HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Free Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in ISI Web of Science Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via ISI Web of Science (21) Citing Articles via Google Scholar Google Scholar Articles by Jannink, J.-L. Articles by Shaw, R.G. Search for Related Content PubMed Articles by Jannink, J.-L. Articles by Shaw, R.G. Agricola Articles by Jannink, J.-L. Articles by Shaw, R.G.

CROP ECOLOGY, PRODUCTION & MANAGEMENT

Index Selection for Weed Suppressive Ability in Soybean

^a 2101 Agronomy Hall, Iowa State Univ., Ames, IA 50011 USA
^b Dep. of Agronomy and Plant Genetics, Univ. of Minnesota, 1991 Buford Circle, 411 Borlaug Hall, St. Paul, MN 55108 USA
^c Dep. of Ecology, Evolution and Behavior, Univ. of Minnesota, 100 Ecology Building, 1987 Buford Circle, St. Paul, MN 55108 USA

jjannink{at}iastate.edu

	ABSTRACT

TOP NOTES ABSTRACT INTRODUCTION Materials and methods Results and discussion Conclusions REFERENCES

 
The economic and environmental costs of weed management in soybean (Glycine max [L.] Merr.) have led to interest in developing weed suppressive soybean varieties to enhance traditional herbicide and tillage-based approaches. We evaluated 104 inbred progeny from three crosses among elite soybean lines to determine optimal selection criteria for weed suppressive ability (WSA). We grew the lines in 1996 and 1997 at Becker, MN, an irrigated sandy site, and Rosemount, MN, a rainfed silt loam site, in a split-plot, with and without white mustard (Brassica hirta Moench). We measured soybean height 7 wk after emergence (WAE), light interception 5 and 7 WAE, specific leaf area 7 WAE, and date of full bloom. We harvested aboveground mustard biomass 8 WAE and calculated each soybean line's WSA as the difference between mustard biomass when grown in competition with that line and the overall mean mustard biomass. We estimated genetic correlations between soybean morphological traits, WSA, and the agronomic traits lodging, maturity date, and yield. Soybean early height's heritability and genetic correlation with WSA made it an ideal selection criterion. Indirect selection on height increased predicted selection efficiency by 70% relative to direct selection on mustard dry weight. Restricted index selection combining information on early height and lodging or yield eliminated undesirable correlated responses of lodging and yield while maintaining genetic gain for early height and WSA. Nevertheless, continuing rapid gains in agronomic performance while incorporating WSA may be difficult.

Abbreviations: ANOVA, analysis of variance • R2, date of full bloom • R8, date of physiological maturity • REML, residual maximum likelihood • RGR, relative growth rate • SLA, specific leaf area • WAE, weeks after emergence • WSA, weed suppressive ability

	INTRODUCTION

TOP NOTES ABSTRACT INTRODUCTION Materials and methods Results and discussion Conclusions REFERENCES

 
SOYBEAN PRODUCERS in the USA are highly dependent on herbicides for weed management. In 1997, 97% of all land planted to soybean received a herbicide treatment (USDA/NASS, 1998). Herbicide costs averaged U.S. $65 per hectare, equivalent to 35% of the variable cost of production, a percentage double that of corn, Zea mays L. (Hunst and Howse, 1997). Over 35 million kilograms of active ingredient were applied; of the four most commonly applied herbicides, imazethapyr (2-[4.5-dihydro-4-methyl-4-(1-methylethyl)-5-oxo-1H-imidazol-2-yl]-5-ethyl-3-pyridinecarboxylic acid) readily leaches while pendimethalin [N-(1-ethylpropyl)-2,6-dinitro-3,4-xylidine], glyphosate [N-(phosphonomethyl) glycine], and trifluralin (,,-trifluoro-2,6-Dinitro-N,N-dipropyl-p-toluidine) are prone to leaving the site of application through surface runoff, potentially adversely affecting water quality (USDA/NRCS, 1998). Besides herbicides, growers rely on cultivation as a weed management tool. Cultivation, however, disturbs the soil, aggravating erosion losses and increasing the sediment load of surface waters. These costs bring together farmers and environmentalists in calling for more diverse, ecologically sound options to manage weeds (Liebman and Dyck, 1993). Diverse options may be even more important if over-emphasis on a few management methods leads to shifts in weed species or evolution within weed populations toward resistance to these methods, reducing their efficiency (Jordan and Jannink, 1997).

Two traits inherent to the crop genotype contribute toward weed management: weed tolerance, the ability to maintain high yield despite weed competition, and weed suppressive ability (WSA), the ability to reduce weed growth through competition. Callaway (1992) documents genetic variability for both traits in numerous crop species and many authors suggest breeding to improve the traits (Garrity et al., 1992; Callaway and Forcella, 1993; Kropff and van Laar, 1993; Wortmann, 1993; Liebman and Gallandt, 1997; Bussan et al., 1997). Two arguments favor focussing breeding effort on WSA over weed tolerance to aid weed management (Jordan, 1993). First, suppressing weeds reduces weed seed production and benefits weed management in future growing seasons while tolerating weeds only benefits the current growing season. Second, weed pressure from unsuppressed weeds increases the likelihood of crop yield loss, irrespective of the crop's tolerance. For a given initial weed infestation, a weed suppressive genotype may prevent the rise of excessive weed pressure and thereby also confer within-season benefits.

Direct selection for WSA would entail growing each genotype in the presence of weeds to measure weed seed biomass or total biomass as a selection criterion. The labor requirements and high error of such measurements make indirect selection for WSA more attractive. The efficiency of using a trait as an indirect selection criterion to increase WSA depends on its heritability and genetic correlation to WSA through the following formula:

where r_{I, WSA} is the genetic correlation between the selection criterion, I, and WSA, and h²_I and h²_WSA are the heritabilities of I and WSA, respectively (Falconer, 1989). The literature documents relationships between several plant traits and competitive ability, in particular height (Jennings and Aquino, 1968; Kropff and van Laar, 1993, Chapter 8), various measures of leaf area or light interception (Garrity et al., 1992; Jordan, 1993; Wortmann, 1993; Forcella, 1987), specific leaf area (SLA) (Poorter, 1990), and maturity (Hinson and Hanson, 1962; Monks and Oliver, 1988). We lack, however, reports of these traits' heritabilities and genetic correlations to WSA, and therefore of their suitability as indirect selection criteria. Moreover, most previous studies measuring genetic variation in these traits have been conducted at only one location. Before engaging in breeding efforts to increase weed suppressive traits, we need evidence that they are not unduly affected by genotype x environment interaction.

Two further questions attend the inclusion of such selection criteria into a practical breeding program. First, the efficiency of index selection depends on the accuracy of the estimates of genetic parameters used in designing the index (Baker, 1986; Hayes and Hill, 1981; Sales and Hill, 1976). This problem applies not only to the population on which the estimates were made, but also to other populations on which the breeder may desire to impose selection but for which estimates are not available. That is, the utility of a selection index increases if evidence exists that it can be applied to new populations without undue loss of efficiency. Second, the desirability of selection for WSA depends on its relationship with other agronomically important traits. Index selection provides the flexibility to restrict correlated responses in some traits to zero while maximizing the response in the desired trait (Lin, 1978; Baker, 1986). In the context of restricted selection indices, desirability depends on whether predicted gains from selection under the imposed restrictions warrant the selection effort.

We conducted this study to determine the value of indirect selection for increasing WSA in a practical soybean breeding program. Our specific objectives were to identify soybean traits that provide efficient indirect selection criteria for WSA and to assess the extent of genotype x environment interaction affecting them. We evaluated how consistently different soybean breeding populations would respond to these selection criteria, and predicted correlated responses of agronomically important traits.

	Materials and methods

TOP NOTES ABSTRACT INTRODUCTION Materials and methods Results and discussion Conclusions REFERENCES

 
Germplasm, Planting, and Management
In 1992, Bussan et al. (1997) measured elite breeding lines for canopy area. Parent lines M89-1006 and M90-1682 showed low canopy area while M89-792, M88-250, and M89-1946 showed high canopy area. In addition, M90-1682 carried the dtl allele for determinate growth habit (Bernard, 1972). We calculated coefficients of coancestry among parents using methods and pedigrees given in Allen and Bhardwaj (1987). We advanced progeny from three crosses, M89-1006 x M89-792, M90-1682 x M89-792, and M88-250 x M89-1946 (denoted Cross 1, Cross 2, and Cross 3 below) by modified single-seed descent to the F4 stage. We threshed 35 randomly selected individual plants from each cross in 1995, and sent seed to La Platina, Chile, for seed increase. Seed from one line of Cross 1 was lost during the return trip.

In 1996 and 1997, we planted the 104 progeny lines, five parental lines, and two checks, cv. Evans and cv. Kato, in a sets-in-reps design with three replicates at two locations, Becker, MN (irrigated Hubbard loamy sand: sandy mixed Udorthentic Haploborall) and Rosemount, MN (Waukegan silt loam: fine silty mixed mesic Typic Hapludoll). Each of seven sets contained 16 lines: five lines from each cross and either a parent or a check. Dates for field management tasks are given in Table 1 . We applied fertilizer prior to planting at Becker but not at Rosemount. Soil samples were taken to a depth of 0.2 m, oven dried at 35 °C and analyzed for available nitrate at the Univ. of Minnesota soil testing lab.

We thinned mustard plants back to one per position. We counted the number of soybean plants within 0.1 and 0.2 m of each mustard and scored mustard plants for initial size on a scale of 1 (about 30 mm in diameter) to 7 (about 200 mm in diameter). About 8 WAE, mustard plants were harvested individually, oven dried at 60°C, and weighed. Though some mustard plants were missing, all plots retained at least three plants in 1996 and at least five plants in 1997.

The following soybean traits were measured in the weed-free subplots on dates given in Table 2 . Early height was measured as the distance from the ground to the base of the central leaflet on the highest expanded leaf of three random plants. Early and mid light interception were measured using a LI-COR model LI-191 SA line quantum sensor (LI-COR Inc., Lincoln, NE¹) at the base of the canopy and a LI-COR model LI-190 SA point sensor above the canopy that integrated photosynthetic photon flux density over a 3-s period. Four and three measurements were taken per plot for early light interception and mid light interception, respectively. Specific leaf area was measured by sampling 10 central leaflets from the seventh or eighth nodes experiencing full sun per plot. The total area of the 10 leaflets was measured with a LI-COR LI-3100 area meter, the leaflets were then dried and weighed and SLA calculated as the quotient of area by weight. Date of full bloom (R2 stage) was observed by visiting each location biweekly starting in the second week of July. Plots were noted as entering the R2 stage when >50% of plants carried an open flower in one of the two upper-most nodes (Fehr et al., 1971). Date of physiological maturity (R8 stage) was observed by visiting each location weekly, starting in the second week of September. Plots were noted as entering the R8 stage when 95% of the plants carried only mature-color pods. After the R8 stage, lodging was scored on each plot on a scale of 1 (no lodging, erect) to 5 (plants completely prostrate). Yield was harvested on an end-trimmed 2.4-m length of plot. Seed was oven dried overnight at 35 °C in cloth bags then weighed.

Analysis of Variance
Preliminary analysis indicated that for many variables measured, year x location effects were greater than zero. Given the lack of simple year and location effects, the two years and locations were grouped in further analyses as environments. Environment, block, and line within cross effects were considered random, set and cross effects were considered fixed. Exact F tests for the significance of environment, population, and environment x population interactions are not available, given their expected mean squares (Steel and Torrie, 1980, p. 357). These effects were therefore tested using approximate F tests. Parents and checks were analyzed in a separate analysis including only effects for environment, block, line, and their interactions.

Maximum Likelihood
To estimate genetic (line) and error covariance components, we performed residual maximum likelihood (REML) as implemented in programs described by Shaw (1987, 1991) and Shaw and Shaw (1998) that iteratively solve REML estimation equations. We modified the programs to allow for comparisons among >2 populations and for estimation of line x environment variance components. In the estimation of variance components, environment, block, set, cross, and their interactions were considered fixed effects. Broad-sense genetic (line), line x environment, and error components were considered random. The programs also provided asymptotic error matrices of the covariance parameters.

In a first analysis, we estimated covariances among six traits related to soybean weed suppressive ability: mustard suppression, early height, early light interception, mid light interception, date of full bloom, and specific leaf area. To increase the power and generality of this analysis, we used data pooled over the three crosses. In a second analysis, we estimated covariances among mustard suppression, early height, and three agronomic traits: lodging score, date of physiological maturity (R8), and yield. Because of the computer memory requirements of these analyses, we were limited to analyzing four traits at a time. To assemble covariance matrices containing six (first analysis) and five (second analysis) traits, we analyzed all possible sets of four traits and averaged the results. We determined the probability that a variance or covariance equaled zero using its asymptotic standard error rather than by a likelihood ratio test (Shaw, 1987).

In the second analysis, we first estimated separate covariances for the three crosses for all of the random effects. Likelihood ratio tests, however, revealed no significant differences in error covariances among crosses (data not shown). In subsequent analyses, we constrained the error covariances to equality across the three populations. Thus, in a single analysis, we obtained pooled estimates of error covariance but separate line and line x environment covariances for each cross. These covariances were used to design selection indices as described below.

We note that the REML equations may yield solutions that do not lie in the parameter space for (co) variance estimates (e.g., negative variances or correlations outside of [-1,1]). They are called unfeasible estimates. Hill and Thompson (1978) show that unfeasible estimates of multiple trait covariance matrices are to be expected. We used the cutting plane algorithm (Shaw and Geyer, 1997) to constrain our estimates to feasibility.

Selection Indices
We designed restricted selection indices using formulae given in Walsh and Lynch (1999)(Chapter 19). Restricted indices maximize gain for a desired trait (in our case, WSA) while restricting correlated responses in other traits to zero (Tallis, 1962; Lin, 1978). We calculated index coefficients using the genetic (line), line x environment, and error covariance matrices assuming that selection would be conducted on the basis of an evaluation in one environment with lines replicated three times. We predicted response to direct selection on mustard dry weight, indirect selection on early height, or index selection on early height while restricting the correlated response of lodging score, R8, or yield to zero.

We obtained confidence intervals for the predicted response using the asymptotic parametric bootstrap procedure developed by Shaw and Geyer (1997). This procedure uses the Fisher information matrix, i.e., the matrix of second derivatives of the log likelihood function with respect to the parameters to be estimated. This matrix provides the asymptotic variances and covariances of the gradient of the log likelihood with respect to the parameters (Searle et al., 1993). Using this property, we parametrically bootstrapped the gradient and maximized the log likelihood with respect to each sample gradient. The solution of each maximization could be constrained to feasibility and considered as a candidate parameter set. For each parameter set in turn, we calculated the response to selection using the index of interest and formed a distribution of the predicted responses. We derived confidence intervals from these distributions.

	Results and discussion

TOP NOTES ABSTRACT INTRODUCTION Materials and methods Results and discussion Conclusions REFERENCES

 
Early season conditions differed substantially between the two years. Precipitation between 15 April and 30 May was 138 and 98 mm in 1996 and 45 and 64 mm in 1997 at Becker and Rosemount, respectively. Excessive moisture delayed planting in 1996 and, at Becker, caused nitrogen to leach down the sandy soil: available nitrate to a depth of 0.2 m was 3.2, 17.7, 11.4, 10.4 µg g^-1 in the four environments in the order given above. Presumably because these conditions differentially affected the two locations, preliminary analyses showed that year, location, and year x location interactions affected early season measurements related to WSA (data not shown). For example, mustard dry weights averaged 2.1, 18.7, 15.1, and 10.3 g in the four environments in the order given above, suggesting that mustard responded strongly to nitrogen availability (Liebman, 1989) rather than to year or location main effects. Because environment effects were not clearly interpretable as year or location effects, we considered the year and locations combinations as environments in further analyses.

Analysis of variance (ANOVA) results in Table 3 show that, besides strong environment effects, crosses differed for early height and R2, and environment x cross interactions affected mustard dry weight, mid light interception, and R2. Patterns of environment x cross interaction were difficult to explain (data not shown). Within-cross genetic variation affected all traits and was little influenced by environment: when significant line x environment variance was detected (i.e., for early height and R2), it was a small fraction of the line variance as shown by the F ratios given in Table 3. This result suggests that lines selected for WSA will show broad environmental adaptability for this trait, an important result for the use of WSA as a reliable component of weed management (Liebman and Gallandt, 1997).

View this table: [in this window] [in a new window]   Table 5 Heritabilities (genetic matrix diagonal), error proportion of variance (error matrix diagonal), genetic and error correlations between traits related to weed suppressive ability: mustard dry weight (DW), early height, early and mid light interception, date of full bloom (R2) and specific leaf area (SLA). The final column shows the predicted efficiency of improving soybean WSA by indirect selection on the trait relative to direct selection on mustard dry weight with the 5% worst-case lower bound given in parentheses

Several aspects of the error correlations are noteworthy (Table 5). Strong positive error correlations arose between soybean height and light interception indicating that although these traits are not genetically related, they do respond jointly to micro-environmental variation. Lesser but significant error correlations were found between these growth traits and SLA: in favorable environments for growth, soybeans produced larger, thinner leaves. Finally, positive and marginally significant (P < 0.10) error correlations arose between mustard dry weight and soybean growth traits suggesting that micro-environmental variation affected soybean and mustard similarly. This observation cautions against the simple use of phenotypic correlations in the place of genetic correlations when assessing the relationship between traits. Because of the large error associated with mustard dry weight and its positive error correlation with soybean growth traits, phenotypic correlations between mustard dry weight and soybean height, early and mid light interception, -0.22, -0.03, and -0.01, respectively are small in magnitude. Estimating only phenotypic correlations among traits might lead to the conclusion that the traits are marginally relevant to weed suppressive ability (e.g., Bussan et al., 1997).

On the basis of the predicted efficiency of indirect selection for decreased mustard dry weight using early height (Table 5), we explored further the agronomic consequences of using this trait as a selection criterion within each cross. Among the three crosses, we note that in Cross 3, we do not find evidence for line variation in weed suppressive ability: neither the heritability of mustard dry weight nor any of the genetic correlations between it and other traits are significant (Table 6) . Cross 2 differed from the others in its genetic correlations between early height and yield and mustard suppression and yield (Table 6). The magnitude and direction of these correlations more strongly suggests a genetic tradeoff between weed suppressive ability and yield in this cross than in the others. Cross 2 segregated for the dtl allele that confers determinate growth habit and subsequently pleiotropically affects final height and yield (Bernard, 1972; Saindon et al., 1990). We did not, however, find evidence for pleiotropic effects of dtl on either early height or mustard suppression (data not shown). The cause of the strong negative relationship between yield and early height and mustard suppression in this cross remains unclear.

View this table: [in this window] [in a new window]   Table 6 Heritabilities (italic) of mustard suppression (Sup.) and early height, and their genetic correlations with the agronomic traits lodging, date of maturity (R8) and yield measures in 1997

	Conclusions

TOP NOTES ABSTRACT INTRODUCTION Materials and methods Results and discussion Conclusions REFERENCES

 
Within the array of approaches available to implement integrated weed management, the competitive suppression of weeds by crops can make several small but cumulative contributions (Liebman and Gallandt, 1997). Weed suppression is preventative in that it decreases the weed seed rain and therefore tends to reduce weed infestations in subsequent years (Jordan, 1993). When used in conjunction with herbicide or cultivation, it may allow the farmer to decrease application rates or frequencies, (Christensen, 1994) or the number of cultivation passes. While competitive suppression will rarely kill weeds outright, results from this study showing small or inexistant line x environment interaction affecting traits related to WSA suggest that it will act reliably across environments. Moreover, competitive suppression can function independently of weather conditions that might hinder the application of other management practices.

With respect to the genetic improvement of soybean WSA, we found that within-cross genetic variation exists for early-growth traits that affect soybean's ability to suppress weeds. Soybean height 6 to 7 wk after emergence showed moderately high heritability, strong genetic correlation to WSA, and was quick and simple to measure. Predicted responses per cycle of selection show that gains in mustard suppression may result from selecting on early height. These characteristics make it an ideal indirect selection criterion, particularly in a practical soybean breeding program where labor needs at the time to measure early height are not as high as in the spring or the fall.

With respect to the consequences on soybean agronomic traits of selecting on early height, our results suggest the following broader interpretations. The relationship between early height and agronomically important traits differed among the three crosses studied. Particularly Cross 2 showed a strong negative correlation between early height and yield. This result indicates that it may be important to examine such relationships on a cross-by-cross basis. Nevertheless, we were always able to design selection indices that provided gains in early height and WSA while restricting the correlated response in yield to zero. For both Cross 1 and Cross 2 that were most variable for early height, predicted correlated responses to selection using all proposed indices gave some indication of a tradeoff in agronomic performance for improved WSA. When changes in maturity were restricted to zero, negative (though non-significant) correlated responses in yield were predicted. When changes in yield were restricted to zero, positive (though non-significant) correlated responses in maturity were predicted. Though we cannot make general inferences from this single study, these results suggest that continuing rapid gains in agronomic performance while also incorporating WSA into soybean may be difficult. Just as certain varieties are singled out for their disease or insect resistance, so we may need to single out varieties for their WSA and count on recovering the cost of lost agronomic performance through improved weed management.1998; 1998

	ACKNOWLEDGMENTS

 
This research was conducted with funds from the United States Department of Agriculture USDA/95/31315-2092 and from the Minnesota Soybean Research and Promotion Council. The senior author was supported by a National Science Foundation Graduate Research Fellowship during part of the study. We thank Frank Shaw for statistical guidance and Sheri Huerd, Phil Schaus, Darryld Oistad, Eric Boelke, Alvaro Rivera, Art Killam, Joshua Larson, Dana Blumenthal, Emily Pullins, Kathryn Hamilton, and Heather Laflin for field assistance. Jeff Gunsolus, Steve Simmons, and Deon Stuthman gave valuable comments on this manuscript.

	NOTES

TOP NOTES ABSTRACT INTRODUCTION Materials and methods Results and discussion Conclusions REFERENCES

 
Contribution No. 00-13-0146 of the Minnesota Agric. Exp. Stn., Univ. of Minnesota, St. Paul, MN.

¹ Mention of any product is for scientific purposes only and does not imply endorsement by the Minnesota Agricultural Experiment Station.

Received for publication July 15, 1999.

	REFERENCES

TOP NOTES ABSTRACT INTRODUCTION Materials and methods Results and discussion Conclusions REFERENCES

 

• Allen, F.L., and H.L. Bhardwaj. 1987. Genetic relationships and selected pedigree diagrams of North American soybean cultivars. Tenn. Agric. Exp. Stn. Bull. 652.
• Baker R.J. Selection indices in plant breeding. Boca Raton, FL: CRC Press, Inc, 1986.
• Bernard R.L. Two genes affecting stem termination in soybeans. Crop Sci. 1972;12:235-239.[Abstract/Free Full Text]
• Bussan A.J., Burnside O.C., Orf J.H., Ristau E.A., Puettmann K.J. Field evaluation of soybean (Glycine max) genotypes for weed competitiveness. Weed Sci. 1997;45:31-37.
• Callaway M.B. A compendium of crop varietal tolerance to weeds. Am. J. Alternative Agric. 1992;7:169-180.
• Callaway M.B., Forcella F. Crop tolerance to weeds. In: Francis C.A., Callaway M.B., eds. Crop improvement for sustainable agriculture. Lincoln: Univ. Nebraska Press, 1993:100-131.
• Christensen S. Crop weed competition and herbicide performance in cereal species and varieties. Weed Res. 1994;34:29-36.
• Falconer D.S. Introduction to quantitative genetics, 3rd ed Essex, England: Longman Scientific and Technical, 1989.
• Fehr W.R., Caviness C.E., Burmood D.T., Pennington J.S. Stage of development descriptions for soybeans, Glycine max (L.) Merrill. Crop Sci. 1971;2:929-931.
• Forcella F. Tolerance of weed competition associated with high leaf area expansion rate in tall fescue. Crop Sci. 1987;27:146-147.[Abstract/Free Full Text]
• Garrity D.P., Movillon M., Moody K. Differential weed suppression ability in upland rice cultivars. Agron. J. 1992;84:586-591.[Abstract/Free Full Text]
• Gizlice Z., Carter T.E., Jr., Burton J.W. Genetic diversity in North American soybean: I. Multivariate analysis of founding stock and relation to coefficient of parentage. Crop Sci. 1993;33:614-620.[Abstract/Free Full Text]
• Harper J.E. Nitrogen metabolism. In: Wilcox J.R., ed. Soybean: Improvement, production, and uses. Madison, WI: ASA, CSSA, and SSSA, 1987:497-533.
• Hayes J.F., Hill W.G. Modification of estimates of parameters in the construction of genetic selection indices (`Bending'). Biometrics 1981;37:483-493.
• Hill W.G., Thompson R. Probabilities of non-positive definite between-group or genetic covariance matrices. Biometrics 1978;34:429-439.
• Hinson K., Hanson W.D. Competition studies on soybeans. Crop Sci. 1962;2:117-123.[Free Full Text]
• Hunst, M., and G. Howse. 1997. Minnesota agricultural statistics. Minn. Agric. Stat. Service. St. Paul, MN.
• Hunt R., Cornelissen J.H.C. Components of relative growth rate and their interrelations in 59 plant species. New Phytol. 1997;135:395-417.
• Jennings P., Aquino R.C. Studies on competition in rice. III. The mechanism of competition among phenotypes. Evolution 1968;22:529-542.
• Jordan N. Prospects for weed control through interference. Ecol. Appl. 1993;3:84-91.
• Jordan N.R., Jannink J.-L. Assessing the practical importance of weed evolution: A research agenda. Weed Res. 1997;37:237-246.
• Kropff M.J., van Laar H.H. Modelling crop-weed interactions. Wallingford, UK: CAB International, 1993.
• Li B., Suzuki J.-I., Hara T. Latitudinal variation in plant size and relative growth rate in Arabidopsis thaliana. Oecologia 1998;115:293-301.[ISI]
• Liebman M. Effects of nitrogen fertilizer, irrigation, and crop genotype on canopy relations and yields of an intercrop/weed mixture. Field Crops Res. 1989;22:83-100.
• Liebman M., Dyck E. Weed management: A need to develop ecological approaches. Ecol. Appl. 1993;3:39-41.
• Liebman M., Gallandt E. Many little hammers: ecological approaches for management of crop-weed interactions. In: Jackson L.E., ed. Ecology in Agriculture. San Diego: Academic Press, 1997:291-343.
• Lin C.Y. Index selection for genetic improvement of quantitative characters. Theor. Appl. Genet. 1978;52:49-56.
• Lynch M., Walsh B. Genetics and analysis of quantitative traits. Sunderland, MA: Sinauer Assoc, 1997.
• Marrow P., Johnstone R.A., Hurst L.D. Riding the evolutionary streetcar: where population genetics and game theory meet. Trends Ecol. Evol. 1996;11:445-446.
• McWhorter C.G., Hartwig E.E. Competition of johnsongrass and cocklebur with six soybean varieties. Weed Sci. 1972;20:56-59.
• Monks D.W., Oliver L.R. Interactions between soybean (Glycine max) cultivars and selected weeds. Weed Sci. 1988;36:770-774.
• Poorter H. Interspecific variation in relative growth rate: on ecological causes and physiological consequences. In: Lambers H., et al. , ed. Causes and consequences of variation in growth and productivity of higher plants. The Hague: SPB Academic Publishers, 1990:45-68.
• Reich P.B., Walters M.B., Ellsworth D.S. From tropics to tundra: Global convergence in plant functioning. Proc. Natl Acad. Sci. (USA) 1997;94:13730-13734.[Abstract/Free Full Text]
• Saindon G., Voldeng H.D., Beversdorf W.D. Adjusting the phenology of determinate soybean segregants grown at high latitude. Crop Sci. 1990;30:516-521.[Abstract/Free Full Text]
• Sales J., Hill W.G. Effect of sampling errors on efficiency of selection indices. 2. Use of information on associated traits for improvement of a single important trait. Anim. Prod. 1976;23:1-14.
• Searle S.R., Casella G., McCulloch C.E. Variance components. New York: John Wiley and Sons, 1993.
• Shaw F.H., Geyer C.J. Estimation and testing in constrained covariance component models. Biometrika 1997;84:95-102.[Abstract/Free Full Text]
• Shaw R.G. Maximum-likelihood approaches applied to quantitative genetics of natural populations. Evolution 1987;41:812-826.[ISI]
• Shaw R.G. The comparison of quantitative genetic parameters between populations. Evolution 1991;45:143-151.
• Shaw, R.G., and F.H. Shaw. 1998. Quercus: programs for quantitative genetic analysis using maximum likelihood. Available by anonymous ftp from ftp://evolution.umn.edu/pub/quercus/quercus/ (verified March 6, 2000).
• Steel, R.G.D., and J.H. Torrie. 1980. Principles and procedures of statistics. A biometrical approach. 2nd ed. McGraw-Hill, New York.
• Tallis G.W. A selection index for optimum genotype. Biometrics 1962;18:120-122.
• USDA/NASS 1998. Agricultural chemical usage. USDA, National Agricultural Statistics Service. U.S. Gov. Print. Office, Washington, DC.
• USDA/NRCS 1998. Selected pesticide properties database and SPISP ratings. USDA, Natural Resources Conservation Service. U.S. Gov. Print. Office, Washington, DC.
• van der Werf A., Nuenen M.V., Visser A.J., Lambers H. Contribution of physiological and morphological plant traits to a species' competitive ability at high and low nitrogen supply. Oecologia 1993;94:434-440.
• Villar R., Veneklaas E.J., Jordano P., Lambers H. Relative growth rate and biomass accumulation in 20 Aegilops species. New Phytol. 1998;140:425-437.
• Walsh, B., and M. Lynch. 1999. Selection and evolution of quantitative traits. http://nitro.biosci.arizona.edu/zbook/volume_2/vol2.html (verified March 6, 2000).
• Wortmann C.S. Contribution of bean morphological characteristics to weed suppression. Agron. J. 1993;85:840-843.[Abstract/Free Full Text]

This article has been cited by other articles:

				W. Yan and J. Fregeau-Reid Breeding Line Selection Based on Multiple Traits Crop Sci., March 19, 2008; 48(2): 417 - 423. [Abstract] [Full Text] [PDF]

				M. M. Williams II and J. L. Lindquist Influence of Planting Date and Weed Interference on Sweet Corn Growth and Development Agron. J., June 5, 2007; 99(4): 1066 - 1072. [Abstract] [Full Text] [PDF]

				D. L. Zhao, G. N. Atlin, L. Bastiaans, and J. H. J. Spiertz Cultivar Weed-Competitiveness in Aerobic Rice: Heritability, Correlated Traits, and the Potential for Indirect Selection in Weed-Free Environments Crop Sci., January 24, 2006; 46(1): 372 - 380. [Abstract] [Full Text] [PDF]

This Article Abstract Full Text (PDF) Free Alert me when this article is cited Alert me if a correction is posted Services Similar articles in this journal Similar articles in ISI Web of Science Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via ISI Web of Science (21) Citing Articles via Google Scholar Google Scholar Articles by Jannink, J.-L. Articles by Shaw, R.G. Search for Related Content PubMed Articles by Jannink, J.-L. Articles by Shaw, R.G. Agricola Articles by Jannink, J.-L. Articles by Shaw, R.G.