See

discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/233954589

Tillage Effects on Weed Seed Return and

Seedbank Composition

Article in Weed Science May 1996

CITATIONS READS

141 211

4 authors, including:

David R. Clements Stephen D. Murphy

Trinity Western University University of Waterloo
61 PUBLICATIONS 1,183 CITATIONS 19 PUBLICATIONS 675 CITATIONS

SEE PROFILE SEE PROFILE

Clarence J Swanton
University of Guelph
244 PUBLICATIONS 6,712 CITATIONS

SEE PROFILE

All content following this page was uploaded by Clarence J Swanton on 30 July 2015.

The user has requested enhancement of the downloaded file. All in-text references underlined in blue are added to the original document
and are linked to publications on ResearchGate, letting you access and read them immediately.
Weed Science Society of America

Tillage Effects on Weed Seed Return and Seedbank Composition

Author(s): David R. Clements, Diane L. Benott, Stephen D. Murphy and Clarence J. Swanton
Reviewed work(s):
Source: Weed Science, Vol. 44, No. 2 (Apr. - Jun., 1996), pp. 314-322
Published by: Weed Science Society of America and Allen Press
Stable URL: http://www.jstor.org/stable/4045684 .
Accessed: 21/12/2012 14:28

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at .
http://www.jstor.org/page/info/about/policies/terms.jsp

.
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of
content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms
of scholarship. For more information about JSTOR, please contact support@jstor.org.

Weed Science Society of America and Allen Press are collaborating with JSTOR to digitize, preserve and
extend access to Weed Science.

http://www.jstor.org

This content downloaded on Fri, 21 Dec 2012 14:28:08 PM

All use subject to JSTOR Terms and Conditions
Weed Science, 1996. Volume44:314-322
TillageEffectson WeedSeed Returnand SeedbankCompositionI
DAVIDR. CLEMENTS,DIANE L. BENOIT,STEPHEND. MURPHY,and CLARENCEJ. SWANTON2
Abstract.Weedseed returnand seedbankcomposition,with
particularreferenceto commonlambsquarters, werestudied
in four tillage systems establishedon a site near Fingal,
Ontario.Thetillagetreatmentsweremoldboardplow,chisel
plow,ridge-till,and no-till.Thecroppingsystemwas a corn-
soybeanrotation.Tillageeffectson weedpopulationcompo-
sitionwereassessedafterall weedcontrolmeasureshadbeen
implemented.More than 60% of the weed seedbankwas
concentratedin the upper 5 cm of soil in chisel plow and
no-till. The seedbankof the moldboardplow system was
more uniformlydistributedover depth and largerthan the
other systems. Common lambsquarterscomprised more
than50%of the seedbankin all systemsexceptridge-till,but
only dominatedthe abovegroundweed populationin chisel
plow.Seedbankpopulationsof commonlambsquarterswith
moldboardplowingwere greaterthan those with ridge-till
and no-till, and chisel plow seedbank populationswere
greaterthan those in ridge-till.Chiseland moldboardplow
systemsgenerallyhadhigherabovegroundplantpopulations
of commonlambsquartersthan the othertwo systems.Seed
productionperplantby commonlambsquarterswasequiva-
lent amongthe four systems,but estimatedseedproduction
per unit areawas higherin moldboardplowandchiselplow
systemsthan in the other systems.Populationsof common
lambsquartersand similarspeciesmay producemoreseeds
andpersistin moldboardplowandchiselplowsystems;these
weedsmay producefewerseedsper unit area and be easier
to managein no-till and ridge-tillsystems.Nomenclature:
Common lambsquarters, Chenopodiumalbum L. #3
CHEAL;corn,ZeamaysL.; soybean,Glycinemax(L.)Merr.
Additionalindexwords:Conservationtillage, no-till, ridge-
till, seedbankprofile,weedspectrum,CHEAL.
INTRODUCTION
Availableevidence demonstratesthat adoptionof conserva-
tion tillage systems resultsin alteredweed species composition
and abundance (5, 9, 17, 21, 28, 30). Because adoption of
conservation tillage represents a reduction in agroecosystem
disturbance,these changes in weed communities representa
form of ecological succession (28). It has been estimatedthat
yield, weed populations,and soil characteristicsdo not reach
'Received for publicationJune9, 1994, andin revisedformAugust27, 1995.
2Asst.Prof.,Dep. Biology, TrinityWesternUniversity,Langley,BC, Canada,
V2Y lYl, Weed Sci., Agric. CanadaRes. Stn., Saint-Jean-sur-Richelieu,
QUE,
CanadaJ3B 3E6; Res. Assoc., and Assoc. Prof., Dep. CropSci., Univ. Guelph,
Guelph,ON, CanadaNIG 2W1.
3Lettersfollowing this symbol are a WSSA approvedcomputercode from
Composite List of Weeds, Revised 1989. Available from WSSA, 1508 West
UniversityAve., Champaign,IL, 61821-3133.
314
This content downloaded on Fri, 21 Dec 2012 14:28:08 PM
All use subject to JSTOR Terms and Conditions
equilibriumuntil the managementregime has been established
for 4 to 10 yr (10, 18). Therefore,studiesareneededto assess the
effects of managementon weed populations.
To understandand managethese successionalchanges more
effectively,thereis a needfor dataon processessuchas how weed
dispersalchangesduringlong time periods,the populationecol-
ogy of abovegroundweeds, and how the resultantweed seed-
banks will be affected (28). In conventionaltillage, many weed
species rely on a single regenerativestrategyconsisting of ger-
minationfroma large,persistentsoil seedbank.Partof the reason
the seedbank is persistent is that most seeds are buried by
moldboardplowing (1, 5, 8, 36). For many species, buriedseeds
remain viable and will germinate in subsequent years when
returnedto a suitabledepthby tillage. In contrast,seeds remain
on or nearthe soil surfacein undisturbedsoils, i.e., with conser-
vation tillage (1, 5, 24, 27, 36). Adequate abovegroundweed
control may cause seeds present in the upper layer of soil to
diminish within a few years (19). Thus, species that rely on
regenerationfrom seedbanksand historically have been prob-
lematic in conventional tillage systems may become less of a
problemin conservationtillage systems (7). Despite its potential
importancein determiningrelativeweed pressuresamongdiffer-
ent tillage regimes, relatively few studies have accountedfor
verticaldistributionof weed seeds in the soil (24).
Common lambsquartersrelies on regenerationfrom seed-
banksand often flourishesin conventionaltillage systems. Seed
productionmay range from 30 000 to 176 000 seeds per plant
for common lambsquartersif shielded from herbicide applica-
tions in soybean (12). Commonlambsquartersseeds buriedfor
20 yr still had 23% viability (22), and buried common lamb-
squartersseeds have survived 39 yr (31). Worldwide,common
lambsquartershas been among the most importantweeds in
many crops, and was rankedseventhin abundancein corn (20).
In soybean, 2 common lambsquartersm-l of row for 5 wk of
interferenceaftercrop emergencecauseda 5%yield loss (12). A
recent survey of corn, soybean, and winter wheat (Triticum
aestivumL.) fields in Ontariolisted common lambsquartersas
the second most abundantweed (16). Derksen et al. (9) found
that common lambsquarterswas more common in conventional
tillage systems than in minimum or no-till systems in western
Canada;in Ontario,common lambsquartersmainly was associ-
ated with minimumtillage (30).
Althoughcommonlambsquartersgenerallycan be controlled
with herbicides,its persistentseedbankmay cause management
difficultiesif it escapes treatmentsor developsresistance(2, 12).
Weed seedbanksare an indicator,i.e., they representthe cumu-
lative effects of many years of managementbut allow sampling
within a single year.We were interestedin observingchanges in
the populationsof weed escapes from actual managementre-
gimes, ratherthantryingto extrapolateresultsfromuncontrolled
 WEED SCIENCE

Table1. Herbicideuse duringthe studyperiod.Herbicideswere appliedat ratesrecommendedfor Ontarioin no-till andridge-tillsystems. See text for additionalyear-
and plot-specificherbicidetreatments.
Year Crop Herbicideapplieda Applicationmethod

1986 Soybean Metolachlor/metribuzin PRE

1987 Corn Metolachlor/bentazon+ atrazine POST
1988 Soybean Metolachlor/metribuzin PRE
1989 Corn Metolachlor+ atrazine PRE
1990 Corn Metolachlor/2,4-D/dicamba+ atrazine PRE/PRE/POST
1991 Soybean Glyphosate/2,4-D/imazethapyr PRE/PRE/POST
1992 Corn Metolachlor/2,4-D PRE/POST
1993 Soybean Imazethapyr POST

aMetolachlor:2-chloro-N-(2-ethyl-6-methylphenyl)-N-(2-methoxy-1-methylethyl)acetamide;metribuzin:4-amino-6-(1,1-dimethylethyl)-3-(methylthio)-1,2,4-
triazin-5(4H)-one;bentazon:3-(1-methylethyl)-(1H)-2,1,3-benzothiadiazin-4(3H)-one
2,2-dioxide; atrazine:6-chloro-N-ethyl-N'-(1-methylethyl)-1,3,5-triazine-2,4-
diamine; 2,4-D: (2,4-dichlorophenoxy)acetic acid; dicamba, 3,6-dichloro-2-methoxybenzoic acid; glyphosate: N-(phosphonomethyl)glycine;imazethapyr:
2-[4,5-dihydro-4-methyl-(1-methylethyl)-5-oxo-lH-imidazol-2-y]-3-pyridinecarboxylic
acid.

weeds (32). To accomplishthis,we usedon-farmstudies.Advan-
tages of on-farmresearchincludea relativelylargesamplingarea
and the ability to monitor long-term effects of actual farmer
practices;this is importantin conservationtillage becauseof the
4-10 yr requiredfor the system to attainequilibrium(10, 18).
Effectively,long-termon-farmstudies such as ours allow repli-
cation over time. The disadvantagesinclude a lack of spatial
replicationwithin and across sites.
The objectiveof this studywas to compareweed seed return,
seedbankcomposition,andverticaldistributionof weed seeds in
the soil amongfourtillage systemsestablishedfor eight seasons:
moldboardplow,chisel plow,ridge-till,andno-till.We were able
to examinecommonlambsquartersin moredetailbecauseit was
the most prevalentweed in the seedbanksof most fields studied.
To examine the effect of tillage on potentialweed seed return,
we analyzedthe effect of tillage, in corn and soybeanfields, on
reproductiveallocation for different size classes of common
lambsquarters.
ANDMETHODS
MATERIALS
Study sites and agronomic practices. A long-termtillage study
was initiatedon a privatefarmin 1985 nearFingalin southwest-
ernOntario,Canada.The field was a Beverley silt loam soil with
36% sand, 45% silt, and 19% clay. The soil organic matter
content was 3.2% and pH was 7.3. The field was divided into
four plots each 23 m wide and 200 m long. Each plot was
managedwith one of four differenttillage systems: 1) conven-
tional tillage with an autumn moldboardplow and a spring
discing, 2) conservationtillage using an autumnchisel plow and
a springdiscing,3) ridgetillageusing a 'Buffaloridgecultivator'
in the springandin mid-Julyfor ridge rebuildingand4) a no-till
system wherea modified 'JohnDeere 7000'4 no-till drill seeder
was used. The same seederwas used in all tillage systems, with
adjustmentsmadeto accommodateseedingrequirementsof each
tillage system. Mean percent residue cover (? SE) during the
study periodwas 5.2% (?3.6%) for conventionaltillage, 33.0%
4JohnDeere Co., Moline IL.
Volume44, Issue 2 (April-June)1996
(?19.2%) for chisel plow, 45.9% (?17.6%) for ridge-till, and
68.6% (?15.4%) for no-till.
The croppingsystem was a corn-soybeanrotation,except for
1989 to 1991 where corn followed corn. Corn was seeded at
68,000 seeds ha-' in 75 cm wide rows in 1987, 1989, 1990, and
1992. The hybrids used were Pioneer 3737 (1987), Northrup
King 3624 (1989 and 1990), andPioneer3751 (1992). Soybean
was seeded at 375,000 seeds ha-1in 1986, 1988, 1991, and 1993.
The cultivarsused were Elgin (1986 and 1988), Pioneer 9161
(1991), andNorthrupKing S20-20 (1993). Agronomicpractices
and herbicide treatmentswere consistent with those recom-
mendedfor field crops in Ontario.Therefore,all assessmentsof
weed populations were for weeds that escaped control treat-
ments. Herbicideuse across all tillage systems duringthe eight
years of our study is presented in Table 1. In addition, the
co-operatorsalso used 2,4-D (2,4-(dichlorophenoxy)aceticacid)
andglyphosate[N-(phosphonomethyl)glycine] in some cases. In
1986, 1987, 1992, and 1993 the no-till plot was sprayedwith
glyphosate as a burn-downtreatment,2 wk before planting. In
1988 and 1993, glyphosatewas appliedin the ridge-tillplot. In
1988, 2,4-D was appliedin the no-till plot. In 1990, 1991, and
1992, 2,4-D was appliedpostemergenceon all plots exceptchisel
plow. Average machine-harvestedyields duringthe trial in the
moldboardplow stripwere 9500 and 2900 kg ha-' for corn and
soybean, respectively.Yields in the other systems were within
10%of the moldboardplow yields, but were higherin the case
of ridge-tillsoybean.
Seedbankanalysis.Soil samplesweretakenon May 10, 1992,
afterseedbedpreparationbut priorto seedingand herbicide
application.One hundredfifteen cm- by 1.9-cmcores were
sampled from each of the four tillage strips. The cores were
collected every 4 m along each of the two middlecroprows. Half
of the ridge-till samples was from the ridge; the otherhalf was
sampledwithinthe furrow.We used a corerfittedwith an acetate
sleeve to divide the sample into 0-5, 5-10, and 10-15 cm
segments(1, 5). Eachsegmentwas placedseparatelyin 3-cm
deepcontainerslinedwithpeat.Thecontainers wereplacedin a
= =
growthroom(T 25 C;photoperiod 16:8 light:dark)on June
1 1992,andwatereddaily(5). Seedlingemergencewas moni-
toredover a periodof 6 mo andthe soil was stirredonce per
315

This content downloaded on Fri, 21 Dec 2012 14:28:08 PM

All use subject to JSTOR Terms and Conditions
 CLEMENTSET AL.: TILLAGEEFFECTSON WEED SEED RETURNAND SEEDBANKCOMPOSITION

Table2. Occurrenceof weeds in the seedbankand abovegroundamong the four tillage systems in 1992 and 1993.

Tillage systemsa

Code Scientific name Commonname MP CP RT NT

CHEAL ChenopodiumalbumL. Commonlambsquarters S-Ab S-A S-A S-A

ABUTH AbuthilontheophrastiMedic. Velvetleaf S-A S-A A
PLAMA Plantago majorL. Broadleafplantain S-A S-A S S-A
AMBEL AmbrosiaartemisiifoliaL. Commonragweed S-A S-A S-A A
SETVI Setaria viridis (L.) Beauv. Greenfoxtail S A S S-A
SENVU Senecio vulgarisL. Commongroundsel S S
PHYHE Physalis heterophyllaNees Clammygroundcherry - S S A
TAROF TaraxaxumofficinaleWeberin Wiggers Dandelion A S-A A A
CERVU CerastiumvulgatumL. Mouseearchickweed S-A
SOOCA Solidago canadensisL. Canadagoldenrod S
FRAVI Fragaria virginianaMill./Duch. Wild strawberry - S
ECHCG Echinochloacrus-galli (L.) Beauv. Bamyardgrass A _ S
AGRRE Elytrigiarepens(L.) Nevski Quackgrass A A A A
ASCSY Asclepias syriaca L. Commonmilkweed A A A
CYPES CyperusesculentusL. Yellow nutsedge A A A
POLPE Polygonumpersicaria L. Ladysthumb A A A
SINAR Sinapsis arvensisL. Wild mustard A A
AMARE AmaranthusretroflexusL. Redrootpigweed - A
PANDI PanicumdichotomiflorumMichx. Fall panicum A
PTLRC Potentilla recta L. Sulfurcinquefoil A
CIRAR Cirsiumarvense (L.) Scop. Canadathistle A A
ERIAN Erigeronannuus(L.) Pers. Annualfleabane A
OXAST OxalisstrictaL. Yellow woodsorrel A
RUBID Rubusidaeus L. Europeanred raspberry - A A
MEDLU Medicago lupulinaL. Black medic A A
DAUCA Daucus carota L. Wild carrot - A A
DIGIS Digitaria ischaenum(Schreib.ex. Schweig.) Smooth crabgrass - A
Shreb.ex Muhl.
SETLU Setariaglauca(L.) Beauv. Yellow foxtail - A
DACGL DactylisglomerataL. Orchardgrass - - A
PRTQU Parthenocissusquinquefolia(L.) Planch. Virginiacreeper A
SONOL SonchusoleraceusL. Annualsowthistle - - - A
aMP= moldboardplow, CP chisel plow, RT = ridge-till,NT = no-till.
bPresence/absence where S = seedbank,A = aboveground,S-A = presentin both the seedbankand aboveground.

month (3). This was followed by a period of 3 mo at -4 C in
darkness(beginning1 October),and3 additionalmo of monitor-
ing germinationin the growthroom (3). Seedlings were identi-
fied by species.
Weed counts and sampling. Weed density was evaluated on
September9, 1992, and September10, 1993, using 0.75 by 0.3
m quadratsplaced every 4 m. Weeds were counted in 100
quadratsin each tillage system. On October7, 1992, and Sep-
tember30, 1993, commonlambsquarterswas harvestedto deter-
mine seed production. At this date, common lambsquarters
developmentwas at a stage where few seeds had shattered.The
harvestedplantswere dividedinto two size classes:less thanand
greaterthan30 cm tall. Foreach tillage treatment,10 plantsfrom
each size class were randomlyselected.Because of the relatively
low density of common lambsquartersin some tillage systems,
it would have been impossible to locate many more than 10
plants within a given size class, and thus it was felt that the
sample was representativeof the population.Plant height was
measuredat harvest.All harvestedplantswere air driedat 80 C
for 3 d.
Following drying,commonlambsquarterswas separatedinto
inflorescenceand vegetative componentsand weighed. Repro-
ductive allocationwas calculatedas the dry weight of the inflo-
rescence dividedby the sum of inflorescenceandvegetativedry
weight. Forplantsover 30 cm tall, the numberof seeds produced
was estimatedby dividingthe total seed weight by the weight of
a 100 seed subsampleweighed for each plant.
Analysis. Analyses of variance(ANOVAs)were used to test the
effect of tillage practiceused on the above-groundweed popu-
lations, seedbanks,vertical distributionof seeds in the soil and
the reproductiveeffort and seed productionin lambsquarters.
Means were separatedby LSD(O05). Weed density data were
log-transformedto adjustfor heterogeneityof variance.Vertical
distributionof seeds was expressed as a proportionof the total
seed bank.
RESULTSANDDISCUSSION
Weed spectrum. Over the four tillage systems, three weed
species appearedonly in soil samples, 19 were observedexclu-
sively in the abovegroundsampling and 12 appearedin both
samples (Table 2). All of the species observed only in above-
groundsamples were relativelyuncommonat our site; hence, it

316 Volume44, Issue 2 (April-June) 1996

This content downloaded on Fri, 21 Dec 2012 14:28:08 PM

All use subject to JSTOR Terms and Conditions
 WEED SCIENCE
is not surprisingthat these species did not appearin the soil
samples.
In 1992, the moldboardplow system had the lowest number
of weed species, with 8 abovegroundand 5 in the seedbank;
however,in 1993 there were 11 species identifiedaboveground
in moldboardplowed fields. In no-till, 16 were recordedabove-
groundin 1992 and 10 in 1993. In a separatestudyset up on part
of the same site (differentherbicidetreatments),Benoit et al. (3)
recordedthe lowest numberof species in the moldboardplow
system (five species abovegroundand in the soil), and recorded
thatthe highestnumberof species [19] occurredabovegroundin
theno-till system.Thetrendtowardgreaterspecies diversitywith
conservationtillage is supportedby ecological theory that pre-
dicts greaterspecies diversity at intermediatelevels of distur-
bance (7, 28).
Species identifiedin the moldboardplow system were found
in all systems [e.g., common lambsquarters,common ragweed
(AmbrosiaaretemisiifoliaL.), and quackgrass(Elytrigiarepens
(L.) Nevski)], whereasotherspecies wererestrictedto one or two
of the conservationtillage systems (Table2). Dandelion(Tarax-
acum officinaleWeberin Wiggers)was foundin all systems and
was most abundantin the least disturbedsystems, ridge-till and
no-till. Six species were unique to no-till, three were unique to
ridge-till,andthreewerefoundonly in chisel plow (Table2). The
abovegrounddensities of these unique species never exceeded
0.5 plantsm-2.
Vertical distribution of seeds. The verticaldistributionof seeds
in the soil variedamongtillage systems (Figure 1). The top 5 cm
of soil contained37 and 33% of the seedbankin moldboardand
ridge-till systems, respectively; in no-till and chisel plow sys-
tems, the top 5 cm contained most of the seeds (74 and 61%,
respectively).In ridge-tillsystems, the highest concentrationof
seeds (45%) was in the 5- to 10-cm layer.The 5 to 10 cm layer
containedthe highest concentrationof seeds on both the ridge
Soil Depth
(cm)
0 700 seeds m-2in the top 10 cm of soil in ridge-tilledor moldboard
37% 61% 33% 74%
5 years after establishmentof differenttillage practices. We de-
25% 23 45% 9 reflect a greaterinfestationof weeds with long-lived seeds.
10_
38% 6 22 18
15
Moldboard Chisel Ridge-till No-till
Plow Plow
Figure 1. Effect of tillage methodson the verticaldistributionof total weed seed
aftersix growingseasons in a corn-soybeanrotation.LSD (0.05) for chisel plow
= 20%, no-till = 25%, moldboardplow and ridge-tillwere non-significant.
Volume44, Issue 2 (April-June) 1996
This content downloaded on Fri, 21 Dec 2012 14:28:08 PM
All use subject to JSTOR Terms and Conditions
4,000
* Commonlambsquarters
I Remainingspecies
3,000 -
a
E
j2,000 a
b
ab bb
1,000 bc
0
Moldboard Chisel Ridge-till No-till
plow plow
Figure 2. Seedbank populationsin the four tillage systems in 1992, for total
weeds and common lambsquarters.Bars with the same letter are not different
accordingto the LSD test; lettersrefer to the barimmediatelybelow them (i.e.,
upperletterrefers to total weeds, lower letterrefersto common lambsquarters).
and in the furrow.The top 5 cm contained33% of the seedbank
on the ridge; 27% of the seedbankwas in the furrow.
Differencesin the verticaldistributionof seeds betweenmold-
boardplow tillage andchisel or no-till systemsis consistentwith
other reports (1, 5, 36). Published informationon the vertical
distributionof seeds in ridge-till is limited (15), and further
studiesareneededto explainthe patternsobserved.The patterns
may relateto "scalping"of the top layer of soil on the ridges by
the ridge-till cultivator.
Weed seedbanks. Moldboard plowed fields had the largest
seedbank,2667 seeds m-2 in the top 15 cm; this was greater(P
> 0.05) than the seedbanks of the other three tillage systems
(Figure 2). The seedbank population in the top 10 cm in the
moldboardplow seedbankpopulationwas approximately1500
m-2. In comparison,ForcellaandLindstrom(15) detected200 to
plowed fields. Their study is comparableto ours in that a corn-
soybean rotationwas used and the fields were sampled 7 to 8
tected more seeds in moldboardplowed fields and this may
For ridge-tillfields, we detected 800 weed seeds m-2; this is
above the range(200 to 700 seeds m2) detectedby Forcellaand
Lindstrom (15) in their ridge-till fields. They suggested that
layby cultivation, used to build-up ridges in mid-season, can
stimulategerminationof weeds; if not controlled,more weeds
may set seed in ridge-till than conventional tillage systems in
continuouscorn. In our study,the significantdifferencein seed-
bank populationsbetween moldboardplowed and ridge-tilled
fields may suggest that the weeds were controlled and this
reducedweed seed return.Additionally,it may indicatethatthe
layby cultivation in our study occurred after the peak weed
317
 CLEMENTSET AL.: TILLAGEEFFECTSON WEED SEED RETURNAND SEEDBANKCOMPOSITION
germinationperiod.As a consequence,weed seeds or seedlings
may become desiccated.
Timing of cultivation also affected weed germinationand
seedling survivorshipin other conservationtillage systems (4,
11). Desiccationof weed seeds or otherfactorssuch as increased
seed predationon the numerousseeds thatareclose to the surface
in chisel plow and no-till fields may explainreducedweed seed
returnand smallerseedbanks.
Above-ground weed populations. Above-groundweed popu-
lationswere rankeddifferentlythanthe seedbankpopulations.In
1992, no-till and chisel plow systems had higherweed densities
than the other two systems (Figure 3). In 1993, this ranking
changed, in orderof decreasingweed density, to no-till, mold-
boardplow andridge-till,andchisel plow (Figure4). The differ-
ence in rankingby tillage practice between the above-ground
populations and seedbanks can be attributedto the increased
densityof perennialweeds, (e.g., quackgrass)in no-till systems.
These weeds may be better adaptedto less-disturbedenviron-
ments and tend to reproduceasexually ratherthan producing
seeds thatwill be returnedand detectedin the seed bank.Addi-
tionally,therewere relativelylargedandelionpopulations(0.9 to
12.3 mi-2) presentin no-till and ridge-till systems that inflated
total weed density, particularlyin 1993. The agamospermous
dandeliontends to produceseeds that withstanddesiccation in
no-till and ridge-till fields but the seeds do not persist long in
seedbanks(35).
There was a shift in weed density in 1993. Weed density in
moldboardplowed andridge-tilledfields increaseddramatically
(Figures3 and 4). Some of this was relatedto increasedinfesta-
tions of lambsquarters(moldboardplow) and dandelion (ridge
till) but densityof all otherspecies also increased(Figures3 and
4). It is possible thatthis increasewas relatedto weathercondi-
tions thatwere moreconduciveto weed genminationandgrowth
20
* Commonlambsquarters
FLDandelion
15 1XRemaining species
E no-till systems (Figure2). Therewere morelambsquarters'seeds
X 10
a chisel plow vs. ridge-tillsystems (Figure2).
-
5 no-till andridge-till,whereasdensitiesrangedas high as 5.6 m-2
..
.. _b.......
Moldboard Chisel Ridge-till No-till
plow plow
Figure 3. Abovegroundpopulationsin the four tillage systems in 1992, for total
weeds, dandelion,and common lambsquarters.Bars with the same letterare not
differentaccordingto the LSD test; letters refer to the bar immediatelybelow
them (i.e., upper letter refers to total weeds, lower letter refers to common
lambsquarters).
318
This content downloaded on Fri, 21 Dec 2012 14:28:08 PM
All use subject to JSTOR Terms and Conditions
20
U Common lambsquarters a
1 Dandelion
el Remaining species
15 -
b~~~~~~
. _
t~~~~~~ ... .....,Wf
1fo.{
Moldboard Chisel Ridge-till No-till
Plow Plow
Figure 4. Abovegroundpopulationsin the four tillage systems in 1993, for total
weeds, dandelion,and common lambsquarters.Bars with the same letterarenot
differentaccordingto the LSD test; letters refer to the bar immediatelybelow
them (i.e., upper letter refers to total weeds, lower letter refers to common
lambsquarters).
in 1993, i.e., it was much warmerin 1993. This is supportedby
the increasedweed densities that occurredin no-till; however,
weed density decreased in chisel plowed systems. It is also
possible that there were herbicide failures in all but the chisel
plow systems in 1993 but it seems unlikelythatthis would occur
consistently across all fields. The most likely factor was the
changefrom corn in 1992 to soybeansin 1993. Soybeanshave a
smallercanopy and are less effective competitorsfor light. As a
consequence,more weed seedlings reachedmaturityand weed
density increased.
Commonlambsquarterspopulations.Commonlambsquarters
was the dominantweed in the seedbankfor chisel plow (69%of
total seeds), no-till (74%), and moldboardplow (76%) systems;
in ridge-till, lambsquarterscomprised 35% of the seedbank
(Figure2). In 1992, the numberof commonlambsquartersseeds
in the seedbank did not differ between moldboardand chisel
plow systems, chisel plow and no-till systems or ridge-till and
in moldboardplow than in ridge-till and no-till systems and
Common lambsquartersdensities did not exceed 1.1 m-2 in
in the chisel plow system (Figures 3 and 4). In the chisel plow
system, common lambsquartersformed 67 and 46% of the
abovegroundweed density in 1992 and 1993, respectively.By
contrast, common lambsquarterswas less than 50% of weed
density in the other three systems. Over 1992-1993, common
lambsquarterscomprisedless than22, 15, and 16%of the above-
groundweed populationsobserved in moldboardplow, no-till,
and ridge-till systems, respectively. Common lambsquarters
plants were most abundantin the chisel plow system in 1992
Volume44, Issue 2 (April-June) 1996
 WEED SCIENCE

Table3. Reproductiveallocationand seed productionper plantby height classes of common lambsquarterscollected in corn on 7 October 1992a.

Vegetative Inflorescence Reproductive Total Weight Seed

Tillage system biomass weight allocationb seed weight per seed number

mg mg
Below 30 cm size class
Moldboardplow 234 a 72 a 0.35 ab 12 a 0.8 a 17 a
Chisel plow 96 a 43 a 0.26 b 9a 0.4 c 18 a
Ridge-till 114 a 85 a 0.40 a 32 a 0.5 bc 52 a
No-till 106 a 62 a 0.38 a 18 a 0.7 ab 28 a

Above 30 cm size class

Moldboardplow 472 b 272 b 0.35 ab 130 b 0.7 ab 205 b
Chisel plow 487 b 289 b 0.35 ab 147 b 0.7 a 234 b
Ridge-till 1090 ab 383 b 0.36 a 148 b 0.6 ab 234 b
No-till 1510 a 921 a 0.26 b 386 a 0.5 b 828 a

aMeanswithin a columnin a size class followed by the same letterare not significantlyat the 0.05 level as determinedby LSD (0.05).
bReproductiveallocation= inflorescenceweight/(inflorescenceweight + vegetative biomass).

(Figure 3) and in both the chisel plow and moldboardplow
systems in 1993 (Figure4).
In their study on the same site from 1986 to 1989, Benoit et
al. (3) observed that common lambsquarterscomprised 47 to
81%of the seedbank,wherethe highestpercentageswere in the
no-till andchisel plow systems.If herbicideinputswerereduced,
or if controlby preemergenceherbicideswas ineffective, com-
mon lambsquarterswas among the most commonly observed
weeds (3). Regardlessof the method,controlof common lamb-
squarterswas mostdifficultin the chisel plow system(3). Among
the fields surveyedby FrickandThomas(16) (about200 for each
system), common lambsquartersdensity averaged 1.4, 3.7, and
3.3 m72 for conventional,chisel plow, and no-till fields, respec-
tively.
Reproductive allocation of common lambsquarters. In gen-
eral, tillage system did not affect reproductiveallocation of
common lambsquarters(Tables 3 and 4). Similarly there was
little effect of tillage on seed productionperplant.One exception
was higherseed productionin no-tillrelativeto othersystemsfor
common lambsquarterstaller than 30 cm in 1992 (Table 3).
Commonlambsquarterstallerthan30 cm had greatervegetative
biomass than common lambsquartersin moldboardplow and
chisel plow in 1992, and had greaterinflorescenceweight, total
seed weight,andseed numberperplantthancommonlambsquar-
tersin the otherthreesystems (Table3). Reproductiveallocation
of the > 30 cm plants in 1992 was relatively low in the no-till
system, and was less than that for the ridge-till system. Weight
per seed was not correlatedwith plant size although in 1992
plants 30 cm in the no-till plot had relativelylow-weight seeds.
In 1993, the only parameterthat variedwith tillage system was
weight per seed, but there were no consistent trends in seed
weight across years (Tables3 and 4).
Because a single weed may producelarge numbersof seeds
(26), we examined the contributionto seed productionby the
few, largerplantsin 1992 (0.007 m2 in ridge-tillandmoldboard
plow systems,0.022 m72in the chisel plow systemand0.033 m2
in no-till). Estimatedseed productionby commonlambsquarters
in no-till was 58 m2, with nearlyhalf of this coming fromplants
in the largersize class (Table5). In other systems, particularly
chisel plow, the smaller plants contributedmost of the seed

Table4. Reproductiveallocationand seed productionby height classes of common lambsquarterscollected in soybeanson 30 September1993a.

Vegetative Inflorescence Reproductive Total Weight Seed

Tillage system biomass weight allocationb seed weight per seed number

mg mg

Below 30 cm size class

Moldboardplow 324 a 188 a 0.32 a 60 a 0.6 a 104 a
Chisel plow 311 a 104 a 0.26 a 10a 0.4 b 33 a
Ridge-till 401 a 273 a 0.34 a 110 a' 0.5 a 212 a
No-till 286 a 125 a 0.33 a 28 a 0.6 a 46 a

Above 30 cm size class

Moldboardplow 4856 ab 5126 a 0.47 a 3014 a 0.6 ab 5403 a
Chisel plow 3731 b 3721 a 0.40 a 1928 a 0.6 a 3598 a
Ridge-till 4542 ab 3587 a 0.45 a 1672 a 0.6 ab 2932 a
No-till 8170a 4330a 0.37a 1975 a 0.5b 3986a

aMeanswithin a column in a size class followed by the same letterare not differentat the 0.05 level has determinedby LSD (0.05).
bReproductive allocation= inflorescenceweight/(inflorescenceweight + vegetative biomass).

Volume44, Issue 2 (April-June) 1996 319

This content downloaded on Fri, 21 Dec 2012 14:28:08 PM

All use subject to JSTOR Terms and Conditions
 CLEMENTSET AL.: TILLAGEEFFECTSON WEED SEED RETURNAND SEEDBANKCOMPOSITION
Table 5. Common lambsquartersseed productionestimates for four tillage that our methodof assessmentunderestimatedseed production,
systemsa.
the reducedseed productionlikely reflectedthe combinedeffect
Seed production of crop competition and reduced plant size due to herbicide
Below Above %of applicationsundermanagedconditions(32). Longer-termstud-
30 cm 30 cm seedbank ies dedicated to weed seed returnto seedbanks are needed to
size size Total popula- clarify this issue.
Tillage system Year class class seeds tion
Common lambsquarterspopulationsamong tillage prac-
no.m-2 tices. Throughthree phases of the life-cycle: seedbank,above-
Moldboardplow 1992 2 2 4 <1 ground plant, and reproductionper area, no-till, and ridge-till
1993 54 11,238 11,292 557 systems exhibitedlower populationsof common lambsquarters
Chisel plow 1992 101 5 106 9 than moldboardplow and chisel plow systems. Differences in
1993 41 4,498 4,539 383
Ridge-till 1992 31 2 33 9
seed productionper area were not a function of differences in
1993 42 880 922 242 seed productionper plant, but rather were related to higher
No-till 1992 31 27 58 8 abovegrounddensities. Similarly, differences in aboveground
1993 2 638 640 91 densities likely were relatedto differencesin seedbankpopula-
aSeedproductioncomputedfrom informationon seed productionper plant tions.
in Tables 3 and 4, and weed density from Figure 3 and Figure 4. In 1992 the The problemsassociatedwith large weed seedbanksare best
density of plants above 30 cm was extremelylow, and density estimatestaken
from the October harvest were used in the calculation of seed productionby
illustratedby the moldboardplow plot. In 1992, the moldboard
above 30 cm plants,whereasin 1993, approximateratios of below to above 30 plow plant populationswere relatively low; however, in 1993,
cm plants in October were used to estimate the relative contributionsto seed therewere largenumbersof common lambsquartersin the same
productionusing density informationfrom Figure4. plot and seed productionwas relatively high. Tillage practices
that disturbthe soil increasethe amountof environmentalvari-
ation; in years where environmentalconditions are favorable,
seed germinationandproductionmaybe high. Moreimportantly,
productionin 1992. In 1993, when largeplantsexhibitedhigher
this means thatyear-to-yearvariationin weed density and weed
density andhigherseed production,the majorityof seeds (95 to
seed return will increase and lead to less predictable weed
100%)were producedby the above 30 cm size class. Harrison
problems. The reductionin seedbanks in no-till and ridge-till
(12) also determinedthat seed productionincreasedwith plant
size and that growth of common lambsquartersvaried greatly may be caused by shallowerburialof seeds underconservation
with environmentalconditions(12). tillage systems, differences in herbicide use and management
In 1992, the estimated seed productionby common lamb- among systems, and differencesin weed growth and mortality
squartersin all systemscomprisedless than 10%of the common resultingfrom differencesin tillage practices.
lambsquartersseeds accumulatedfrom previous years in the In our study and in others (1, 5, 8, 36), seeds were concen-
seedbank(Table5). In 1993, a warmer,drieryear, seed produc- tratedin the upper5 cm of soil in conservationtillage systems.
tion per area was much higher in all systems, especially in Theoretically,this factor alone would increase, ratherthan de-
moldboardplow andchisel plow plots (Table5). Moldboardand crease, weed populationsbecause of increasedemergencefrom
chisel plow plots hadgreaterpopulationsof commonlambsquar- shallowerdepths.Weaveret al. (33) recordedhigheremergence
tersthanno-till andridge-tillplots (Figure4). The estimatedseed of common lambsquartersfor seeds sown on the surface than
productionper unit area was nearly 6 and 4 times the 1992 seeds buried1.5 cm deep. In our study,the herbicidesapparently
seedbankpopulationsfor moldboardplow and chisel plow sys- controlledweeds emerging in the conservationtillage systems
tems, respectively.Estimated seed productionper area in the suchthatover time seedbankswere reducedin comparisonto the
no-till plot was only 91% of the 1992 seedbank;thus, during moldboardplow seedbank.This may indicatethatmanagement
1992 and 1993, the common lambsquartersseedbankmay have was effective enough to preventexcessive seed production(24).
been declining.Theno-till seedbankwas muchsmaller(one third Populationpatternsof commonlambsquartersin the chisel plow
the size) than the moldboardplow seedbank(Figure 2). In the system were similar to those observed in the moldboardplow
ridge-tillplot, estimatedseed productionwas approximately2.5 system, despite the shallow burialof seeds in chisel plow. One
times the seedbankin 1992. However,the ridge-tillplot also had differencebetweenthe chisel plow system andthe otherconser-
the smallestcommonlambsquartersseedbank(Figure2); at 386 vation tillage systems was the lack of glyphosate or 2,4-D
m-2, this was about one fifth the size of the moldboardplow treatmentsin the chisel plow plot. This could affect common
seedbank. lambsquarterspopulationsover time because a relatively large
Seed productionper plantin both yearswas ordersof magni- proportionof individualsmay emerge priorto planting(13).
tude below the range of 30,000 to 176,000 Harrison(12) re- Timinganddegreeof soil disturbancemay determinewhether
cordedfor unsprayedcommonlambsquarters. This may indicate more or less weed seed emergence is stimulated or whether
thatcommon lambsquartersescaping herbicidetreatmentspose tillage effects interact with herbicide effects. Common lamb-
a reducedthreatto futurecontrol,even at the densitiesobserved squartersgerminates at cooler temperaturesthan most weed
in the chisel and moldboardplow systems. While it is possible species (6, 33, 34), andthuscontrolmeasuresthatact earliermay

320 Volume44, Issue 2 (April-June) 1996

This content downloaded on Fri, 21 Dec 2012 14:28:08 PM

All use subject to JSTOR Terms and Conditions
 WEED SCIENCE
be more effective in reducing populations. Alternatively,the
potentialdelay in emergencewith lower soil temperaturesunder
conservationtillage conditions may be a factor (14). Further
studies are needed to examine the interactionof tillage systems
with timingof emergence.Even thoughthe chisel plow does not
buryseeds as deeply as the moldboardplow, therestill is a higher
amountof disturbancerelative to no-till or ridge-till;this may
result in higher levels of emergence or seedling survival. The
additionaldisturbanceto constructridges also may stimulate
weed seed germination(15). Althoughseed productionalso may
increasein ridge-tillfields (15), this did not occur in our study.
The lack of soil disturbancein no-till may resultin mortalityof
weed seeds for species such as common lambsquarters,if they
arevulnerableto desiccation,seed predationor otherfactorsthat
seeds are exposed to on the soil surface(24, 28).
Theamountsof residueleft overwinteron the soil in this study
ranked, from lowest to highest, was moldboardplow, chisel
plow,ridge-till,andno-till.Residueremainingon thesoil surface
may affect weed population dynamics, e.g. interactionswith
herbicideapplications,allelopathy,and shadingeffects andmay
resultin reducedweed seedling emergence(23, 25, 29).
Populationsof common lambsquartersand similar species
that persist in conventional tillage systems may be managed
more readily in reduced tillage systems, as indicated by this
on-farmresearchwhere actualmanagementpracticeswere em-
ployed over 8 yr.However,given thattheseresultswereobtained
from a single site, furtherinformationon the long-termeffect of
tillage on common lambsquarterspopulationswould be useful.
This would expeditethe developmentof managementstrategies
to reduce populations of weeds with seedbank regeneration
strategiesand to managethe successionaldynamicsof weeds in
conservationtillage (28).
ACKNOWLEDGMENTS
WethankHughMartinof the OntarioMinistryof Agriculture
and Food and the farmoperatorsAl and Paul Jones and Dennis
Zettlerfor cooperatingin the Tillage 2000 project.Kevin Chan-
dler provided technical assistance. Douglas Buhler and two
anonymous reviewers provided helpful comments on earlier
versionsof this manuscript.
CITED
LITERATURE
1. Ball, D.A. 1992. Weed seedbankresponse to tillage, herbicides,and crop
rotationsequence.Weed Sci. 40:654-659.
2. Bandeen,J. D., G. R. Stephenson,and E. R. Cowett. 1982. Discovery and
distributionof herbicide-tolerantweeds in NorthAmerica.Pages 9-30 in H.
M. LeBaronandJ. Gressel,eds. HerbicideResistancein Plants.JohnWiley
& Sons, Toronto,ON.
3. Benoit,D. L., C. J. Swanton,K. Chandler,andD. A. Derksen.1991. Changes
in weed populationsand seed bankthroughtwo cycles of a maize-soyabean
rotationin Ontario,Canada.Proc. Brit. Crop Prot. Conf.-Weeds-1991.
1:403-410.
4. Buhler, D. D., J. L. Gunsolus, and D. F. Ralston. 1992. Integratedweed
managementtechniquesto reduce herbicide inputs in soybean. Agron. J.
84:973-978.
5. Cardina,J., E. Regnier,and K. Harrison.1991. Long-termtillage effects on
seed banksin threeOhio soils. Weed Sci. 39:186-194.
Volume44, Issue 2 (April-June) 1996
This content downloaded on Fri, 21 Dec 2012 14:28:08 PM
All use subject to JSTOR Terms and Conditions
6. Chu, C. C., P. M. Ludford,J. L. Ozbun, and R. D. Sweet. 1978. Effects of
temperatureand competitionon the establishmentand growth of redroot
pigweed and common lambsquarters.Crop. Sci. 18:308-3 10.
7. Clements, D. R., S. F..Weise, and C. J. Swanton. 1994. Integratedweed
managementand weed species diversity.Phytoprotection75:1-18.
8. Cousens, R. and S. R. Moss. 1990. A model of the effects of cultivationon
the verticaldistributionof weed seeds within the soil. WeedRes. 30:61-70.
9. Derksen, D. A., G. P. Lafond, A. G. Thomas, H. A. Loeppky, and C. J.
Swanton. 1993. Impact of agronomic practices on weed communities:
tillage systems. Weed Sci. 41:409-417.
10. Dick, W. A. andT. C. Daniel. 1987. Soil chemical and biological properties
as affected by conservation tillage: Environmentalimplications. Pages
125-147 in T. J. Logan, J. M. Davidson, J. L. Baker,and M. R. Overcash,
eds. Effects of ConservationTillage on GroundwaterQuality.Lewis Pub-
lishers, Inc., Chelsea, MI.
11. Eadie, A. G., C. J. Swanton,J. E. Shaw,andG. W. Anderson.1992. Banded
herbicideapplicationsandcultivationin a modified no-till corn (Zea mays)
system. WeedTechnol.6:535-542.
12. Harrison,S. K. 1990. Interferenceand seed productionby common lamb-
squarters(Chenopodiumalbum) in soybeans (Glycine max). Weed Sci.
38:113-118.
13. Harvey,S. J. and F. Forcella. 1993. Vernalseedling emergence model for
common lambsquarters(Chenopodiumalbum).Weed Sci. 41:309-316.
14. Hayhoe, H. N., L. M. Dwyer, D. Balchin, and J.L.B. Culley. 1993. Tillage
effects on com emergencerates. Soil Till. Res. 26:45-53.
15. Forcella,F. and M. J. Lindstrom.1988. Weed seed populationsin ridge and
conventionaltillage. Weed Sci. 36:500-503.
16. Frick,B. andA. G. Thomas. 1992. Weedsurveysin differenttillage systems
in southeasternOntariofield crops. Can. J. Plant Sci. 72:1337-1347.
17. Froud-Williams,R. J. 1988. Changesin weed florawith differenttillage and
agronomic managementsystems. Pages 213-236 in M. A. Elder and M.
Liebman, eds. Weed Management in Agroecosystems: Ecological Ap-
proaches.CRC Press, Boca Raton,Florida.
18. Gebhardt,M. R., T. C. Daniel, E. E. Schweizer,and R. R. Allmaras. 1985.
Conservationtillage. Science 230:625-630.
19. Jan, P and R. Faivre-Dupaigre.1977. Incidencedes facon culturalessur la
flore aventice. Proc. EWRS Symposium on Different Methods of Weed
Controland TheirIntegrations,Uppsala 1:57-64.
20. Holm, L. G., D. L. Plucknett,J. V. Pancho, and J. P. Herberger.1977. The
world's worst weeds. East-West Center Book, Univ. Press of Hawaii,
Honolulu,Hawaii. 609 p.
21. Kapusta,G. and R. F. Krausz. 1993. Weed control and yield are equal in
conventional,reduced-,andno-tillagesoybean(Glycinemax)after 11years.
WeedTechnol.7:443-451.
22. Lewis, J. 1973. Longevity of crop and weed seeds: survivalafter 20 years
in soil. Weed Res. 13:179-191.
23. Mohler,C. L. 1991. Effects of tillage andmulchon weed biomassand sweet
corn yields. Weed Technol.5:545-552.
24. Mohler,C. L. 1993. A model of the effects of tillage on emergenceof weed
seedlings. Ecol. Appl. 3:53-73.
25. Mohler,C. L. andM. B. Callaway.1992. Effects of tillage andmulchon the
emergenceand survivalof weeds in sweet corn. J. Appl. Ecol. 29:21-34.
26. Norris,R. F. 1992. Case history for weed competition/populationecology:
Barnyardgrass(Echinochloacrus-galli) in sugarbeets(Beta vulgaris).Weed
Technol.6:220-227.
27. Roberts,H. A. and F. G. Stokes. 1965. Studies on the weeds of vegetable
crops. V. Final observationson an experimentwith differentprimaryculti-
vations.J. Appl. Ecol. 2:307-315.
28. Swanton,C. J., D. R. Clements,andD. A. Derksen. 1993. Weed succession
underconservationtillage: a hierarchicalframeworkfor researchand man-
agement.Weed Technol.7:286-297.
29. Teasdale,J. R., C. E. Beste, and W. E. Potts. 1991. Response of weeds to
tillage and cover crop residue.Weed Sci. 39:195-199.
30. Thomas,A. G. and B. L. Frick. 1993. Influenceof tillage systems on weed
abundancein SouthwesternOntario.WeedTechnol.7:699-705.
31. Toole, E. H. and E. Brown. 1946. Final results of the Durvel buried seed
experiment.J. Agric. Res. 72:201-210.
32. Weaver,S. E. 1991. Size-dependenteconomic thresholdsfor threebroadleaf
weed species in soybeans. 5:674-679.
321
 CLEMENTSET AL.: TILLAGEEFFECTSON WEED SEED RETURNAND SEEDBANKCOMPOSITION
33. Weaver,S. E., C. S. Tan,and P. Brain. 1988. Effect of temperatureand soil 35. Wilson, R. G. 1988. Biology of weed seeds in the soil. Pages 28-39 in M.
moistureon time of emergenceof tomatoesand four weed species. Can. J. A. Altieri, and M. Liebman eds. Weed Managementin Agroecosystems:
Plant Sci. 68:877-886. Ecological Approaches.CRC Press, Boca Raton,Florida.
34. Wiese, A. M. andL. K. Binning. 1987.Calculatingthe thresholdtemperature 36. Yenish,J. P.,J. P. Doll, andD. D. Buhler.1992. Effects of tillage on vertical
of developmentfor weeds. Weed Sci. 35:177-179. distributionand viability of weed seed in soil. Weed Sci. 40:429-433.

322 Volume44, Issue 2 (April-June) 1996

This content downloaded on Fri, 21 Dec 2012 14:28:08 PM

All use subject to JSTOR Terms and Conditions
View publication stats