1995

Agriculture Ecosystems & Environment ELSEVIER

Agriculture, Ecosystems and Environment 56 (1995) 93-108

Crop-weed interference as influenced by a leguminous or synthetic fertilizer nitrogen source: I. Doublecropping experiments with crimson clover, sweet corn, and lambsquarters Elizabeth Dyck *, Matt Liebman, M. Susan Erich Department of Applied Ecology and Environmental Sciences, University of Maine, Orono, ME 04469-5722, USA Accepted 7 September 1995

Abstract Field experiments were undertaken to assess the effect of N source (incorporated legume residue vs. synthetic fertilizer) on crop-weed interference. In a 2 year study, a doublecropping system was used in which a crimson clover ( Trifolium incarnatum L. ) green manure was followed by a crop of sweet corn (Zea mays L.) grown alone or with lambsquarters (Chenopodium album L.). Inclusion of several rates of ammonium nitrate fertilizer in the experiment allowed determination of the clover's N equivalency value ( 55 kg N ha- ~) and contrast of the clover treatment with a comparable rate of N fertilizer addition (45 kg N ha- ~). Soil NO3-N concentration in the experiment at one week after corn and lambsquarters planting was 52% lower in the clover than the fertilizer treatment. Differences in nitrate levels between the two treatments tended to decrease at subsequent sampling dates. At two weeks after emergence, drymatter accumulation of lambsquarters was 72% lower in the clover than the fertilizer treatment and remained 39% lower at final harvest. In contrast, sweet corn biomass accumulation in the clover treatment was 31% lower than in the fertilizer treatment at 2 weeks after emergence but recovered to levels attained in the fertilizer treatment as the growing season progressed. As a result of reduced lambsquarters growth, loss of corn drymatter accumulation to weed interference was 8% in the clover treatment as compared to 28% in the fertilizer treatment. Results of a second experiment in which crimson clover was followed by lambsquarters grown alone also showed a weed suppressive effect of the legume N source in comparison to use of fertilizer N. These experiments demonstrate that use of legume green manure has the potential to reduce the need for herbicide as well as synthetic fertilizer applications in subsequent crops. Keywords: Trifolium incarnatum L.; Zea mays L.; Chenopodium album L.; N fertilizer; Legume green manures

I. Introduction Legume green manures have been assessed primarily in terms of their ability to provide nitrogen (N) to a subsequent crop (e.g. Hargrove, 1986; Hesterman, 1988). Relatively little research has focused on the effect of a legume N source on crop-weed interference. However, there are at least two areas of research that * Corresponding author. 0167-8809/95 /$09.50 © 1995 Elsevier Science B.V. All rights reserved SSDIO 1 6 7 - 8 8 0 9 ( 9 5 ) 0 0 6 4 3 - 5

suggest that substitution of legume green manure for fertilizer N may reduce weed interference, and thus the necessity for herbicide use, in the subsequent crop. On one hand, chemical constituents of a legume green manure or products of its decomposition could inhibit weed growth. Aqueous and volatile extracts of such legumes as sweetclover (Melilotus spp.), berseem clover (Trifolium alexandrinum L.), crimson clover (Trifolium incarnatum L.), and hairy vetch ( Vicia uil-

94

E. Dyck et al. / Agriculture, Ecosystemsand Environment56 (1995) 93-108

losa Roth) have been shown to reduce germination and early seedling growth in bioassay studies (McCalla and Duley, 1948; White et al., 1989; Bradow and Connick, 1990). Moreover, sensitivity to allelochemicals derived from legumes has been found to vary among species; White et al. (1989), e.g. found water extracts of crimson clover and hairy vetch to negatively affect germination and radicle elongation of pitted morning glory (Ipomoea lacunosa L.) and wild mustard (Brassica kaber [D.C.] L.C. Wheeler) more than that of corn (Zea mays L.). Such a differential effect, i.e. greater suppression of weed than crop growth, is essential if the allelopathic potential of a legume green manure were to be used successfully as a weed management strategy. On the other hand, differences in temporal availability of N between synthetic and organic N sources could lead to reduced interference from weeds in systems using a legume green manure. This theory (presented in detail by Liebman and Dyck, 1993) is based on a synthesis of several research findings and assumptions. First, application of synthetic N fertilizer under weed-infested conditions may be of greater benefit to weed than crop growth and, as a result, may depress crop yield (Okafor and De Datta, 1976; Carlson and Hill, 1985; Liebman, 1989). This differential effect of N fertilizer on crop and weed growth is probably linked to the capacity of many arable weeds for early rapid growth and N uptake (Haynes et al., 1991; Seibert and Pearce, 1993), characteristics that confer a competitive advantage in the use of the immediately available N that occurs with synthetic fertilizer application. Second, manipulation of the timing of N availability, i.e. use of split applications of N fertilizer during the growing season rather than as a single application at planting, has been shown to benefit crop growth under weedinfested conditions (Alkaemper et al., 1979). And third, legume residue, in comparison to synthetic N fertilizer, has been characterized as a form of slow or delayed release N (Ladd et al., 1983; Mueller, 1987). In this context, use of a legume N source may decrease weed interference with crop growth through a slow release of N, which could deprive weeds of an early competitive advantage while benefiting those crops whose maximal rates of growth and N uptake occur later in the growing season. Previous work on the effect of legumes on weed management has largely focused on legumes as

smother crops (Nelson et al., 1991; Samson, 1991 ) or as mulches in no-tillage systems (Janke et al., 1989; Hoffman et al., 1993). In these studies, legume treatments have been compared to non-leguminous cover crops or to control treatments unamended with cover crop residue. However, to determine the effect of a legume as a N source on crop-weed interference, it is more appropriate to compare use of the legume green manure to that of synthetic fertilizer N. Accordingly, a series of experiments was conducted to determine the effect of N source (legume green manure vs. N fertilizer) on weed and crop growth. Because the investigators' interest was on the effect of hypothesized qualitative differences between N sources and because amount of N can also influence crop-weed interactions (Lawson and Wiseman, 1977; Carlson and Hill, 1985), a legume green manure treatment was compared with an application rate of N fertilizer that supplied an equivalent amount of N to the subsequent test crop. It was hypothesized that the legume N source would result in ( 1 ) reduced weed growth and (2) enhanced crop growth and yield under weedy conditions in comparison to application of fertilizer N.

2. Methods and materials 2.1. Experimental system The above hypotheses were tested within a doublecropping system of crimson clover followed by sweet corn. The doublecropping system was used to facilitate temporal replication of the experiment. Crimson clover, which is capable of rapid biomass accumulation (Knight and Hoveland, 1985), was selected as the green manure crop to optimize biomass production within the doublecropping system. Corn was used as the test crop because corn production systems are heavily dependent both on N fertilizer and herbicides (Economic Research Service, 1993). Sweet corn was used in particular because it had performed well in past seasons at the experiment station following early summer plowdown of overwintering legume green manures, e.g. alfalfa (Medicago sativa L.) and red clover (Trifolium pratense L.). To further accommodate the constraints of a double-cropping system, a short season variety (70 days) of sweet corn was used. Lambsquarters (Chenopodium album L.), a major pest of corn

E. Dyck et aL /Agriculture, Ecosystems and Environment 56 (1995) 93-108

95

Table 1 Dates of major field operations in Experiments 1 and 2 Field operation

Crimson clover planted Crimson clover incorporated Sweet corn and lambsquarters planted Experiment irrigated First weeding Second weeding

Experiment 1

Experiment 2

1989

1990

1989

11 May 10 July 12-13 July 22 July 4--10 August 29-31 August

2 May 10 July 11 July 16 July 2-4 August 15-16 August

22 May 24 July 26 July a I August 18-23 August 14-15 September

~'Lambsquarters only was planted in Experiment 2.

(Holm et al., 1977), was chosen as the infesting weed species. Although most recent studies of the effect of legume residues on weeds have been conducted within notillage systems, in these experiments the effect of incorporated legume residue was studied for two reasons. First soil incorporation of residue has been shown to accelerate or intensify microbially mediated transformations of N (e.g. immobilization (Varco et al., 1993) and rate of N release from plant tissue (Wilson and Hargrove, 1986)) that can in turn affect subsequent plant growth. Second, incorporation, particularly shallow incorporation, serves to concentrate residue in the seed and seedling rooting zones, which may intensify effects of residue on early plant growth. To further ensure crop and weed interaction with the legume green manure, in these experiments crop and weed seeds were planted into freshly incorporated legume residue. Two experiments were undertaken. In Experiment 1 (conducted for 2 years), crimson clover was followed by sweet corn grown alone and in association with lambsquarters. To allow separation of direct effects of N source on weed growth from that of crop interference, Experiment 2 (conducted for 1 year) was undertaken in which crimson clover was followed by lambsquarters grown alone.

2.2. Site description, experimental design, and field operations Experiment 1 was initiated in 1989 at the University of Maine Sustainable Agriculture Research Center in Stillwater, Maine. The site consisted of both a Adams fine sandy loam (sandy, mixed, frigid Typic Haplor-

thods) and a Boothbay silt loam (fine, silty, mixed, frigid aquic Dystric Eutrochrepts), and the experiment was blocked accordingly. In the 5 years before the start of experimentation, a series of field crops of small grains and legumes had been grown on the site. In 1990 the experiment was repeated on the same site used in 1989, each plot receiving the same experimental treatment in both years. The experiment used a split-plot randomized block design with four replications. Main plots consisted of six N treatments: crimson clover residue, a 0 N treatment (receiving neither residue nor N fertilizer), and ammonium nitrate applied at rates of 45, 90, 135, and 180 kg N ha- 1. The 0 N treatment and various fertilizer rates were included to allow estimation of a N fertilizer equivalency value of the crimson clover residue. Subplots were weed-free corn or corn grown with lambsquarters. Plots measured 3.0m × 9.1 m to accommodate four corn rows spaced 76 cm apart. Specific dates of major field operations are given in Table 1. In early May of each year the entire site was rotovated once. Crimson clover (cv. Dixie) seed inoculated with Rhizobium trifolii was then sown in appropriate plots by hand at a rate of 84 kg ha J (a high seeding rate was used to ensure sufficient clover stand development within the doublecropping system). Nonlegume treatments were kept in bare fallow by rotovating twice. At onset of flowering, the clover was mowed and its residue incorporated with a rotovator to a depth of 10-15 cm. After treating all plots with P and K fertilizer at rates recommended by the University of Maine Soil Testing Lab (in 1989, 84 kg ha -1 each of P205 and K20; in 1990, 84 kg ha -~ of P205 and 196 kg ha- 1 of K20), the entire experiment was disk har-

96

E. Dyck et al./ Agriculture, Ecosystems and Environment 56 (1995) 93-108

rowed (in 1989) or rotovated (in 1990). Within 2 days of clover incorporation, sweet corn (cv. Sugar Buns) was sown in all plots. Plots were heavily seeded with corn in both years and then thinned to assure adequate stands. In 1989, corn was planted mechanically. In 1990, in an attempt to achieve more even spacing of corn plants within the row, corn was seeded by hand. In each year, on the same day as corn planting ammonium nitrate treatments were broadcast into appropriate plots and then shallowly raked into the soil to a depth of 2-3 cm. Lambsquarters seed (collected from the experimental farm the previous year and acid-pretreated to remove hardseededness (Ellis et al., 1985) ) was then sown by hand over the corn rows in appropriate plots at a rate of 1.1 g m - 2. In 1989, clover plots were seeded with double the amount of lambsquarters seed used in the 0 N and N fertilizer treatments in anticipation of possible inhibitory effects of the residue on emergence. In 1990, equal amounts of lambsquarters seed were sown in all weedy corn treatments in an effort to document possible residue effects on emergence. To ensure adequate crop and weed establishment, plots were irrigated in both years with overhead sprinklers within 12 days after planting; total irrigation water delivered was 28 mm in 1989 and 21 mm in 1990. Plots were handweeded and hoed twice in each season to remove all weeds from the corn alone treatment and all weeds but lambsquarters from the corn + lambsquarters treatment. To simulate field conditions in which mechanical cultivation is used for weed control, iambsquarters plants in the weedy plots were confined to bands 30 cm wide centered on the corn rows. Corn and lambsquarters were thinned between 1-2 weeks after corn emergence. Target densities were 6.5 plants m -2 and 26 plants m -2 for corn and lambsquarters, respectively. The corn density is in accordance with the standard planting practices for sweet corn in the northeastern U.S. (Erhardt, 1991). The weed density required to cause substantial crop yield loss can vary between sites and years; to insure a competitive interaction between crop and weed, the target lambsquarters density selected was considerably higher that the 8.7 lambsquarters plants m -2 reported by Beckett et al. (1988) to produce significant yield loss in field corn. Experiment 2 was conducted in 1989 on a Boothbay silt loam at a site adjacent to Experiment 1. A corn-

pletely randomized design was used with four replications. Nitrogen treatments were identical to those used in Experiment 1, while plot size was 3.0 m X 4.6 m. Dates of major field operations are given in Table 1. Field operations and plot management techniques were the same as those used in Experiment 1, with the following exception. Although lambsquarters stands were kept free from other weeds throughout the growing season, no attempt was made to thin the lambsquarters stands in each plot to a constant density. Instead, counts of plant number were made at each sampling date to check for potential differences in density between plots or treatments.

2.3. Sampling and analytical procedures In both Experiments 1 and 2, crimson clover biomass was sampled immediately before clover incorporation. Shoot material was collected in each legume plot from two 0.05 m 2 quadrats. In Experiment 1, emergence counts of sweet corn, lambsquarters, and other weeds were taken each year before thinning and weeding occurred using five 0.23 m 2 quadrats per plot. Aboveground corn and lambsquarters biomass were sampled at approximately 2, 4, and 6 weeks after corn emergence from the two center corn rows of each plot using two 0.33 m 2 quadrats. At the end of the growing season, six 0.33 m 2 quadrats were collected. Four of the quadrats were bulked and the freshweight of each species determined. The remaining two quadrats were bulked and both freshweight and dryweights determined in order to generate a conversion factor of percent drymatter for each species on a plot by plot basis. At each biomass harvest, heights of sampled corn and lambsquarters plants were determined to the furthest extended leaf tip. Stem diameter measurements (taken slightly above the soil surface) were made of corn plants at all biomass harvests and of lambsquarters plants at the final three harvests. Sweet corn and lambsquarters leaf number were determined at the first harvest date. In Experiment 2, lambsquarters biomass was sampled at five dates during the growing season using two 0.167 m 2 quadrats. Height and stem diameter measurements and leaf counts were taken as in Experiment 1. In both experiments, samples used in dryweight determination were dried for several days at approximately 65°C before weighing. To determine whole

E. Dyck et al. / Agriculture, Ecosystemsand Environment56 (1995) 93-108

97

Table 2 Crimson cloverbiomassaccumulation, N concentration,and N fertilizerequivalencyvalues in Experiment 1 Year

Aboveground drymatteraccumulation (kg ha- ~)

N concentration (g kg - 1)

Predictionequationa

r2

CalculatedN fertilizer equivalencyvalue (kg N ha- ~)

1989 1990

5136 4758

24.0 24.6

y = 0.009x +5.507 y = 0.026x +9.070

0.58 0.83

52 58

~Wherey is the total N accumulationin weedfreecorn and x is the N fertilizerrate. plant N content, dried plant tissue samples were ground to pass through a 40-mesh screen and subjected to the block digestion procedure outlined by Wall and Gehrke (1975). The digests were analyzed for NH4+ on a Lachat automated ion analyzer. Soil samples were collected at four dates after corn planting in both years of Experiment 1. Using randomly generated coordinates, five soil cores, taken to a depth of 25 cm, were collected and bulked to form a composite sample for each experimental plot. The samples were immediately frozen and stored until they could be extracted. Samples were extracted using 5 g (oven dryweight basis) soil/50 ml 1N KCL and a shake time of 30 min. Soil extracts were analyzed for both NH4-N and NO3-N content on a Lachat automated ion analyzer. 2.4. Statistical analysis

Total crop N accumulation (drymatter accumulation × crop N concentration), which is a more sensitive indicator of external N supply than yield (Hargrove, 1986; Hesterman et al., 1992), was used as a basis for estimating an N fertilizer equivalency value for crimson clover. In each year of Experiment 1, a fertilizer equivalency value was calculated using the following steps: ( 1 ) total N accumulation of corn grown under weedfree conditions in the non-legume treatments was analyzed using polynomial contrasts for significant linear, quadratic, or cubic trends, (2) the appropriate model was then used to generate a prediction equation by regressing mean treatment values of corn N accumulation on N fertilizer rate, and (3) the mean value of N accumulation in weedfree corn following crimson clover was inserted into the equation and an estimate of N fertilizer equivalency of crimson clover calculated. The crimson clover treatment was then compared to the N fertilizer treatment closest to its equivalency value using single degree of freedom contrasts within an analysis of variance. Data generated in Experiments 1 and

2 were analyzed using the repeated measures option in the GLM procedure of SAS (1990). Results are reported as between effects (treatments averaged across sampling dates) and within effects (treatment by sampling date interactions). Multivariate analysis was used to test hypotheses concerning sampling date by treatment interactions.

3. Results and discussion 3.1. Experiment1 3.1.1. Crimson clover biornass accumulation and N fertilizer equivalency values Crimson clover aboveground biomass accumulation averaged 4947 kg ha- 1 over both years of Experiment 1 (Table 2). The average N concentration of crimson clover shoots was 24.3 g kg - t. In both years of the experiment, trend analysis data indicated a significant linear relationship between N accumulation in weedfree corn and N fertilizer rate ( P < 0 . 0 5 ) . Prediction equations generated from subsequent regression analysis provided estimates of N fertilizer equivalency as ammonium nitrate for crimson clover of 52 and 58 kg N ha - t , respectively, in 1989 and 1990 (Table2). Nitrogen fertilizer equivalency values of legumes for sweet corn have apparently not been previously calculated. However, the values for sweet corn found in this study appear reasonable in light of average values for N contribution of crimson clover reported in studies for no-tillage field corn systems in the southeastern USA (72 kg N ha -1 (Smith et al., 1987)) and in Maryland (63 kg N h a - t ; Holderbaum et al., 1990). In subsequent analysis of soil, lambsquarters, and sweet corn data, the crimson clover treatment was therefore contrasted with the 45 kg N h a - t treatment, which in both 1989 and 1990 was the N rate closest to the clover's estimated N fertilizer equivalency value. For pur-

98

E. Dyck et al./ Agriculture, Ecosystems and Environment 56 (1995) 93-108

poses of comparison, results of the 0 N treatment have also been reported.

3.1.2. Soil inorganic N Because of divergent sampling dates between years, soil data for each year of Experiment 1 were analyzed separately. Temporal patterns of plant-available N in the soil differed significantly between N sources (45 kg N h a - ~ and crimson clover) in both 1989 and 1990 (Table 3 and Table 4). In 1989, initial NO3-N and NH4-N concentrations were 65% lower in the clover than the fertilizer treatment. At subsequent sampling dates, differences in nitrate levels between the two treatments steadily decreased while ammonium concentration in the two treatments differed by less than 1 mg kg - 1. In 1990, NO3-N concentration in the legume treatment was 40% lower than in the fertilizer treatment at eight days after planting (DAP). By 59 DAP, however, nitrate concentrations in the clover treatment were twice those in the fertilizer treatment. Levels of NH4N in 1990 did not differ between the two treatments. Concentration of NHa-N was not affected by the presence of weeds in either year (Table 4). However, in both years, NO3-N concentrations were reduced in all weedy treatments as the growing season progressed: weed presence had lowered nitrate levels 42% at 53 DAP in 1989 and 67% at 59 DAP in 1990. No significant differences in the effect of weeds on NO3-N levels were detected between clover and fertilizer treatments (N source X weed effects) in either year. In both years, differences in total soil mineral N (NO3-N + NH4-N) between clover and fertilizer treatments paralleled differences in NO3-N concentrations in terms of both weed and N source effects. 3.1.3. Emergence of lambsquarters, other weeds, and sweet corn In both years of Experiment 1, weed emergence in the clover treatment was delayed or suppressed when compared to emergence in the N fertilizer treatment. Although crimson clover plots had been sown with a higher rate of lambsquarters seed in 1989, lambsquarters emergence did not differ between clover and fertilizer treatments (Table 5). In 1990, when an equal rate of lambsquarters seed was sown in all treatments, lambsquarters emergence in the clover treatment was 34% lower than in the fertilizer treatment at 13 DAP, although by 21 DAP lambsquarters densities were

equivalent in the two treatments (Table 5 ). Emergence of weeds other than lambsquarters (of which the maj or species were Amaranthus retroflexus L. and Capsella bursa-pastoris [L.] Medic.) was strongly suppressed in the crimson clover treatment in both years: counts of other weeds were 60% lower in the clover than the fertilizer treatment at 18 DAP in 1989 and 74% lower at the single count of other weeds taken at 13 DAP in 1990. Emergence data also indicate a possible inhibitory effect of crimson clover on sweet corn emergence. Although corn emergence in 1989 did not significantly differ between N source treatments, in 1990 sweet corn emergence in the clover treatment was 29% lower than in the N fertilizer treatment at 13 DAP.

3.1.4. Lambsquarters and sweet corn stand densities Lambsquarters densities at final harvest were 35 plants m - 2 in 1989 and 81 plants m - 2 in 1990. Lambsquarters number exceeded the target density in both years because of flushes of emergence that occurred after thinning. Despite increase in density after stand establishment, analysis of lambsquarters counts taken at each biomass harvest showed no difference in density between the crimson clover and fertilizer treatments (data not shown). After thinning, sweet corn density was 6.8 plants m - 2 in 1989 and 7.3 plants m - 2 in 1990. 3.1.5. Lambsquarters and sweet corn growth 3.1.5.1. Seasonal effects on lambsquarters and corn growth Despite the use of a short season cultivar, sweet corn failed to develop marketable ears before killing frosts in either 1989 and 1990. Treatment effects on lambsquarters and corn growth have been analyzed in terms of aboveground biomass accumulation,, height and stem diameter development, and leaf number. Similar sampling dates and comparable N fertilizer equivalency values in each year allowed 1989 and 1990 plant growth data to be analyzed jointly. Although a joint analysis of 1989 and 1990 lambsquarters and sweet corn biomass accumulation data showed a significant year effect, there was no evidence of an interaction of year with N source or related effects (year X N source × weed, year × N source × sampling date, year X N source X weed X sampling date). Analysis of lambsquarters and sweet corn stem diameter and height data produced similar results. Because of this

E. Dyck et aL /Agriculture, Ecosystems and Environment 56 (1995) 93-108

99

Table 3 Soil inorganic N concentrations (mg N k g - ~ soil) at four sampling dates in 1989 and 1990 in Experiment 1 1989 N treatment

9 DAP a

29 DAP

53 DAP

84 DAP

Weedfree corn

Weedy corn Weedfree corn

Weedy corn Weedfree corn

Weedy corn Weedfree corn

Weedy corn

15.4 26.8 8.6

14.6 25.8 9.5

29.4 28.6 12.7

19.1 34.5 17.0

9.6 16.2 8.3

4.1 8.4 5.9

2.9 5.4 2.8

1.2 1.8 2.7

1.7 8.9 2.3

1.1 7.0 3.2

1.3 0.7 1.3

1.4 0.8 1.4

0.8 0.8 0.8

0.6 0.7 0.8

2.7 2.7 2.4

2.7 2.7 3.3

NO3-N 0N 45 kg N h a - ~ Crimson clover

NH4-N 0N 45 kg N h a - l Crimson clover 1990 N treatment

8 DAP

22 DAP b

37 DAP

59 DAP

Weedfree corn

Weedy corn Weedfree corn

Weedy corn Weedfree corn

Weedy corn

Weedfree corn

Weedy corn

9.3 16.9 10,8

9.7 20.0 11,5

11.6 21.1 17.1

10.8 21.0 18.1

6.6 11.2 11.5

6.1 8.9 11.2

1.3 2.7 6.6

0.5 1.1 1.5

1,4 7.3 4.1

1.2 5.4 3.0

34.6 36.0 35.7

33.6 20.7 30.1

1.1 1.8 1.0

1.1 0.8 1.0

1.1 1.2 1.6

1.2 1.1 1.4

NO3-N 0N 45 kg N ha - l Crimson clover

NH4-N 0N 45 kg N h a - ~ Crimson clover

aDays after planting.~The high levels of NI-~-N occurring at 22 DAP in 1990 are a likely result of a heavy rainfall and subsequent flooding of the experimental site on 16 DAP, which may have precipitated a flush of ammonification and temporarily impeded nitrifying activity.

similarity in N source effect between years, results from a single year (1989) have been used to illustrate N source x sampling date effects (Fig. 1).

3.1.5.2. Lambsquartersgrowth Use of the crimson clover treatment substantially reduced lambsquarters growth in the weedy corn plots when compared to that of N fertilizer (Table 6). The suppressive effect of the clover treatment was immediate and strong: at the first sampling date (approximately two weeks after corn and lambsquarters emergence) in 1989 and 1990, lambsquarters biomass in the crimson clover treatment was 64% and 81% lower, respectively, than in the fertilizer treatment. Lambsquarters growth was also characterized by a significant N source X sampling date interaction (Table 6): in both years drymatter accumulation in the

clover treatment gradually increased in relation to that in the fertilizer treatment after the first sampling date, as can be seen when biomass is plotted on a natural log scale (Fig. la). However, despite decreased differences between treatments as the growing season progressed, lambsquarters remained 37% and 42% lower, respectively, in the clover than the fertilizer treatment at final sampling dates in 1989 and 1990. In both years lambsquarters growth in the clover treatment never exceeded that in the 0 N treatment (Table 6). Lambsquarters plants grown with crimson clover consistently had both smaller stem diameters (P<0.0001) and were shorter (P<0.0001) than plants grown with N fertilizer (Fig. lb and lc). Both lambsquarters stem diameter and height showed an interaction of N source with sampling date (P < 0.05). However, while the difference in stem diameter between N source treatments steadily decreased with

E. Dyck et al. / Agriculture, Ecosystems and Environment 56 (1995) 93-108

100

Table 4 Results of multivariate analysis of variance for repeated measures on soil inorganic N levels at four sampling dates in 1989 and 1990 in Experiment 1" 1989

1990

NH4-N

NO3-N

NH,,-N + NO3-N

NH4-N

NOa-N

NH4-N + NOa-N

ns ns ns

* **** ns

** **** ns

ns ns ns

ns **** ns

ns **** ns

*** ns ns

* *** ns

** *** ns

ns ns ns

** ** * * ns

** ** * * ns

Between effects N source (45 kg N - ~vs. crimson clover) Weed b N source × weed

Within effects N source × sampling date Weed × sampling date N source × weed × sampling date

~Before analysis, data were In-transformed to reduce heterogeneity of variances among sampling dates.bWeed signifies presence or absence of lambsquarters.Significance levels: ****, P < 0.0001; ***, P < 0.001; * *, P < 0.01; *, P < 0.05; ns, not significant.

Table 5 Emergence of sweet corn, lambsquarters, and other weeds (plants m -2) at two sampling dates in 1989 and 1990 in Experiment 1 N treatment 1989

1990

12 DAPa

18 DAP

! 3 DAP

21 DAPb

Sweet corn

Lambsquarters

Other weeds

Sweet corn

Lambsquarters

Other weeds

Sweet corn

Lambsquarters

Other weeds

4.7 4.9

28.0 25.2

17.4 21.1

6.6 5.9

39.5 40.6

56.6 52.6

15.7 16.0

240.6 271.5

16.4 45.5

129.7 119.8

Crimson 3.7 17.9 6.7 6.0 clover Multivariate analysis for repeated measuresd

31.3

20.8

11.4

177.9

11.8

116.0

0N 45 kg N ha- 1

1989 ~

N source (45 kg N ha- ~ vs. crimson clover) N source × sampling date

Sweet corn

Lambsquarters~

Other weeds

1990 f

Sweet corn

Lambsquarters

Other weeds

Sweet corn

Lambsquarters

Other weeds

ns ns

ns ns

*** ns

*** -

** **

* -

aDays after plantingbCounts of lambsquarters only were taken at this date.~The substantial reduction in lambsquarters densities from first to second count dates in all treatments was caused by an intense rainstorm at 16 DAP.dBecause of the differences in number of sampling dates between years, emergence data for each year were analyzed separately.eBefore analysis, emergence data in 1989 were transformed in the following manner: (y + 0.5)°5.fEmergence counts of sweet corn and other weeds in 1990, which were taken on a single date, were subjected to univariate analysis of variance.Significance levels: ***, P < 0.001; **, P < 0.01; *, P < 0.05; ns, not significant.

s a m p l i n g d a t e ( F i g . l b ) , d i f f e r e n c e s in h e i g h t b e t w e e n

at t h e first s a m p l i n g d a t e w a s 4 1 % l o w e r in t h e c l o v e r

fertilizer and clover treatments increased between the

t h a n t h e f e r t i l i z e r t r e a t m e n t ( 4 . 3 vs. 7.3 l e a v e s p e r p l a n t

first a n d s e c o n d s a m p l i n g d a t e s b e f o r e d e c r e a s i n g a g a i n

in

b y final h a r v e s t ( F i g . l c ) . I n b o t h y e a r s , l e a f n u m b e r

P
1989,

3.9

vs.

6.6

leaves

per

plant

in

1990;

E. Dyck et al. /Agriculture, Ecosystems and Environment 56 (1995) 93-108

Larnbsquarters 13 a. Bi 11.25 -

1.8 b.

7

S t e m ~

c. Height

o m a s s / ~ 1.4-

.

oo

/

/

6-

1-

0.7 E E c:

0.6--

7.75 -

101

5-

0 4

40.2-

620

Sweet

I

I

I

30

40

50

20406080

I

I

60

70

80

20

I

'I'

30

"0

50

20 "0 60 80

i

i

60

70

3 .... 20

80

I ' ~ 30 40

I 50

20"06080 t I 60 70 80

Corn

12.5,,..

d.

B i o m a s ~ 2.75 -

c~

I

7.25

e. Stem Diameter

f. Heightf

2.25 -

10.5 -

_= 8.5-

2'

20

f

o.ts-W----- 1

o.1t\

0.051

//

I

i

I

I

30

40

50

i

i

!

20`06080

'i

I

60

70

k

°'1 80

20

i

,

,

30

40

50

..q

I 5.25

21'°?'° 60

70

/

6.25 -

I

'2'1¢ 0o:t ,

0

6.5-

1.-,5- !

80

1

°15V-q

'~ 20

20 40 60 80

I

I

I

I

"t

30

40

50

60

70

80

Days after Planting Fig. 1. Biomass, stem diameter, and height ( natural log scale ) of lambsquarters and sweet corn (averaged across weedfree and weedy treatments ) at three or four dates in 45 k g N h a - ] synthetic fertilizer ( o ) and crimson clover (e) treatments in 1989 in Experiment 1. Insets illustrate N source × sampling date interactions, i.e. the relative difference between treatments (in I N fertilizer value] - in [crimson clover value] ) over the sampling dates. N source × sampling date interactions for the joint analysis of 1989 and 1990 data were significant at P < 0.05 for each growth parameter in each species.

3.1.5.3. Sweet corn

growth

Sweet corn biomass accumulation was characterized by an N source x weed X sampling date interaction (Table 6). The interaction stemmed from the differential response of fertilizer-grown and clover-grown sweet corn to weed presence at the last two sampling dates in both years. At 50 DAP in 1989, corn biomass was 33% lower in the weedy than in the weedfree N fertilizer treatment while corn grown with crimson clover remained unaffected by the presence of weeds. By final sampling date, loss of corn drymatter to weeds was 36% in the fertilizer treatment as compared to 14% in the clover treatment. In 1990, the effect of weeds was again delayed and reduced in the crimson clover treatment: biomass loss to weed interference was 19% at 54 DAP and 21% at 86 DAP in the fertilizer treatment as compared to no loss and a 2% loss, respectively, at these same dates in the clover treatment.

Sweet corn biomass accumulation also manifested a strong N source X sampling date interaction (Table 6, Fig. ld). As with lambsquarters, the clover treatment initially negatively affected corn drymatter accumulation; at approximately 2 weeks after emergence in 1989 and 1990, corn biomass in the clover treatment was 25% and 38% lower, respectively, than in the fertilizer treatment. However, at final harvest in these same years, biomass of clover-grown corn was 2% and 7% higher, respectively, than that of corn grown with N fertilizer. The greater growth of weedy corn in the clover treatment undoubtedly contributed to the decrease in relative difference between the clover and fertilizer treatments when averaged across weedy and weedfree treatments at the last two sampling dates in either year. However, comparison of weedfree fertilizer and clover treatments at successive sampling periods also shows

E. Dyck et al. / Agriculture, Ecosystems and Environment 56 (1995) 93-108

102

Table 6 Aboveground biomass accumulation (g m -2) at four sampling dates in 1989 and 1990 in Experiment 1 1989 N treatment 22 DAW

36 DAP

50 DAP

77 DAP

Weedfreecorn Weedy Lambs- Weedfree Weedy Lambs- Weedfree corn quarters corn corn quarters corn

Weedy corn

Lambs- Weedfree quarters corn

Weedy corn

Lambsquarters

0N 45 kg N ha-l

0.88 1.40

0.71 1.24

1.29 1.06

12.10 12.36

10.52 18.63

18.04 18.55

58.4 95.7

33.2 64.0

101.4 108.1

239 305

145 195

294 325

Crimson clover

0.67

1.31

0.38

14.51

9.15

9.37

74.4

89.8

61.0

274

235

203

1990 N treatment 23 DAP

0N 45 kg N ha- J

37 DAP

54 DAP

86 DAP

Weedfreecorn Weedy Lambs- Weedfree Weedy Lambs- Weedfree corn quarters corn corn quarters corn

Weedy corn

Lambs- Weedfree quarters corn

Weedy corn

Lambsquarters

3.22 3.75

3.20 5.22

1.11 1.91

Crinxson 2.81 2.78 0.37 clover Multivariate analysis for repeated measuresb

37.78 49.77

47.44 51.34

18.39 23.98

279.0 340.9

202.9 276.1

165.8 143.4

609 669

469 529

152 214

52.76

55.03

7.61

288.2

321.0

63.2

646

635

124

Lambsquarters

Sweet corn

**** -

ns **** ns

* * -

**** ** ** t

Between effects N source (45 kg N ha- ~vs. crimson clover) Weed ~ N source × weed

Within effects Year N source x sampling date Weed × sampling date N source × weed × sampling date

aDays after planting.~Before analysis, data were In-transformed to reduce heterogeneity of variances among sampling dates?Weed signifies presence or absence of lambsquarters. Significance levels: ****, P < 0.0001; **, P < 0.01; *, P <0.05; t, P < 0.10; ns, not significant.

a pattern of initially reduced growth in the clover treatment followed by recovery to levels approaching or equivalent to those in the N fertilizer treatment (Table 6), which suggests that accelerated growth in the clover treatment after the first sampling date contributed to the diminished difference between the two treatments. Sweet corn stem diameter was also affected by a N source × sampling date interaction (P < 0.0001). The

response of corn stem diameter to N source resembled that of biomass accumulation: in both years initially thinner corn in the legume treatment recovered to have stem diameters that were equivalent to or thicker than corn stem diameters in the N fertilizer treatment (Fig. le). Corn stem diameter was also negatively affected by weed interference in all treatments as the experimental season progressed (data not shown,

P
E. Dyck et aL /Agriculture, Ecosystems and Environment 56 (1995) 93-108

Height development in sweet corn showed a consistent response to N source (P < 0.01). Although the relative difference in height between the two treatments tended to decrease as the growing season progressed (Fig. l f, P < 0.05 ), clover-grown corn was shorter than corn grown with fertilizer at all sampling dates in both years. Corn height was increased by the presence of weeds in all treatments (data not shown, P < 0.01), a probable result of corn stem etiolation in response to shading from lambsquarters within the corn row. Corn leaf number at the first sampling date was reduced by 9% and 4%, respectively, in 1989 and 1990 in the clover as compared to the fertilizer treatment (data not shown, P
3.1.5.4. Nitrogen concentration and accumulation in lambsquarters and sweet corn Nitrogen concentration at final harvest in both lambsquarters and sweet corn was characterized by a year × N source interaction (Table 7). For both weed and crop, N concentration did not differ between N source treatments in 1989 but was higher in the crimson clover treatment than the fertilizer treatment in 1990. The presence of weeds produced equivalent reductions in sweet corn N concentration in fertilizer and clover treatments in both years. Results of analysis of N accumulation at final harvest paralleled that of biomass accumulation for both lambsquarters and corn (Table 7). In 1989 and 1990, N accumulation in clover-grown lambsquarters was 38% and 29% lower, respectively, than in lambsquarters grown with N fertilizer. Sweet corn N accumulation were characterized by a N source × weed interaction: in 1989 and 1990 weed interference reduced N accumulation in corn by 47% and 35%, respectively, in the fertilizer treatment as compared to 26% and 12%, respectively, in the legume treatment. 3.2. Experiment 2 Biomass accumulation in the crimson clover phase of Experiment 2 was 6267 kg h a - ~while N concentration before clover incorporation was 23.1 g kg-1. To facilitate comparison between Experiments 1 and 2, Experiment 2 was analyzed by contrasting the crimson clover treatment with the fertilizer treatment of 45 kg N h a -j.

103

Although equal applications of lambsquarters seed had been made to all experimental units, season-long lambsquarters density in the crimson clover treatment was lower (P < 0.05) than in the N fertilizer treatment (68 vs. 107 plants m - 2, respectively). Patterns in drymatter accumulation in Experiment 2 were consistent with those from both years in Experiment 1, i.e. biomass accumulation of lambsquarters was significantly lower in the crimson clover plots than in the N fertilizer plots throughout the experiment (Table 8, Fig. 2a), remaining 34% lower at the final sampling date. Lower biomass accumulation in the clover as compared to the fertilizer treatment in Experiment 2 could be explained by the difference in densities between the two treatments. However, lambsquarters biomass considered on a per plant basis was also significantly lower in the clover than the fertilizer treatment (data not shown; P < 0.05), which indicates that the reduced growth of lambsquarters in the legume treatment resulted from factors other than reduced plant density. Lambsquarters stem diameter was negatively affected (P < 0.10) by the clover treatment, although the effect clearly did not persist throughout the experimental season (Fig. 2b). Lambsquarters plants were shorter in the clover than the fertilizer treatment (P < 0.001 ). Height development in the two treatments showed a pattern similar to that in Experiment 1; i.e. the difference in relative heights between the two treatments initially increased (from 23 to 51 DAP) before decreasing at the end of the season (Fig. 2c). Leaf number at first sampling date was lower (P < 0.01 ) in the clover than the fertilizer treatment (4.1 vs. 5.1 leaves per plant, respectively). Results of analysis of N concentration and accumulation at final harvest in Experiment 2 are also in keeping with Experiment 1. While N concentration in clover-grown lambsquarters was higher than that in lambsquarters grown with N fertilizer, N accumulation remained lower in the clover than the fertilizer treatment (Table 8).

3.3. Nitrogen source effects on crop-weed interference The results of Experiment 1 demonstrate a differential effect of a crimson clover N source on crop and weed growth in comparison to that of fertilizer N. Although the growth of both lambsquarters and sweet

E. Dyck et al./ Agriculture, Ecosystems and Environment 56 (1995) 93-108

104

Table 7 N concentration (g k g - 1) and N accumulation ( g m -2) in sweet corn and lambsquarters at final harvest dates in 1989 and 1990 in Experiment 1 N treatment

N concentration

N accumulation

1989

1990

1989

1990

weedfree corn

weedy corn

lambsquarters

weedfree corn

weedy corn

Iambsquarters

weedfree corn

weedy corn

lambsquarters

0N 45 kg N ha-1

21.1 22.1

16.5 18.5

25.4 28.2

15.1 14.9

11.6 12.1

23.5 21.9

5.06 6.75

2,42 3.60

7.49 9.12

Crimson clover

21.8

18.3

28.1

16.3

14.5

25.8

5.97

4.39

5.69

weedfree corn

weedy corn

lambsquarters

9.18 9.94

5.50 6.48

3.53 4.38

10.57

9.31

3.12

Multivariate analysis for repeated measures N concentration

N accumulation

Sweet corn

Lambsquarters

Sweet corn

Lambsquarters

Between effects N source (45 kg N ha- l vs. crimson clover)

t

ns

t

***

Weed

****

_

****

_

ns

-

*

**** ** ns ns

**** t -

**** ns * ns

a

N source x weed

Within effects Year N source X year Weed × year N source X weed x year

**** ns -

aWeed signifies presence or absence of lambsquarters.Significance levels: ****, P<0.0001; ***, P<0.001; **, P<0.01; *, P<0.05; j, P < 0.10; ns, not significant

Table 8 Aboveground biomass accumulation at five sampling dates and N concentration and N accumulation at final harvest in Experiment 2 N treatment

Biomass accumulation (g m -2) 23 DAP ~

37 DAP

0N 0.650 10.31 45 kg N ha- ~ 0.901 12.08 Crimson clover 0.336 5.69 Multivariate analysis for repeated measures

N source (45 kg N h a - ' vs. crimson clover) N source × sampling date

N concentration (gkg -t)

51 DAP

66 DAP

79 DAP

79 DAP

N accumulation ( g i n -2) 79 DAP

78.6 110.4 47.0

104.1 209.0 119.8

186.7 335.0 222.1

31.8 31.3 34.8

5.93 10.39 7.73

Biomass accumulationb

N concentrationc

N accumulation

*** ns

* -

t -

aDays after planting.~Before analysis, biomass data were In-transformed to reduce heterogeneity of variance among sampling dates.ON concentration and accumulation data were subjected to univariate analysis of variance. Because of recalcitrant heterogeneity of variances, N accumulation data were rank-transformed before analysis.Significance levels: ***, P < 0.001; *, P < 0.05; t, P < 0.10; ns, not significant.

E. Dyck et al. /Agriculture, Ecosystems and Environment 56 (1995) 93-108

13 "

a. 8 i o r n a ~

b. StemDiameter

6.5

c.

105

Fleight

""

1.3 55-

0.8

1.2

0.6

c

0.4

35-

0.3

5 20

I

!

40

60

80

0,6"" 0.4

4.5=

20

r 40

20 40 60 80 r 60 80

20 4O

2.5 20

I

40

610

6O 80 80

Days after Planting Fig. 2. Lambsquarters biomass, stem diameter, and height (natural log scale) at four or five dates in 45 kg N ha- ~syntheticfertilizer ( o ) and crimson clover (e) treatments in Experiment 2. Insets illustrate N source× sampling date interactions, i.e. the relative difference between treatments (In N fertilizer value ] - In [crimsonclover valuel ) over the samplingdates. Nitrogensource× samplingdate interactions were not significant. corn was initially suppressed by the clover treatment, the inhibitory effect of the clover was more severe on lambsquarters growth than on that of sweet corn. At 2 weeks after emergence, e.g. reduction in biomass accumulation (averaged over 2 years) in the clover treatment was 72% for lambsquarters and 31% for sweet corn in comparison to lambsquarters and sweet corn in the fertilizer treatment. Reduction in growth also persisted through the growing season in clover-grown lambsquarters as compared to clover-grown sweet corn (Fig. 1). Continued suppression of lambsquarters growth in the clover treatment throughout the growing season was also seen in Experiment 2, which suggests that sweet corn interference was not responsible for the persistence of reduced lambsquarters growth in the clover treatment in Experiment 1. As a result of reduced interference from lambsquarters, in both years of Experiment 1 use of a crimson clover N source increased total sweet corn biomass grown under weed-infested conditions by 20% in comparison to use of 45 kg N h a - 1. In fact, biomass accumulation of weed-infested sweet corn grown with 180 kg N ha ~ did not exceed biomass accumulation of weed-infested, clover-grown sweet corn in either year (220 vs. 235 g m - 2 in 1989 and 581 vs. 635 g m -2 in 1990), a result which suggests that increased fertilization is not an efficient substitute for the weed suppression gained from the clover treatment.

3.4. Evidence as to the causes of the inhibitory effect of crimson clover residue The major symptoms of the inhibitory effect of crimson clover in both species were suppressed growth and development, particularly during the period between planting and the first sampling date (approximately two weeks after emergence). Although the experiments described above were not primarily designed to investigate mechanisms, soil sampling results in Experiment 1 provide some evidence as to the factor(s) responsible for the inhibitory effect of the clover treatment on early plant development. In terms of the hypothesis of temporal differences in N availability between legume and fertilizer N sources, soil nitrate levels were initially lower in the legume than the fertilizer treatment. Low initial N availability could help to explain the clover's suppression of early growth of sweet corn and lambsquarters, and, since nitrate is know to stimulate germination in a number of weed species, including lambsquarters (Williams and Harper, 1965; Karssen and Hilhorst, 1992), could also explain suppression of weed emergence. However, soil sampling results of Experiment 1 are not entirely consistent with the N availability hypothesis: in 1990, initial soil nitrate levels between the clover and the 0 N treatments were equivalent even though emergence and early growth of the lambsquarters were lower in the clover than the 0 N treatment (Tables 3, 5 and 6), a result which sug-

106

E. Dyck et al. / Agriculture, Ecosystemsand Environment 56 (1995) 93-108

gests that factors in addition to or other than low N availability contributed to the inhibitory effect of the clover treatment. The existence of a mechanism other than low nitrate availability is also suggested by the consistent lag in height development of clover-grown lambsquarters in Experiments 1 and 2 and the reduced height of clovergrown sweet corn in Experiment 1. Nitrogen deficiency results in general stunting of growth and therefore should have similarly affected biomass accumulation and stem diameter development. The lag in height development may indicate inhibition of cell division and elongation or interference with plant growth regulators, effects which have been produced by allelochemicals in bioassay studies (Einhellig, 1986). Results of plant tissue analysis for N concentration do not provide further evidence as to the cause(s) of the inhibitory effect of the clover treatment. Interpretation of results of tissue analysis is complicated by the fact that N concentration in plant tissue is affected by a number of factors in addition to external N supply, including plant growth (Glass and Siddiqi, 1984). Because N concentration decreases with biomass accumulation and because N addition stimulates growth, an increased N supply may result in decreased plant tissue N concentration (Greenwood et al., 1986). In this context, the cause of the higher N concentrations found in clover-grown than fertilizer-grown lambsquarters at final harvests in Experiment 1 (in 1990) and Experiment 2 remains ambiguous. 3.5. Conclusions and future research needs

The substantial reduction in weed interference with crop growth that resulted from substitution of legume green manure for synthetic N fertilizer in this study demonstrates that legume green manures have the potential to reduce the need for herbicide as well as synthetic N fertilizer applications in subsequent crops. However, the results of these experiments also indicate potential drawbacks to use of a legume N source, including possible inhibition of corn emergence (further investigated in an experiment reported in Dyck and Liebman, 1994). Additionally, use of crimson clover as an N source, while substantially reducing weed growth, did not entirely eliminate weed interference with crop growth nor did it prevent lambsquarters plants from setting seed by the end of the growing

season, a result that could lead to increased weed densities in a cropping system in subsequent years. Clearly, use of crimson clover needs to be considered as just one component of a weed management strategy to be used in conjunction with other weed-suppressing practices, including, e.g. cultivation and tillage, crop rotation, and use of weed-suppressive crop varieties. A major limitation of this study is that the effect of N source on crop-weed interference could not be assessed in terms of crop reproductive yield. This resulted from the mistaken assumption that springplanted crimson clover could emulate the performance of such overwintering legumes as alfalfa in developing sufficient early season biomass accumulation to allow early summer plowdown and subsequent successful production of a sweet corn crop. In areas with short growing seasons such as Maine, the effect of crimson clover vs. fertilizer N on weed-crop interference should be further tested in more agronomically feasible doublecropping systems, e.g. crimson clover followed by a cold-tolerant or transplanted vegetable crop. (The effect of crimson clover on crop-weed interference in a rotation system in which the clover is grown the year preceding a cash crop is reported in Dyck and Liebman, 1995) Further investigation is also needed to determine whether the results found in this study are species specific or are more generally applicable in other cropping systems, i.e. through the use of other legume species (particularly those with potentially high N fertilizer equivalency values), with other crops (including those with growth strategies differing from that of corn, e.g. small grains or fast-growing brassica species), and with other weeds (including other broadleaved species as well as annual and perennial grasses). Additional information also needs to be generated on how specific management practices contribute or detract from the weed-suppressive effect of crimson clover found in this study, e.g. the effect of residue placement (surface Or subsurface), amount of residue incorporated, depth of incorporation, and relative timing of legume incorporation and crop planting. Finally, determination of the mechanism(s) responsible for the weed-suppressive effect of crimson clover in this study could lead to better manipulation of cropping systems to maximize both N supply and weed control from use of a legume N source.

E. Dyck et al. /Agriculture, Ecosystems and Environment 56 (1995) 93-108 Acknowledgements

T h e a u t h o r s w o u l d like to t h a n k W . H a l t e m a n for statistical a d v i c e a n d G. D i c k e y , G. E l d r e d , J. G o n t o s k i , K. G e r o w , C. Grallert, J. Kline, E. L a n g h a m , E. M e l lander, K. M e l l a n d e r , B. M u r p h y , K. Sader, R. Stafford, a n d V. T r a c y for t h e i r i n v a l u a b l e h e l p w i t h the field w o r k r e q u i r e d for this study. T h i s w o r k w a s s u p p o r t e d by g r a n t s f r o m the J e s s i e S m i t h N o y e s F o u n d a t i o n ; the Maine Department of Agriculture, Food, and Rural R e s o u r c e s ; a n d the N a t i o n a l S c i e n c e F o u n d a t i o n ( D E B 9 1 2 2 8 3 3 ) a n d is c o n t r i b u t i o n 1914 o f the M a i n e A g r i c u l t u r a l a n d F o r e s t E x p e r i m e n t Station.

References Alkaemper, J., Pessios, E. and Long, D.V., 1979. Einfluss der duengung auf die entwicklung und naehrstoffnahme verschiedener unkrauter in mais. Proc. European Weed Res. Soc., pp. 181-192. Beckett, T.H., Stoller, E.W. and Wax, L.M., 1988. Interference of four annual weeds in corn (Zea mays). Weed Sci., 36: 764-769. Bradow, J.M. and Connick, W J., 1990. Volatile seed germination inhibitors from plant residues. J. Chem. Ecol., 16: 645--666. Carlson, H.L. and Hill, J.E., 1985. Wild oat (Avenafatua) competition with spring wheat: effects of nitrogen fertilization. Weed Sci., 34: 29-33. Dyck, E. and Liebman M., 1995. Crop-weed interference as influenced by a leguminous or synthetic fertilizer nitrogen source: II. Rotation experiments with crimson clover, field corn, and lambsquarters. Agric. Ecosystems Environ., 56: 109-120. Dyck, E. and Liebman, M., 1994. Soil fertility management as a factor in weed control: the effect of crimson clover residue, synthetic nitrogen fertilizer, and their interaction on emergence and early growth of lambsqnarters and sweet corn. Plant and Soil, 167: 227-237. Economic Research Service, 1993. Agricultural Chemical Usage: 1992 Field Crops Summary. USDA, Washington, D.C., Ag CH I (93). Einhellig, F.A.. 1986. Mechanisms and modes of action of allelochemicals. In: A.R. Putnam and C.-S. Tang (Editors), The Science of Allelopathy. Wiley, New York, pp. 171-188. Ellis, R.H.. Hong, T.D. and Roberts, E.H., 1985. Handbook of Seed Technology for Genebanks I: Principles and Methodology. International Board for Plant Genetic Resources, Rome, 210 pp. Erhardt, W.H. ( Editor ), 1991. New England Vegetable Management Guide. Cooperative Extension Service, University of Maine, Orono, 78 pp. Glass, A.D.M. and Siddiqi, M.Y., 1984. The control of nutrient uptake rates in relation to the inorganic composition of plants. Adv. Plant Nutr., 1: 103-147. Greenwood, D.J., Neeteson, J.J. and Draycott, A., 1986. Quantitative relationships for the dependence of growth rate of arable crops

107

on their nitrogen content, dry weight, and aerial environment. Plant and Soil, 91: 281-301. Hargrove, W.L., 1986. Winter legumes as a nitrogen source for notill grain sorghum. Agron. J., 78: 70-74. Haynes, B., Koide, R.T. and Elliot, G., 1991. Phosphorus uptake and utilization in wild and cultivated oats (Arena spp. ). J. Plant Nutr., 14: ll05-1118. Hesterman, O.B., 1988. Exploiting forage legumes for nitrogen contribution in cropping systems. In: W.L. Hargrove (Editor), Cropping Strategies for Efficient Use of Water and Nitrogen. Am. Soc. Agron., Madison, WI, pp. 155-166. Hesterman, O.B., Griffin, T.S., Williams, P.T., Harris, G.H. and Christenson, D.R., 1992. Forage legume-small grain intercrops: nitrogen production and response of subsequent corn. J. Prod. Agric., 5: 340-348. Hoffman, M.L., Regnier, E, and Cardina, J., 1993. Weed and corn (Zea mays) responses to a hairy vetch (Vicia villosa) cover crop. Weed Technol., 7: 594-599. Holderbaum, J.F., Decker, A.M., Meisinger, J.J., Mulford, F.R. and Vough, L.R., 1990. Fall-seeded legume cover crops for no-tillage corn in the humid east. Agron. J., 82: 117-124. Holm, L.G., Plucknett, D.L., Pancho, J.V. and Herberger, J.F.. 1977. The World's Worst Weeds: Distribution and Biology. University Press of Hawaii, Honolulu, 609 pp. Janke, R. and Peters, S., 1989. Cover crops, crop rotation, and weed control in sustainable cropping systems. In: Proc. Third PennJersey Tillage Conference, 22 February, Lehigh University, Bethlehem, PA, USA, pp. 1-8. Karssen, C.M. and Hilhorst, H.W.M., 1992. Effect of chemical environment on seed germination. In: M. Fenner (Editor), Seeds: The Ecology of Regeneration in Plant Communities. CAB International, Wallingford, UK, pp. 327-348. Knight, W.E. and Hoveland, C.S., 1985. Arrowleaf, crimson, and other annual clovers. In: M.E. Heath, R.F. Barnes and D.S. Metcalfe (Editors), Forages: The Science of Grassland Agriculture, 4th edn. Iowa State University Press, Ames, IA, pp. 136-145. Ladd, J.N., Amato, M., Jackson, R.B. and Butler, J.H.A., 1983. Utilization by wheat crops of nitrogen from legume residue decomposing from soils in the field. Soil Biol. Biochem., 15: 231-238. Lawson, H.M., and Wiseman, J.S., 1977. Competition between weeds and transplanted spring cabbage: Effects of nitrogen topdressing. Hortic. Res., 19: 25-34. Liebman, M., 1989. Effects of nitrogen fertilizer, irrigation, and crop genotype on canopy relations and yields of an intercrop/weed mixture. Field Crops Res., 22: 83-100. Liebman, M. and Dyck, E., 1993. Crop rotation and intercropping strategies for weed management. Ecol. Appl., 3: 92-122, McCalla, T.M. and Duley, F.L., 1948. Stubble mulch studies: Effect of sweetclover extract on corn germination. Science, 108: 163. Mueller, M.M., 1987. The Release and Fate of Clover Nitrogen in Soil. Hakapaino Oy, Helsinki, 53 pp. Nelson, W.A.. Kahn, B.A. and Roberts, B.W., 1991. Screening cover crops for use in conservation tillage systems for vegetables following spring plowing. HortScience, 26: 860-862. Okafor, L.I. and De Datta, S.K., 1976. Competition between upland rice and purple nutsedge for nitrogen, moisture, and light. Weed Sci., 24: 43-46.

108

E. Dyck et al. / Agriculture, Ecosystems and Environment 56 (1995) 93-108

Samson, R.A., 1991. The weed suppressing effect of cover crops. Proc. Fifth Annual REAP Conference, MacDonald College, SteAnne-de-Bellevue, Quebec, Canada, pp. 11-22. SAS, 1990. SAS/STAT User's Guide, Version 6, Fourth Edition. SAS Institute, Cary, NC, 1674 pp. Seibert, A.C. and Pearce, R.B., 1993. Growth analysis of weed and crop species with reference to seed weight. Weed Sci., 4 I: 5256. Smith, M.S., Frye, W.W. and Varco, J.J., 1987. Legume winter cover crops. Adv. Soil Sci., 7: 95-139. Varco, J.J., Frye, W.W., Smith, M.S. and MacKown, C.T., 1993. Tillage effects on legume decomposition and transformation of legume and fertilizer nitrogen- 15. Soil Sci. Soc. Am. J., 57: 750756.

Wall, L.L. and Gehrke, C.W., 1975. An automated total protein nitrogen method. J. Assoc. Official Anal. Chem., 58: 1221-1226. White, R.H., Worsham, A.D. and Blum, U., 1989. Allelopathic potential of legume debris and aqueous extracts. Weed Sci., 37: 674-679. Williams, J.T. and Harper, J.L., 1965. Seed polymorphism and germination: 1. The influence of nitrates and low temperatures on the germination of Chenopodium album. Weed Res., 5: 141-150. Wilson, D.O. and Hargrove, W.L., 1986. Release of nitrogen from crimson clover residue under two tillage systems. Soil Sci. Soc. Am. J., 50: 1251-1254.