from this
study
The importance of initial plant properties in determining N mineralization in
incubation studies and its extrapolation on field N mineralization is to my opinion actually
being overemphasized Varying factors such as i) range ot legumes tested, u) duration of
incubation, in) sampling schedule, iv) soil type, v) soil temperature and iv) and soil
moisture make it difficult to compare the relative importance of initial plant composition
on N release across studies
It seems important to me to differenuate between the fertilizing effect and the N
fertilizer substitution effect of a GM on a crop A high fertilizing effect was expressed by
tomato yields in Ilocos and AVRDC (WS, raised beds) which were doubled with GM
compared to control The N substitution effect estimated at 38 kg N ha-1 (Ilocos) and 30-
120 kg N ha-1 (AVRDC, WS) however seemed modest compared to N inputs (120 - 150
109
kg N ha-1) used in intensive tomato production. Both N fertilization and N substitution
with GM in field studies depend strongly on i) the subsequent crop and its ability to
absorb released N, ii) expectation of yield, iii) legume species used for GM, iv) legume
age and management, v) site and soil type, vi) season.
It would be interesting to optimize tomato yields with GM by N side dressings after
the decline in N release 8 weeks after GM application. Tomato response to GM could be
extended and compared to a wide range of vegetable crops such as fruit vegetables
(tomato, eggplant, pepper etc.); leafy vegetables (lettuce, spinach, cabbage); flower
vegetables (cauliflower, broccoli); and root, bulb and tuber crops (radish, carrot, potato,
onion, garlic etc.). The N response of these vegetables could be compared to crops with a
high N efficiency such as maize or wheat.
In order to avoid soil N accumulation during fallow, a planted fallow treatment
should be included in further experiments for the time that legumes are grown in legume
plots, thus the release of GM-N could be distinguished from accumulated soil N.
15N studies tracing legume N in soil organic matter fractions should be continued to
understand the results obtained at MMSU on legume N accumulation in stable organic
matter fractions.
REFERENCES
Appel, T. 1994. Relevance of soil N mineralization, total N demand of crops and
efficiency of applied N for fertilizer recommendations for cereals- Theory and
application. Z. Pflanzenernahr. Bodenk. 157; 407-414.
Campbell, C. A., R. J. K. Myers, and K. L. Weier. 1981. Potentially mineralizable
nitrogen, decomposition rates and their relationship to tfimperature for five
Queensland soils. Aust. J. Soil Res. 19: 323-332.
Groffman, P. M., D. A. Hendrix, and D. A. Crossley. 1987. Nitrogen dynamics in
conventional and no-tillage agroecosystems with inorganic or legume nitrogen
inputs. Plant Soil 97: 315-332.
Jackson, L. E., and A. J. Bloom. 1990. Root distribution in relation to soil nitrogen
availability in field-grown tomatoes. Plant Soil 128:115-126.
Singh Y., B. Singh, and C.S. Khind. 1992. Nutrient transformations in soils amended
with green manures. Adv. Agron. 20: 237-309.
110
7
Summary
Nitrogen contribution of leguminous green manures to succeeding crops depend on
their ability to accumulate high amounts of biomass and N in a short time, and their ability
to decompose at a rate which matches the N needs of the subsequent crop Factors
affecting legume biomass and N accumulation, such as seeding density, growth duration
and season were evaluated in a field expenment for Medicago sativa L,Desmodmm
intortum (Mill) Urb , Indigofera nnctona L and Glycine max (L ) Merr 60, 75, and
90 days after sowing (d) Nitrogen release in the soil was investigated in an aerobic
incubation expenment with three tropical soils (a silt loamy, mixed, hyperthermic
Fluvaquentic Entochrept, a clayey, kaohmtic, lsohyperthermic Uldsol, and a clayey,
mixed, lsohyperthermic Fluvaquentic Ustropept) from Taiwan and the Philippines
The feasibility of meeting N needs of vegetables with legume green manures (GM)
was tested m a 6 month expenmental cropping pattern in four field experiments two
expenments were performed at the Asian Vegetable Research and Development Center
(AVRDC) in Southern Taiwan one m the wet season (WS) and one in the dry season
(DS), and two expenments in the Philippines one in Ilocos Norte (MMSU) and one in
Bukidnon (BRCI) Two legume species, soybean (Glycine max L Merr), and
indigofera (Indigofera tinctona L) at AVRDC and MMSU, and soybean and mungbean
(Vigna r adiata (L) Wilczek) at BRCI were grown for 60 - 70 d and then either used as
mulch or incorporated into the soil A tomato (Lycopersicum esculentum Mill) crop was
transplanted immediately after GM application and grown to harvest Green manure
amended tomato yields were compared to inorganic fertilzer N treatments ranging from 0-
150 kg N ha 1 The residual effect of the fertilizing method on a crop following the
vegetable crop was determined with maize biomass and N uptake grown for 30 d at
AVRDC Legume and vegetable biomass, yield, and N uptake were studied on two
different bed systems (raised versus low beds) simultaneously at AVRDC Legume
decomposition was mvestigated in a litter bag study, and N release m soil was monitored
through frequent soil sampling at AVRDC and at MMSU Legume N recovery in tomato
and soil organic matter fractions (mobile humic acids, calcium humates) was traced with
!5N at MMSU
111
Seeding density, legume age, legume species, growing season and site were key
factors affecting legume biomass and N accumulation, and chemical composition. In 60 d
most legume biomass and N was accumulated with soybean. The greatest yield advantage
with double compared to normal seeding density was achieved with 60 d legume material.
From 60 and 90 d greatest biomass and N accumulation was accumulated with soybean at
75 d and indigofera at 90 d. For short term GM biomass and N accumulation alfalfa and
desmodium were unsuitable.
Soybean grown for 60 - 70 d accumulated a minimum of 3 t biomass ha-1 and 100
kg N ha-1 in all sites and seasons in Taiwan and the Philippines. A maximum of 6 t
biomass ha-1 and 200 kg N ha-1 was reached in the WS at AVRDC. Indigofera yields
were more variable and always inferior to soybean yields. Average yields were 1 t
biomass ha-1 and 40 kg N ha-1. Small seeds followed by slow emergence of indigofera
make it more vulnerable to variable soil conditions (soil compaction) and rainfall
incidence. These factors make it difficult to use /propagate indigofera as a short term (60 -
90 d) GM.
Extra legume growth from 60 to 90 d changed N release pattern of both legume
species drastically from significant N release to negligible mineralization or net N
immobilization in incubation studies. Initial plant N concentration and C/N were the two
major factors determining net N release in AVRDC and BRCI soils. In field studies
incorporated legumes decomposed faster than if mulched, and more N was released to the
soil. Nitrogen release in soil of all field studies peaked with 80 - 100 kg NO3-N ha-1 for
soybean GM. This peak N release occurred 2 - 6 weeks after GM application in both
seasons at AVRDC and in the WS at BRCI. Results of the incubation study were
confirmed as the N release peak in MMSU soil was delayed by 1 month compared to
AVRDC and BRCI soils.
Under tropical WS conditions at AVRDC up to 50 % of the nitrate is prone to loss
through leaching from 0 - 50 cm soil layer in tomato fields within a month after GM or
fertilizer application .The productivity legumes, vegetables and maize on the raised beds
was significantly greater than on the low beds in WS.
Tomato yields across sites ranged from 3 - 70 t tomato fruit ha"1. The response of
tomato yields to GM and fertilizer N depended on soil N availability and mineralization.
At sites where soil N was high but was readily leached (WS, AVRDC), or soil N
mineralization was low (DS, MMSU), response of tomato yields to GM and fertilizer N
was high. The opposite was true in the DS at AVRDC and the WS at BRCI, where
tomato yields responded neither to GM nor fertilizer applications.
Maize biomass and N-uptake, following the tomato crop, was increased with
soybean GM compared to control in AVRDC WS and DS. 15N experiments showed that
only a small part (9 -15%) of legume N was accumulated in the tomato crop at MMSU,
112
and that 30 -50% remained in the soil as stable fractions, while 30 - 50 % may be lost
from the system.
We may conclude that the tomato yield response to GM-N is high on poor soils and
N can be substituted fully or partially depending on soil N mineralization. The direct
effect of GM-N on tomato yields is marginal on rich soils.
113
8
Zusammenfassung
Der Stickstoffeintrag von Leguminosen-Griindiingern (LGD) in Nachfolgekulturen
hangt von ihrer Fahigkeit ab, viel Biomasse und N in kurzer Zeit zu akkumulieren, sowie
der Uebereinstimmung von LGD-N-Freisetzung und dem N-Bedlirfnis der Nachfolge-
kultur. Am Beispiel von Alfalfa (Medkago sativa L.), Desmodium (Desmodium intortum
(Mill.) Urb.), Indigofera (Indigofera tinctoria L.) und Soja (Glycine max (L.) Merr.),
60, 75, und 90 Tage nach Aussaat (T) wurden in Feldversuchen Faktoren, wie
Saatdichte, Pflanzenalter und Anbausaison untersucht, da diese die Leguminosen-N-
Akkumulation beeinflussen. In aeroben Inkubationsversuchen wurde die LGD-N-
Freisetzung in drei verschiedenen tropischen Boden (ein silt loamy, mixed, hyperthermic
Fluvaquentic Entochrept aus Taiwan; ein clayey, kaolinitic, isohyperthermic Ultisol; und
ein clayey, mixed, isohyperthermic Fluvaquentic Ustropept aus den Philippinen)
bestimmt.
In einem 6 monatigen Versuch wurde in vier Feldversuchen untersucht, ob die
LGD-N-Freisetzung den N-Bediirfnissen von Gemiisekulturen entspricht. Zwei der
Versuche wurden am Asian Vegetable Research and Development Center (AVRDC) in
Siid-Taiwan, zwei weitere Versuche in den Provinzen Ilocos Norte (MMSU) und
Bukidnon (BRCI) auf den Philippinen durchgefiihit. Am AVRDC und am MMSU
wurden Soja und Indigofera, am BRCI Soja und Mungbohne (Vigna radiata (L.)
Wilczek), nach 60-70 T in den Boden eingearbeitet Oder als Mulch auf der Erdoberflache
liegen gelassen. Verpflanzung der Tomatensetzlinge (Lycopersicum esculentum Mill.)
fand sofort nach Applikation der LGD statt. Die Tomatenertrage in LGD-Verfahren
wurden mit jenen in Mineralstickstoffdungerverfahren (0 - 150 kg N ha-1) verglichen.
Nach der Tomatenernte am AVRDC wurde Mais angebaut, und anhand der Mais-
biomasse und N-Aufnahme (30 T) der Residualeffekt der Dungungsmethode auf eine
Zweitkultur bestimmt. Weiterhin wurden Pflanzenbiomasse, -ertrag und N-Aufnahme
von Leguminosen, Tomaten und Mais am AVRDC gleichzeitig auf zwei Beetsystemen
(Hoch- und Tiefbeete) erfasst. Die Bestimmung des LGD-Abbaus erfolgte am AVRDC
und am MMSU anhand einer Litter-bag Studie. Bodenbeprobungen wurden in
regelmassigen Abstanden in alien Feldversuchen durchgefiihrt, um die N-Freisetzung zu
erfassen. Die Aufnahme von LGD-N in Tomaten sowie der Verbleib des LGD-N in der
114
organischen Masse des Bodens wurden am MMSU mit Hilfe von 15N-Untersuchungen
gemessen.
Saatdichte, Pflanzenalter, Leguminosenart, Anbausaison und -ort stellten sich als
Schliisselfaktoren fiir die Leguminosenbiomassen- und N-akkumulation sowie die
chemische Zusammensetzung der Pflanzen heraus. Nach 60 T akkumulierte Soja am
meisten Biomasse und N. Grosste Biomassenvorteile wurden mit doppelter Saatdichte 60
T erreicht. In der Zeitspanne von 60 bis 90 T erreichte Soja 75 T und Indigofera 90 T
hochste Biomassen- und N-Ertrage. Alfalfa und Desmodium eigneten sich nicht als LGD
aufgrund ihrer zu geringen Biomassen- und N-Ertrage in 60 - 90 T.
In alien Feldversuchen, die in Taiwan und den Philippinen durchgefuhrt wurden,
akkumulierte Soja in 60 - 70 T mindestens 3 t Biomasse ha-1 und 100 kg N ha-1.
Maximal-ertrage von 61 Biomasse ha-1 und 200 kg N ha-1 wurden in der Regenzeit (RZ)
am AVRDC erreicht. Indigoferaertrage zeigten grOssere Ertragsschwankungen und lagen
iiberall defer als jene von Soja. Die Durchschnittsertrage lagen bei 1 t Biomasse ha-1 und
40 kg N ha-1. Variable Bodenbedingungen (z.B. Bodenverdichtung) und Starke
Regenfalle in den ersten Tagen nach der Aussat haben einen grossen Einfluss auf das
Gedeihen von Indigofera aufgrund der kleine Saatkorngrosse und dem langsamen
Auflaufen. Daher ist Indigofera fiir Zeitspannen von 60-90 T als LGD weniger geeignet
als Soja.
Die um 30 Tage langere Wachstumszeit (90 versus 60 T) hatte drastische Folgen
auf die N-Freisetzung beider Leguminosenarten in den Inkubationsversuchen. Walirend
bei 60 Tage altem Pflanzenmaterial eine signifikante N-Mineralisation stattfand, wurde bei
dem 90 Tage alten Pflanzenmaterial innerhalb von 10 Wochen meist kein N freigesetzt.
Der N-Gehalt der Leguminosen sowie das C/N-Verhaltnis waren die beiden
Pflanzenparameter, die die N-Freisetzung in den Boden am AVRDC und BRCI am
starksten beeinflussten. Eingearbeitetes Pflanzenmaterial zersetzte sich in alien
Feldversuchen schneller als Mulch und es konnte mehr N freigesetzt werden. Die
hochsten Nitratwerte mit 80-100 kg NO3-N ha-1 wurden bei Soja LGD gemessen. Diese
wurden 2-6 Wochen nach LGD Applikation in beiden Saisons am AVRDC und wahrend
der RZ am BRCI festgestellt. Die N-Freisetzung im MMSU Boden war um ca. 1 Monat
im Vergleich zu den AVRDC- und BRCI-Boden verzSgert, und bestatigte damit die
Resultate der Inkubationsstudie.
Gemass des Chloridtracerversuches in Tomatenkulturen wurden wahrend der RZ
am AVRDC bis zu 50% des in 0-50 cm Tiefe vorhandenen Nitrates innerhalb des ersten
Monats nach LGD oder N-Mineraldiingerapplikation ausgewaschen. Die Produktivitat
von Leguminosen, Tomaten und Mais auf den Hochbeeten war in der RZ signifikant
hoher als jene auf den Tiefbeeten. Die Tomatenertrage betrugen je nach Standort 3 - 70 t
Errata: Missing page 115
Tomatenha-1. Je hoher d\e Bodenstickstoffverfiigbarkeit eines Standortes, dcsto geringer
die Wirkung der LGD und der Mmeraldunger auf die TomatenertragcDer Residualeffekt von Soja LGD crhohte Biomasscncrtragc und N-Aufnahnie von
Mais Die 15N-Versuche am MMSU konnten zeigen, dass 9-15% dcs LGD-N von der
Tomatenkullur aufgenommcn wurde, und dass weilcre 30-50 % dcs LGD-N in relativ
stabilcn Fraktioncn der organischcn Masse des Bodcns /.urtickbliebcn Etwa 30-50 % dcs
LGD-N gingcn via Auswaschung oder VerflUchligung aus dem System veiloien
Aufgrand der Resultate dieser Studie ist zu schliessen, dass LGD auf armen Boden
eine grosse N-Wirkung auf Tomatcncrtrage austlben konnen Der fur die Tomaten-
produktion benotigle N kann mit Soja-LGD je nach Bodensticksloffvcrfugbarkeit voll
odcr tcihveise crsetzt werden Auf nahrstoffrcichcii Bcxlcn ist die LGD N-Wirkung auf
die Tomatenertrage marginal
Acknowledgments
The work reported here was conducted under the 'associate expert program funded by
the Swiss Development Cooperation (SDC) at the Asian Vegetable and Development Centei
(AVRDC) in Taiwan and the International Rice Research Institute (IRRI) in the Philippines
Many people have contributed directly and indirectly to the successful completion of
this thesis and I take the opportunity to express my thanks First, I would like to thank Dr
David J Midmore (AVRDC), who initiated this project and through his great support made
this study possible I like to thank Dr Jurg Benz (SDC) foi his confidence and the financial
and administrative support of this project For his scientific support and his interest in my
work which has been a constant source of inspiration I would like to thank Dr Urs
Schmidhalter (ETHZ) I am indebted to Prof Dr P Stamp of the Institute of Plant Sciences
at ETH Zurich, supervisor of my thesis I would like to thank Dr J K Ladha (IRRI) who
made it possible to move to IRRI/ Philippines for the second part of my project, and
collaborate in his projects at IRRI Consortium sites, as well as for his scientific suppoit
Special thanks go to Dr D Oik (IRRI) for his enthusiasm and collaboration in soil organic
matter fractionation
Results presented here required an enormous amount of field and laboratory woik For
her technical assistance and field work I would like to thank Ms Hsu Chiou Fen and her field
group, Ms J F Kuo, Ms LI Chiang for laboratory analysis, Ms Ma (Soil Science), Mr
Roan and Ms Chang (crop management) foi helpful discussions at AVRDC, Taiwan At the
Mariano Marcos State University (MMSU) in the Philippines I would like to giatefully
acknowledge the following persons Dr TF Maicos, Site Coordinator of the IRRI Rainfed
Lowland Consortium Site, Ilocos Norte, and Dr S R Pascua for his support in conducting
and supervising field experiments For field and laboratory work I would like to thank
especially Mr H Hidalgo (IRRI), Mr Jose Rizal Mercado and Mi Edmundo Tolentino
(MMSU) For the special interest and the financial support of the last field experiment at San
Juan I would like to thank the Bukidnon Resources Co Inc and especially Mr R J Holmer
for his scientific and practical support
Last but not least, I sincerely thank my husband Vincent for his support and patience
Curriculum Vitae
Name
Date and place of birth
Places of origin
Civil status
Carmen Thonmssen Michel
14 July 1966, Sierre (VS)
Arbaz (VS) and Courtedoux (JU)
married
Education
1973-1981
1981-1986
1986-1991
Primary and secondary schools in Sierre
High school in Brig (Matunte type B)
Studies of Agronomy at the Swiss Federal Institute of Technology
(ETH) Zurich, with specialization in Plant Production
Employment
1992-1994
1994-1995
Since 1992
Associate Expert at the Asian Vegetable Research and
Development Center (AVRDC) in Taiwan
Collaborative Research Fellow at the International
Rice Research Institute (IRRI) in the Philippines
Ph D student at the Institute of Plant Sciences at
ETHZ, section Agronomy
Stays abroad
1990
1990
Practical training in the Integrated Pest Management Project (PLI)
at Lac Alaotra, Madagascar (3 months)
Practical training at the Department of Plantpathology at the Welsh
Plant Breeding Station (WPBS) at Aberystwyth
Spoken and written
languages French, german, enghsh