Effects of reduced and conventional tillage on weed communities: Results of a long-term experiment in Southwestern Spain [Efeito do cultivo mínimo e plantio convencional em comunidades de plantas daninhas: Resultados de um experimento de longo prazo no sudoeste da espanha]

Plan a Daninha 2019; 37:e019201336 PARDO, G. e al. E ec s o educed and con en ional illage on weed communi ies: esul s o a long- e m expe imen in ... 1 151103-PD-2016 (9 páginas) PROVA GRÁFICA PARDO, G.1* CIRUJEDA, A.1 PEREA, F.2 VERDÚ, A.M.C.3 MAS, M.T.3 URBANO, J.M.2 A icle PLANTA DANINHA SOCIEDADE BRASILEIRA DA CIÊNCIA DAS PLANTAS DANINHAS 1Cen o de In es igación y Tecnología Ag oalimen a ia de A agón-Ins i u o Ag oalimen a io de A agón-IA2 (CITA-Uni e sidad de Za agoza), Za agoza, Spain; 2 Uni e sidad de Se illa, Se ille, Spain; 3 Escola Supe io d’Ag icul u a de Ba celona, Uni e si a Poli écnica de Ca alunya, Ba celona, Spain. Doi: 10.1590/S0100-83582019370100152 <h p://www.sbcpd.o g> ISSN 0100-8358 (p in ) 1806-9681 (online) * Co esponding au ho : <[email p o ec ed]> Recei ed: May 29, 2018 App o ed: Sep embe 25, 2019 Plan a Daninha 2019; 37:e019201336 Copy igh : This is an open-access a icle dis ibu ed unde he e ms o he C ea i e Commons A ibu ion License, which pe mi s un es ic ed use, dis ibu ion, and ep oduc ion in any medium, p o ided ha he o iginal au ho and sou ce a e c edi ed. EFFECTS OF REDUCED AND CONVENTIONAL TILLAGE ON WEED COMMUNITIES: RESULTS OF A LONG-TERM EXPERIMENT IN SOUTHWESTERN SPAIN E ei o do Cul i o Mínimo e Plan io Con encional em Comunidades de Plan as Daninhas: Resul ados de um Expe imen o de Longo P azo no Sudoes e da Espanha ABSTRACT - An impo an d awback in adop ing minimum illage (MT) and no- illage (NT) echniques is he equen ly obse ed weed shi p omo ing adap ed species and achie ing poo e weed con ol. These changes can be de ec ed bes wi h long- e m expe imen s, and esul s migh di e depending on soil cha ac e is ics and he local lo a. The objec i es o his wo k we e o e alua e he e ec o educed illage on weed seed dis ibu ion in he soil p o ile and o iden i y possible consequences on weed di e si y on a long- e m expe imen main ained du ing 24 yea s in Se ille (Spain) wi h h ee illage sys ems: NT, MT and con en ional illage (CT) including moldboa d plow on a e isol. Fo his pu pose, soil seedbanks a 0-8 cm and 8-16 cm dep hs we e enume a ed in au umn 2005 and in- ield eme ged plan s in au umn 2005 and win e 2006. Shannon di e si y index (H) and e enness (J’) we e calcula ed o seedbank and abo eg ound weed communi ies. To al weed seed densi y was highes o NT and lowes o CT. Some big-seeded species, such as Ch ozopho a inc o ea L., showed highes seed densi y in CT. NT inc eased he ela i e densi y o Ama an hus bli oides S. Wa son seeds in he seedbank and he abundance o eme ged plan s o Mal a pa i lo a L., Anagallis a ensis L. and Pic is echioides L. O e all, MT led o a less di e se seedbank in he 0-8 cm dep h o soil han CT. The equen d ough -induced deep ac u es in he expandable clay soil caused na u al illage, which p obably esul ed in ewe di e ences in weed seed and seedling densi ies among illage ea men s compa ed o wha migh be expec ed in o he soil ypes. Keywo ds: no- illage, minimum illage, moldboa d plow, mechanical weed con ol, seedbank, di e si y indexes. RESUMO - Uma des an agem impo an e na adoção de écnicas de cul i o mínimo (MT) e plan io di e o (NT) é o deslocamen o de plan as daninhas equen emen e obse ado, p omo endo espécies adap adas e, com isso, um con ole de plan as daninhas mais p ecá io. Essas mudanças só podem se de ec adas com expe imen os de longo p azo, e os esul ados podem di e i , dependendo das ca ac e ís icas do solo e da lo a local. Os obje i os des e abalho o am a alia o e ei o da la ou a eduzida sob e a dis ibuição de semen es de plan as daninhas no pe il do solo e iden i ica possí eis consequências na di e sidade de plan as daninhas em um expe imen o de longa du ação man ido du an e 24 anos em Se ilha (Espanha) com ês sis emas de plan io: NT, MT e p epa o con encional (CT), incluindo a ado de ai eca em um Ve issolo. Com esse p opósi o, amos as de solo com 0-8 e 8-16 cm de p o undidade o am cole adas no in e no de 2005 e na p ima e a de ID ID ID ID ID ID Plan a Daninha 2019; 37:e019201336 PARDO, G. e al. E ec s o educed and con en ional illage on weed communi ies: esul s o a long- e m expe imen in ... 2 2006, e as plân ulas de plan as daninhas eme gen es o am egis adas. O índice de di e sidade de Shannon (H) e a uni o midade (J’) o am calculados pa a comunidades de banco de semen es e plan as daninhas acima do solo. A densidade o al de semen es de plan as daninhas oi maio pa a NT e meno pa a CT. Algumas espécies de semen es g andes, como Ch ozopho a inc o ea L., ap esen a am maio densidade de semen es no CT. NT aumen ou a densidade ela i a de semen es de Ama an hus bli oides S. Wa son no banco de semen es e a p esença de plan as eme gidas de Mal a pa i lo a L., Anagallis a ensis L. e Pic is echioides L. Com elação aos esul ados ge ais, a TM le ou a um banco de semen es menos di e si icado na p o undidade de 0 a 8 cm do solo do que a CT. As equen es a u as p o undas na a gila expansí el do solo p o ocam um p epa o na u al, o iginando p o a elmen e menos di e enças dos pa âme os analisados do que em ou os ipos de solo. Pala as-cha e: semeadu a di e a, cul i o mínimo, a ado de ai eca, con ole mecânico de plan as daninhas, banco de semen es, índices de di e sidade. INTRODUCTION P oduc ion echniques like minimum illage (MT) and non- illage (NT) sys ems a e al e na i es o educe soil losses (B ady and Weil, 2002), o main ain o ganic ma e and mois u e, and also o p o ide s abiliza ion o agg ega es a o ing wa e in il a ion (Liebig e al., 2004; Thie elde e al., 2015) e sus con en ional illage (CT). Mo eo e , conse a ion ag icul u e has o he bene i s such as s imula ing bene icial mic o auna ac i i y (Zhang e al., 2015), achie ing mo e ca bon seques a ion han con en ional sys ems (Buchi e al., 2017) and, addi ionally, may sa e money and ime by educing illage ope a ions p io o sowing c ops (Chauhan e al., 2012). Ni ogen e ilize and uel we e ound o be he inpu s wi h he highes educ ion le els in he USA (Ande son, 2017). Howe e , educed illage also in ol es some disad an ages, such as highe su ace compac ion ha can dec ease wa e in il a ion, new weed con ol p oblems, and majo ini ial pu chases o equipmen o di ec sowing, sp aying weeds, and ha es ing c ops. Conce ning weed con ol, ewe managemen op ions exis unde MT and NT and, he e o e, weed p oblems inc ease compa ed o CT o a leas wo easons. Fi s , he absence o soil in e sion causes ecen ly shed seeds o emain close o he soil su ace, in a be e posi ion o ha e a highe p obabili y o es ablishmen (Bà be i and Lo Cascio, 2001), despi e being mo e exposed o p eda o s like ha es e an s (Ba aiba e al., 2009). Second, MT and NT allow a highe su i al a e o he p opagules o pe ennial species and inc ease hei possibili y o sp ou ing, enabling u he de elopmen . Thus, MT and NT migh p omo e in es a ions o species ha eme ge easily a e pa ial bu ial o being le on he soil su ace. Some examples o hese species a e g asses like B omus diand us Ro h o Alopecu us myosu oides L. and he pe ennial species Ci sium a ense L., Con ol ulus a ensis L., Ca da ia d aba (L.) Des ., Sonchus ene imus L. and Mal a pa i lo a L. (Mas and Ve dú, 2003; Sans e al., 2011; Soane e al., 2012). On he o he hand, some weed species may show smalle popula ions in MT and NT sys ems, ei he because hei seeds need o be bu ied a a ce ain dep h o ge mina e, o because hei ge mina ion mechanism is ac i a ed only i he soil is illed. Following Na a e e and Fe nández- Quin anilla (2003), his is he case o Ve onica hede i olia L., Descu ainia sophia (L.) Webb ex P an l and Lamium amplexicaule L. The species Polygonum spp. (F oud-Williams e al., 1983), Fuma ia o icinalis L., Raphanus aphanis um L. and Chenopodium album L. a e also conside ed o be well adap ed o con en ional illage (CT) and hei densi y is expec ed o decline wi h educed illage. Howe e , some species as Papa e hoeas L. seem o be adap ed o bo h sys ems (Ci ujeda e al., 2003; Na a e e and Fe nández-Quin anilla, 2003). Especially he bicide-suscep ible species a e expec ed o dec ease hei densi y o e ime in educed illage because he bicides a e used mo e in ensi ely, unless hey de elop esis ance. Changes in weed communi ies a e expec ed o a ise wi hin a ew yea s o implemen ing MT o NT sys ems, since weed species be e adap ed o no- illage and he bicides will g adually eplace o he species (To esen e al., 2003; P imo e al., 2006; Chauhan e al., 2012) and weed p essu e can be a se ious impedimen o es ablishing he minimum illage sys ems (Lee and Thie elde , 2017). Ne e heless, some o he s udies indica e ha illage educ ion Plan a Daninha 2019; 37:e019201336 PARDO, G. e al. E ec s o educed and con en ional illage on weed communi ies: esul s o a long- e m expe imen in ... 3 does no always lead o changes in weed ichness, bu a he o changes in he ela i e abundance o weed species (De ksen e al., 1993; McCloskey e al., 1996). In any case, e en i a weed shi occu s, i may no be a ibu able o conse a ion ag icul u e alone, because o he ac o s such as c op o a ion, clima e and long-dis ance weed p opagule mo emen s also may impac . Addi ionally, soil ype p obably in luences he e ec s o educing illage ope a ions. Ve isols occupy an a ea o 672,000 ha in Andalusia, Spain (Fe nández, 2011), and 260 million ha wo ldwide (Dudal, 1963), and hey a e e y e ile in yea s wi h enough ain all, al hough wa e a ailabili y in d y yea s is a limi ing ac o . These soils can p oduce high yields in conse a ion illage (Blaise e al., 2015) p o ided weeds a e e ec i ely con olled (Giambal o e al., 2012). Howe e , ew da a a e a ailable o hese ypes o soils ega ding MT. Biodi e si y in ag icul u al sys ems gene ally is accep ed as necessa y o achie e a high s abili y and p o ec ion agains en i onmen al s esses (Al ie i, 1999). In his espec , he impo ance o weed communi ies, as an in eg al pa o ag oecosys ems, is inc easingly claimed o p o ide a wide ange o ecological unc ions, such as he abili y o espond o some dis u bances (Cos anza e al., 1997) and pes con ol (Al ie i, 1999). Whe e MT and NT ha e been adop ed ex ensi ely, a ew p oblema ic weed species, well adap ed o educed illage and o he bicides, apidly inc ease hei popula ion densi ies and he eby educe biodi e si y. Da is e al. (2005), To esen e al. (2003) and Ga cía e al. (2014) desc ibe some in e es ing examples o plan species ha ha e played his ole in di e en a eas o he wo ld. Long- e m expe imen s can p o ide essen ial in o ma ion abou he illage e ec on weed in es a ions. In pa icula , seedbank s udies allow a e ospec i e iew o he ea men e ec s and con ibu e in achie ing a p edic i e insigh in o possible u u e weed p oblems (Fo cella, 1992; Hossain and Begum, 2015). Howe e , long- e m ials ocusing on weeds a e complex and labo ious, as e idenced by he ew exis ing examples in he li e a u e showing da a o mo e han 10 yea s o minimum illage (Table 1). Un o una ely, he newes esea ch on minimum illage ocuses mo e on opics o he han weeds, e en hough weed con ol ha e been shown o be c ucial o adop minimum illage (Lee and Thie elde , 2017). The esul s o illage e ec s on weed communi ies canno be ex apola ed en i ely o o he a eas wi h di e en clima es, soils, and lo as. The e o e, in o de o assess such e ec s in a pa icula zone, local expe imen s a e impo an . Un o una ely, no ele an published da a a e a ailable o sou he n Spain. Howe e , such da a we e collec ed in he long- e m c op and soil managemen ial desc ibed by O dóñez Fe nández e al. (2007), and he esul s ele an o weed seed banks will be summa ized he e. The wo objec i es o his wo k we e: (1) E alua e he e ec o MT, NT and CT on weed seed dis ibu ion in he soil p o ile and on he eme ged weeds analyzing each species indi idually in he longes c op- ee pe iod o a whea -sun lowe -legume o a ion. The longes c op- ee pe iod was be ween whea ha es and sun lowe sowing. (2) Iden i y long- e m consequences o he illage sys ems on weed di e si y a e 24 yea s (1982-2005). The expe imen was pe o med in Table 1 - Summa y o he main s udies in long- e m expe imen s ocusing on weeds conduc ed in ain ed condi ions Au ho Coun y C op Pe iod (yea s) Soil ex u e Sosnoskie e al. (2006) USA Con inuous co n (Zea mays L.), co n-soybean (Glycine max L.), co n-oa (A ena sa i a L.) 35 Sil loam He nández-Plaza e al. (2015) Spain Whea (T i icum aes i um L. and T. du um Des .)- e ch (Vicia sa i a L.) 24 Loam Giambal o e al. (2012) I aly Con inuous whea (T. aes i um), whea - aba bean (Vicia aba), whea -be seem clo e (T i olium alexand ium) 18 Clay ( e isol) Voll e al. (2001) B azil Soybean-whea 16 70% clay Do ado and López-Fando (2006) Spain Ba ley (Ho deum ulga e L.) - e ch, ba ley- sun lowe (Helian hus annuus L.) 16 Sandy loam San ín-Mon anyá e al. (2017) Spain Whea 16 Sandy loam Ca e and I any (2006) Canada Soybean - ba ley 14 Sandy loam Bà be i and Lo Cascio (2001) I aly Whea (T. aes i um) - pigeon bean (Vicia aba L. a . mino ) 12 Sandy loam Da is e al. (2005) USA Co n-soybean-whea (T. aes i um) 12 Sil loam Plan a Daninha 2019; 37:e019201336 PARDO, G. e al. E ec s o educed and con en ional illage on weed communi ies: esul s o a long- e m expe imen in ... 4 he Se ille egion o sou he n Spain. This si e and s udy ep esen ed he ain ed a gillic soils in he a ea, and i was he only second such long- e m expe imen on hea y clay soils in Eu ope (Giambal o e al., 2012). Fou hypo heses we e es ed. (1) Tillage educ ion dec eases weed biodi e si y bo h in he soil seedbank and in he eme ged weed communi y. (2) Di e ences in weed seedbank and eme gence a e expec ed o occu o he di e en illage sys ems. (3) Changes in he weed lo a could happen when educing illage in ensi y, he eby p omo ing hose species adap ed o less illage and o he applied he bicides and dec easing o he species adap ed o illage. (4) These changes may cause a o al highe seedbank and eme ged weed densi y in MT and NT sys ems compa ed o CT. MATERIAL AND METHODS S udy a ea A pe manen long- e m expe imen has been conduc ed since 1982-1983 a Tomejil, nea Se ille (Spain), 37o24’07"N and 05o35’10"W, a 97 m abo e sea le el. The si e is loca ed in he Andalusian e ile plain called “Campiña,” which has a Medi e anean clima e. The soil is a ine mon mo illoni ic Ch omic Haploxe e (Soil Su ey S a , 1999) o med on a Miocene ma l wi h mo e han 700 g kg-1 clay in he 0-0.15 m laye and only 1 g kg-1 sand (O dóñez Fe nández e al., 2007). The domina e expandable clay o his soil is mon mo illoni e, which cha ac e is ically o ms deep issu es du ing d y pe iods ha acili a e desicca ion and p obably gi es weed seeds he oppo uni y o en e deepe soil laye s na u ally, wi hou plowing. The clima e o he expe imen al si e is sub- humid Medi e anean wi h a mean ain all o 575 mm (Figu e 1, da a om he s udy a ea a Tomejil, Se ille). The d y pe iod is om May o Sep embe ; low ain all in May and Sep embe is no e ec i e due o high empe a u es. The mean ai empe a u e is 15.0 oC in au umn, 11.9 oC in win e , 20.8 oC in sp ing and 26.3 oC in summe . The a e age minimum and maximum empe a u es a e 11.5 oC and 25.5 oC, espec i ely. Weed eme gence begins a e he au umn ain all and gene ally s ops in Ap il. Figu e 1 - Omb o he mic diag am o he Tomejil expe imen al ield (Se ille, Spain). A e age da a yea s 2002 o 2017. Expe imen al design Th ee illage sys ems we e compa ed: NT, MT and CT. The elemen a y plo s measu ed 180 m x 15 m (2,700 m2) and we e dis ibu ed in a andomized comple e block design wi h h ee eplica es o each ea men . In all plo s a whea (T i icum aes i um L. and T. du um Des .) – sun lowe (Helian hus annuus L.) - legume ( e ch, Vicia aba L. o peas, Pisum sa i um L.) o a ion was pe o med. Whea , e ch and peas we e sown in Decembe and ha es ed in May; sun lowe was sown in Ma ch and ha es ed in Augus . The o a ion las s h ee yea s, conside ing one yea o each c op. Weed con ol du ing he c opping pe iods was conduc ed when necessa y wi h au ho ized he bicides o he espec i e c op ob aining su icien con ol in all 24 yea s; in many yea s he bicide use was no necessa y in sun lowe , a compe i i e and as -g owing c op. He bicides we e applied wi h a ac o -moun ed ield boom sp aye wi h he ollowing cha ac e is ics: speed 6.5 km h-1, p essu e 2.5 ba , s anda d API, Albuz nozzles, sp ay olume o 200 L ha-1 o glyphosa e and 300 L ha-1 o he o he he bicides. Du ing he 24 yea s, in he c op- ee pe iods, weed con ol in CT was conduc ed mechanically, while he bicides we e used in MT and NT (see Table 2). In au umn, be o e sowing whea , e ch o peas, he no mal ea men s we e glyphosa e + MCPA a 0.6 and 0.5 L ha-1, espec i ely, in p e-sowing o NT and MT wi h an addi ional ea men a highe dosage i needed in we au umns. Plan a Daninha 2019; 37:e019201336 PARDO, G. e al. E ec s o educed and con en ional illage on weed communi ies: esul s o a long- e m expe imen in ... 5 Du ing he longe c op- ee pe iod p e ious o sun lowe sowing, mechanical ea men s we e used in MT and mo e he bicide ea men s we e needed in NT (Table 2). Thus, he pe iod o ime in he ial in which mo e di e ences in weed managemen a e expec ed is he ime be ween he win e c op ha es in May and he sun lowe sowing in Ma ch, which is exac ly he ime pe iod conside ed in his s udy. As empe a u es in summe a e whea ha es a e e y high and ain all is e y sca ce, i ually no weeds eme ge un il he au umn ain all. In he 2005-2006 sampling season he sown c op was sun lowe and he o dina y p e-plan ing ope a ions we e conduc ed o each ea men (Table 2). Soil, seed and seedling sampling and iden i ica ion Soil samples we e aken in he cen al 600 m2 o each plo . A 5 cm diame e cylind ical soil sample o 8 cm dep h was used (Eijkelkamp, sample ing ki wi h open ing holde , model 07.53.SA, Giesbeek Ne he lands) main aining each soil sample in a me al cylinde . Soil cylinde s we e d ied a 50 oC du ing 2 days, la e ex ac ed om he me al con aine , soil placed in plas ic bags and kep a -25 oC un il p ocessing. The pa e n ollowed o sampling was a ixed g id (Amb osio e al., 2004) consis ing o 7 ows 10 m apa and 3 columns 5 m apa . Two sub-samples we e ex ac ed a each o he 21 c ossing poin s o his g id: one sub-sample was o he op 8 cm o soil, and he o he o dep h 8-16 cm. Thus a o al o 378 soil co es we e aken. Sampling was conduc ed in No embe 2005 when soil p epa a ion o sun lowe seeding s a ed. Weed seeds we e sepa a ed om he soil ollowing he p o ocol o Ca e e o (1977) mixing 200 mL o wa e wi h each 100 mL o soil sample, adding 10 g o dispe san s (i.e. sodium hexame aphospha e and 5 g o sodium bica bona e). The liquid sample was shaken du ing 90 minu es a 11 .p.m in a shake (Heidolph model Reax 20, Schwabach, Ge many). The solu ions including he dispe sed seeds we e pou ed h ough a owe o h ee sie es o 1, 0.5 and 0.25 mm mesh. The solid ac ions we e d ied a 35-40 oC du ing 1 hou and la e seeds iden i ied by consul ing images a ailable on he in e ne (HYPPA, 2006). Fo seedling iden i ica ion o plan s eme ging in he ield, Spanish weed lo a by Villa ías (2000) was used. Uniden i ied plan s a he seedling s age we e ansplan ed o po s and g own in he g eenhouses o he Facul y o Ag icul u al Enginee s (EUITA), Se ille, un il he lowe ing s age and hen iden i ied using Ca e e o (2004). Unde he local condi ions, no mally no seedling eme gence occu s du ing he d y summe pe iod; i s a s only a e au umn ains begin. The e o e eme ged weeds we e coun ed be o e soil sample ex ac ion in No embe 2005 (immedia ely be o e seedbed p epa a ion) and a e wa ds in Feb ua y 2006, sho ly be o e sowing he sun lowe c op, in o de o compa e he numbe o seeds in he uppe soil p o ile o dep h 8 cm wi h he eme ged weed lo a. Flo a sampling in No embe documen ed he eme ged weeds du ing he whole pe iod a e ha es ing he p e ious whea c op un il au umn. Sampling in Feb ua y documen ed he su i ing plan s a e illage ope a ions and he bicide applica ions in No embe and desc ibed newly eme ged weeds, hus o cha ac e ize he weed in es a ion sho be o e sowing he sun lowe . No coun s we e conduc ed du ing he c op because weed managemen in ha pe iod was he same o all h ee illage sys ems. Mo eo e , no he bicide was used du ing he sun lowe in he sampling yea because low weed in es a ion was ound, which demons a ed he g ea compe i i e capaci y o his c op. Table 2 - Ope a ions conduc ed o each illage sys em in season 2005-06 a he expe imen al si e Applica ion da e Tillage sys em No- illage Minimum illage Con en ional illage No embe 0.7 L ha -1 glyphosa e 36% + 0.5 L ha -1 MCPA 60% Cul i a o (1) Moldboa d plow (2) Janua y 1 L ha -1 glyphosa e 36% + 0.7 L ha -1 MCPA 60% Vib o-cul i a o (3) Cul i a o P e-sowing (Ma ch 30 h ) 2 L ha -1 glyphosa e 36% + 0.150 L ha -1 oxy luo en 48% 1.5 L ha -1 glyphosa e 36% + 1 L ha -1 MCPA 60% Vib o-cul i a o (1) Cul i a o : e ical illage a dep h 15-20 cm. (2) Moldboa d plow: illage wi h soil in e sion, a dep h 40-45 cm. (3) Vib o-cul i a o : e ical illage a dep h less han 15 cm. Plan a Daninha 2019; 37:e019201336 PARDO, G. e al. E ec s o educed and con en ional illage on weed communi ies: esul s o a long- e m expe imen in ... 6 Thus, coun s in Feb ua y e lec he main plan communi y p io o sun lowe sowing ha would con inue g owing in he c op in NT i he bicides we e no e ec i e and in MT and CT i mechanical ea men s we e ine ec i e. The basic sampling a ea o he ege al communi y was a ci cula quad a wi h a 25 cm diame e , and 21 quad a s we e sampled pe plo . The cen e o each quad a coincided wi h one o he 21 soil sampling poin s in each plo . In he Feb ua y coun s, ecen ly eme ged seedlings and olde plan s wi h mo e han 4 ue lea es we e eco ded sepa a ely. Da a analysis and di e si y measu emen The means o weed and seed densi ies o he 21 samples pe plo we e calcula ed and es ed o no mali y and a iance homogenei y. Seedbank da a o bo h dep hs and di e si y indexes in Feb ua y equi ed ln (x+1) ans o ma ion o ul ill hese equi emen s. ANOVAs we e conduc ed o each species sepa a ely and o he o al weed and seed densi ies conside ing he illage ac o in a comple ely andomized design. Fo he seedbank samples each dep h was analyzed sepa a ely in all cases. A e wa ds Tukey mean sepa a ion es s we e pe o med. Boxplo s we e cons uc ed o weed species ha did no show signi ican di e ences despi e ha ing e y di e en means among ea men s due o da a dispe sal. Means o he aw seed and seedling numbe s pe plo we e used o calcula e he Shannon– Wea e di e si y indices (H) o each epe i ion o he h ee illage sys ems acco ding o he exp ession (Magu an, 2004): whe e S is he numbe o species o species ichness, pi = Ni/N, wi h Ni being he numbe o indi iduals o he species i and N he o al numbe o indi iduals. Also, he Shannon e enness index (J’) was calcula ed as: ‘J’ = H/Hmax whe e Hmax = ln S. High J’ implies g ea e uni o mi y o species’ abundance (Magu an, 2004). Da a o 21 co es and da a o 21 quad a s om each elemen a y plo we e combined in o de o compu e wo di e si y indices. O he calcula ed pa ame e s we e plan densi y (plan s m-2), numbe o weed species pe plo ( ichness), and he numbe o samples wi h weeds pe plo (being adul plan s o seedlings). The calcula ed da a we e analyzed s a is ically acco ding o he expe imen al design using he p og am SPSS 15.0.1 (SPSS, 2006). RESULTS AND DISCUSSION Tillage e ec on weed seed densi y Soil dep h o 0-8 cm The o e all mean weed seed densi y o all species combined in he i s 8 cm was signi ican ly highe unde NT compa ed o CT being app oxima ely wice as high, whe eas ha o MT was in e media e (Table 3), 24 yea s a e applying he di e en illage sys ems, con i ming one o he hypo heses o his wo k. These esul s a i med he obse a ions o To esen e al. (2003) and Conn (2006). When s udying he species sepa a ely, Ch ozopho a inc o ia L. was he only species wi h signi ican ly highe mean seed densi y in CT han in he o he wo illage sys ems, and Polygonum a icula e L. was he only species ha had i s highes seed densi y in MT. No single species had signi ican ly highe seed densi ies in NT han in MT o CT. The esul s o seed banks a e a consequence o he combina ion o he illage sys em and associa ed weed con ol me hods, and e ec s o hese wo managemen a iables canno be Plan a Daninha 2019; 37:e019201336 PARDO, G. e al. E ec s o educed and con en ional illage on weed communi ies: esul s o a long- e m expe imen in ... 7 Table 3 - Mean seed densi y (seeds m-3) ound in he 0-8 cm dep h o soil sampled in No embe 2005 CT MT NT Seed size (mm) Ama an hus bli oides S. Wa son 506 a 202 a 3,596 a 1.2-1.7 (2) Anagallis a ensis L. 707 a 1,010 a 2,156 a 0.9-1.4 x 0.6-1.0 (2) Chamaesyce canescens L. 505 a 202 a 616 a 1.0-1.5 x 0.6-0.9 (2) Chenopodium ul a ia L. 0 a 0 a 103 a 1.0-1.5 (2) Ch ozopho a inc o ia L. 505 a 101 b 103 b 4.0-5.0 x 3.5-4.0 (2) Con ol ulus icolo L. 202 a 0 a 0 a 4.0-5.0 x 3.6-4.3 (2) B assicaceae 0 a 0 a 103 a Poaceae o he han P. pa adoxa 0 a 101 a 103 a Kickxia spu ia L. 0 a 910 a 822 a 0.7-1.0 x 0.5-0.8 (2) Lina ia la i olia Mill. 101 a 101 a 0 a 1.6-2.4 x 1.5-2.0 (2) Papa e hoeas L. 0 a 101 a 205 a 0.7-0.9 (1) Phala is pa adoxa L. 0 a 0 a 103 a 1.7-1.8 (1) Pic is echioides L. 1,112 a 505 a 719 a 0.9-1.0 x 2.5-3.0 (1) Polygonum a icula e L. 707 b 2,122 a 719b 2.0-3.0 (2) To al 4,345 b 5,355 ab 9,348 a CT: Con en ional illage, MT: Minimum illage, NT: No- illage. Mean seed numbe in ows ollowed by he same le e s a e no signi ican ly di e en acco ding o he Tukey es (P<0.05). (1) HYPPA (2006), (2) Cas o iejo (1986-2012). sepa a ed. Thus, when mo e seeds o a ce ain species a e ound in CT his is due o he combina ion o moldboa d ploughing and no he bicide use. In con as , when mo e seeds a e ound in NT, he eason is lack o illage and adap a ion o he plan s o he he bicides used. Mos o he esul s ob ained in o he long- e m s udies compa ing illage e ec s on weed seed dis ibu ion ound mo e weed seeds in he uppe soil laye in NT compa ed o CT (Table 3), con i ming he esul s o he p esen wo k (Do ado and López-Fando, 2006; To esen e al., 2003; Ca e and I any, 2006; Sosnoskie e al., 2006). Howe e , Bà be i and Lo Cascio (2001) did no ind any di e ences in weed seed con en in he uppe soil pa due o illage sys ems and, su p isingly Swan on e al. (2000) and Ruisi e al. (2015) ound he opposi e esul s. This he e ogenei y is p obably caused by he di e en soil ypes as well as by he domina ing weed species beha ing di e en ly. In ac , weed seed size is one o he ele an a iables explaining why some species dec ease and o he inc ease wi h NT. La ge seeds p obably su i e a sho e ime in MT and NT as he e is less con ac be ween seeds and solid soil pa icles p o ec ing hem (Te ps a, 1990; Reuss e al., 2001) and also due o seed p eda ion. This would explain signi ican ly highe mean densi y in CT o he big seeds o C. inc o ia and he same endency o Con ol ulus icolo L. and P. echioides in he p esen ial compa ed o he o he illage sys ems (Table 3). Small-seeded and pe ennial weeds a e gene ally mo e abundan in conse a ion ag icul u e (Bajwa, 2014) al hough He nández-Plaza e al. (2015) showed in hei long- e m expe imen ha CT did no s ic ly allow only big seeds o su i e bu o e ed a g ea e a iabili y in seed size han MT and NT. Howe e , species shi p ocesses a e complex and canno be a ibu ed o seed size, only. Seeds o some species, i.e. Ama an hus bli oides S. Wa son., Anagallis a ensis and P. hoeas, ended o be mo e abundan unde NT al hough a iabili y o da a did no allow s a is ical di e ences (Figu e 2). Howe e , mean densi y o A. bli oides seeds unde NT was 7 imes g ea e han unde CT and almos 18 imes g ea e han unde MT (Figu e 2); mo eo e , he pa ame e s ela i e densi y, ela i e equency and ela i e abundance o seeds di e ed signi ican ly being again highe o NT compa ed o he o he wo sys ems (da a no shown, p<0.05). A highe mean densi y o A. bli oides unde NT has been obse ed by o he au ho s, oo (Zawieja and Ko das, 2003), e en hough in o he cases his species is supposed o be a o ed by illage (Masiunas e al., 1997; De ksen e al., 1993). A possible explana ion o such con adic o y esul s is ha small seeds may ha e a s onge in e ac ion wi h soil pa icles which may in luence seed ge mina ion (Reuss e al., 2001) so ha species like A. bli oides may inc ease o dec ease in impo ance depending on he soil p ope ies. Addi ionally, i egula seed dis ibu ion in he soil is e y common so ha expe imen al esul s may be also condi ioned by his ac and di e ences may be smalle han expec ed in some cases (see in his wo k Figu es 2, Table 3). Plan a Daninha 2019; 37:e019201336 PARDO, G. e al. E ec s o educed and con en ional illage on weed communi ies: esul s o a long- e m expe imen in ... 8 The ob ained esul s (Table 3) ne e heless suppo he hypo hesis ha he e is a species-speci ic esponse o educed illage in he seedbank. Polygonum a icula e L. mean seed densi ies we e signi ican ly highe unde MT han unde he o he wo sys ems, and some o he species including Chenopodium ul a ia L., Kickxia spu ia L. Dumo . and Phala is pa adoxa L. ended o ha e highe mean seed densi y unde NT (Table 3) wi h he comple e absence o seeds o hese species in some illage ea men s; his shows a clea end ei he o emain s eady o o g ow in impo ance in some illage sys ems and o educe o o disappea in o he s. Highe P. a icula e mean seed densi ies we e ound o MT compa ed o he o he wo sys ems (Table 3). In cons as , Ruisi e al (2015) ound P. a icula e associa ed o CT and Do ado and López-Fando (2006) also ound highe P. a icula e seed densi y wi h inc easing illage in ensi y in an expe imen pe o med in cen al Spain, al hough no in all cases. Also in se e al s udies on win e ce eal c ops (i.e. F oud-Williams e al., 1983; Do ado e al., 1999; Ve dú and Mas, 2004), P. a icula e has been desc ibed as a cha ac e is ic species o CT, because highe seedling densi ies we e eco ded unde his sys em. Howe e , Bà be i and Lo Cascio (2001) did no ind any di e ences in he seedbank o his species a e 12 yea s o compa a i e illage in 0-45 cm soil dep h. Such di e se and con adic o y esul s jus i ied he need o epea ing hese kind o expe imen s a di e en si es. Acco ding o Ve dú and Mas (2004) he highe densi y eco ded in he MT soil seedbank in he p esen s udy could be explained by di e ences in seedling ec ui men , seedling su i al and adul plan sizes o eme ged P. a icula e plan s unde he h ee illage sys ems (CT, MT and NT). Disco e ing which o hese ac o s is mainly esponsible would equi e e y de ailed sampling. In he p esen ial he hea y clay soil may ha e p e en ed some seeds om ge mina ing, and he small seed size o his species could con ibu e o explaining he highe seed abundance in MT. Soil dep h o 8-16 cm The same end ound in 0-8 cm was obse ed o hese species in he 8-16 cm dep h o soil bu in his case signi ican di e ences we e only ound o P. a icula e (highe in MT han in CT and NT) wi h simila amoun o seeds compa ed o 0-8 cm dep h. The es o species main ained he same pa e n as in he 0-8 cm soil dep h, bu wi h non-signi ican di e ences among ea men s (da a no shown). In he 8-16 cm dep h NT ended o ha e he highes quan i y o seeds (7,725 seeds m-3) compa ed o CT (4,635 seeds m-3) and MT (5,159 seeds m-3) so ha di e ences seemed o be so ened in dep h. These esul s a e simila o hose ound by Auskalnienë and Auskalnis (2009) who also desc ibed di e ences in he uppe laye (0-10 cm CT: Con en ional illage, MT: Minimum illage, NT: No- illage. Poin s indica e mean alue; hick ho izon al line shows he median. Figu e 2 - Box plo showing seed densi y (seeds m-3) o Ama an hus bli oides (A) and Anagallis a ensis (B) ound in he 0-8 cm dep h o soil sampled in No embe 2005. (A) (B) Plan a Daninha 2019; 37:e019201336 PARDO, G. e al. E ec s o educed and con en ional illage on weed communi ies: esul s o a long- e m expe imen in ... 9 in ha case) wi h less weed seeds in CT han in NT bu no signi ican di e ences in he deepe soil laye (10-20 cm). Addi ionally, hese au ho s also ound highe species numbe in he uppe soil laye o NT and MT (a 0-5 cm) while no di e ences we e ound a 5-20 cm. Ve isolic soils de elop la ge issu es du ing desicca ion, which con ibu ed o na u al illage in all ea men s in he p esen expe imen . This p obably educed he di e ences among illage ea men s a 8-16 cm soil dep h, simila o wha was ound by Ca dina e al. (2002). In ac , his is p obably he eason why weed seeds we e ound in he 8-16 cm dep h o soil in NT e en 24 yea s a e he las soil illage (da a no shown), bu seeds we e no ound a his dep h in o he s udies conduc ed on o he soil ypes. In ligh e soils subjec ed o conse a ion illage a la ge p opo ion o he weed seedbank gene ally emains on o close o he soil su ace a e c op sowing (Chauhan e al., 2012) e en i his illage e ec is no ound in all s udies (e.g., Ca e and I any. 2006). An e ec o soil ex u e is also ha in ligh soils seed do mancy gene ally is educed as he e a e less soil agg ega es and seeds a e less p o ec ed (Te ps a, 1990) and he long- e m illage e ec may be seen ea lie as in hea y soils. On he o he hand, as discussed by Ca e and I any (2006), 24 yea s can be conside ed a oo sho pe iod o ime o y o unde s and any di e ences when s udying long-las ing species as Chenopodim album L. o o he s ha su i e 30-40 yea s (Holm e al., 1977); C. album showed s ill a su i al a e o 28% a e 17 yea s o bu ial (Bu nside e al., 1996). Seedbank di e si y In he 0-8 cm dep h o soil highes mean seedbank di e si y and e enness we e ound o CT and lowes di e si y o MT. Conce ning e enness, MT p esen ed he lowes species abundance uni o mi y in he 0-8 cm dep h o soil, indica ing he exis ence o mo e concen a ed seedbanks. The same endency was de ec ed in he 8-16 cm dep h o soil o bo h indices (Table 4). Howe e , i is no ewo hy ha many species we e ound a mo e homogenous densi ies unde CT compa ed o he o he illage sys ems and only wo species composed mo e han hal o he o al weed seeds ound unde NT and MT demons a ing low di e si y (Table 3). The indings o S o key and Ne e (2018) demons a e ha weed compe i ion is lowe in a si ua ion wi h mo e di e se weed lo a; hus, he ci cums ance unde CT ound in he p esen ial would be mo e desi able om he p oduc i e poin o iew. Table 4 - Shannon di e si y index (H) and Shannon e enness (J’) o he soil seedbank o he h ee illage sys ems in he wo dep hs o soil p o ile Pa ame e /dep h 0-8 cm 8-16 cm CT MT NT CT MT NT H 1.48 (0.08) a 1.12 (0.42) b 1.31(0.22) ab 1.35(0.23) a 1.07(0.54) a 1.13(0.33) a J’ 0.93 (0.02) a 0.74 (0.17) b 0.81 (0.10) ab 0.77(0.08) a 0.64(0.20) a 0.67(0.13) a CT: Con en ional illage MT: Minimum illage NT: No- illage. The b acke s show s anda d e o s. Means in ows ollowed by he same le e s a e no signi ican ly di e en acco ding o he Tukey es (P<0.1). One possible explana ion o inding highe weed seed di e si y in CT in he p esen s udy is ha cen u ies o widesp ead use o his weed con ol echnique in he a ea may ha e selec ed o a la ge numbe o weed species well-adap ed o deep illage leading o he highes di e si y in CT. The use o he bicides, which is he main weed con ol echnique in NT, is ela i ely ecen (s a ing abou 50 yea s ago), so chemical con ol can be expec ed o educe weed di e si y, as he bicide esis ance is no ye a p oblem in he a ea. In MT chemical and mechanical weed con ol is combined, causing he lowes di e si y, con i ming ha plan s o ew species a e expec ed o su i e and ep oduce unde bo h con ol me hods. The obse ed di e ences o e enness can be explained by he weed seed dispe sal me hods o each illage sys em. Thus, in CT he soil is mixed and he seedbank is mo e o less uni o mly dispe sed by illage ope a ions esul ing in highe e enness alues, while in he MT and NT ligh weigh seeds, as o example he wind-dispe sed P. echioides o Conyza spp. (Sa age e al., 2014) a e li le bu ied o no bu ied a all i he MT ope a ion has been al eady ca ied ou so ha i may be plausible ha his kind o seeds a e no easily e enly dis ibu ed unde MT.