Effects of mechanical and chemical methods on weed control, weed seed rain and crop yield in maize, sunflower and soyabean

Crop Protection 64 (2014) 51e59

Contents lists available at ScienceDirect

Crop Protection journal homepage: www.elsevier.com/locate/cropro

Effects of mechanical and chemical methods on weed control, weed seed rain and crop yield in maize, sunflower and soyabean Euro Pannacci*, Francesco Tei Department of Agricultural, Food and Environmental Sciences, University of Perugia, Borgo XX Giugno, 74 - 06121 Perugia, Italy

a r t i c l e i n f o

a b s t r a c t

Article history: Received 11 July 2013 Received in revised form 26 May 2014 Accepted 2 June 2014 Available online

Eight field experiments with maize (Zea mays L.), sunflower (Helianthus annuus L.) and soyabean (Glycine max (L.) Merr.) were carried out in central Italy in order to evaluate the effects of mechanical and chemical methods (spring-tine harrowing, hoeing, hoeing-ridging, split-hoeing, finger-weeding, herbicides in the row þ inter-row hoeing, herbicides broadcast) on weed control, weed seed rain and crop yield. The choice of chemical and mechanical treatments in maize and soyabean compared to sunflower, required to be managed more carefully in order to maximize the weed control reducing yield losses. A global rating of weed control methods, based on their weed control efficacy, was obtained as useful means to assist farmers and technicians to choose the more appropriate weed control method. The combination of herbicides intra-row and hoeing inter-row gave best efficacy (on average 99% of weed control), with a 50% reduction in the chemical load in the environment. Hoeing-ridging gave good results, both inter- and intra-row (on average 93% of weed control); this method was also effective in reducing competitive ability and seed production of uncontrolled weeds. Split-hoeing or finger-weeding showed some limitations giving satisfactory results only when combined. Harrowing gave lowest weed control, although when combined to other mechanical methods, can help achieve a better efficacy. © 2014 Elsevier Ltd. All rights reserved.

Keywords: Integrated weed management Mechanical inter and intra-row weed control Maize Sunflower Soyabean Weed seed rain

1. Introduction Over the last fifteen years, environmental and human health impact of herbicides use, increasing of herbicide resistance, the scarce availability of herbicides for minor crops such as vegetables and the increased of organic farming were the main factors that stimulated the interest to develop new methods for mechanical weed control to use alone or with herbicides in integrated weed control strategies (Melander et al., 2005; Cloutier et al., 2007). Recently, Harker and O'Donovan (2013) stressed as “given herbicide-resistant weed issue and consistent public pressure to reduce overall pesticide use, herbicide alternatives and true integrated weed management (IWM) strategies are urgently required now more than ever.” Furthermore, the authors noticed as the importance of using alternatives to herbicides for weed control was recognized long ago (from 1929); although unfortunately, in modern agriculture, non chemical weed control methods have not always held a prominent place, and too often is common a “false

* Corresponding author. Tel.: þ39 075 5856342; fax: þ39 075 5856344. E-mail address: [email protected] (E. Pannacci). http://dx.doi.org/10.1016/j.cropro.2014.06.001 0261-2194/© 2014 Elsevier Ltd. All rights reserved.

integration” consisting of only chemical control components (i.e. different ways of applying herbicides, applying different herbicidal mode of action). Many others authors have challenged weed researchers to increase emphasis on IWM systems and alternatives to herbicides in order to develop systems that give producers more flexibility and options (Wyse, 1992; Buhler, 1999; Hamill et al., 2004). In this context, the priority to implement a true integrated weed management is mainly required in industrial crops where there is a large availability and application of herbicides. The most important industrial crops in Italy, excluding soft and durum wheat, are maize, soyabean and sunflower with 808,317 ha, 174,934 ha and 107,000 ha of arable areas, respectively (Istat, 2013). In row crops, although weeds between the rows (inter-row weeds) can normally be controlled by ordinary inter-row cultivation, such as hoeing, weeds that grow within the line of row crop plants (intra-row weeds) have a great impact on yield and constitute a major problem for selective control, especially for organic farmers (Vangessel et al., 1995; Melander and Rasmussen, 2001; Ascard and Fogelberg, 2002; Pannacci et al., 2007a; Melander et al., 2012). For intra-row weed control, most mechanical methods are based on old principles, but new implements and improved versions have emerged lately, such as finger-weeder,

52

E. Pannacci, F. Tei / Crop Protection 64 (2014) 51e59

torsion-weeder and intelligent weeders (Van der Weide et al., 2008; Rasmussen et al., 2012). Most of the European researches on mechanical weed control have been published through the workshops proceedings of the Physical and Cultural Weed Control Working Group, organized under the European Weed Research Society (http://www.ewrs.org/ pwc). Some authors have summarized various suitable machinery adjustments for different crops and weed stages, in the form of technical guide for farmers in order to increase a correct use of mechanical weed control methods (Van der Schans et al., 2006). In Italy, the mechanical methods used traditionally for weed control in maize, soyabean and sunflower are hoeing and hoeingridging. Over the last ten years new mechanical weed control methods such as split-hoeing, finger-weeding and harrowing were introduced in order to give farmers more flexibility and options rberi, 2000). However, there is a low availability of (Frondoni and Ba data on the performance of mechanical weed control methods obtained from field experiments in different years and in different crops. These data seems to be necessary and useful to assist farmers and technicians to select the more appropriate weed control method in order to adopt a rational Integrated Weed Management. For these reasons, the aim of this study was to evaluate the effects of mechanical and chemical methods on weed control, weed seed rain and crop yield in maize, sunflower and soyabean in central Italy. The mechanical treatments involved in this study were chosen with the aim to compare weed control methods traditionally used in Italy (i.e. hoeing and hoeing-ridging) with weed control methods relatively new such as split-hoeing, finger-weeding and harrowing. Several initial studies have supported this choice, showing that these mechanical methods may have application in maize, soybean and sunflower (Balsari et al., 2002; Raffaelli et al., 2002a, 2002b; Pannacci and Covarelli, 2003). 2. Materials and methods From 2002 to 2005, eight field experiments with maize, sunflower and soyabean were carried out in central Italy (Tiber valley, Perugia, 42 570 N e 12 220 E, 165 m a.s.l.) on a clay-loam soil (24.8% sand, 30.4% clay and 0.9% organic matter). The trials were carried out according to good ordinary practices, as concerns soil tillage and seedbed preparation (Bonciarelli and Bonciarelli, 2001); in all cases, soft winter wheat was always the preceding crop. Experimental design was always a randomized block with three replicates and plot size of 45 m2 (3 m width). In each crop, different weed control methods were compared (Table 1) and untreated plots were added as checks. Harrowing was applied before other mechanical treatments (hoeing, hoeing-ridging and split-hoeing þ finger-

weeding) with the aim to reduce initial competitiveness of weeds toward crops (Table 1). However, in 2004 and 2005 harrowing was also applied alone in order to know its weed control ability. Herbicides applied in broadcast applications and in band on the row integrated with hoeing in inter-row were added as chemical control and as integrating of chemical and mechanical control, respectively (Table 1). Herbicide were always applied with a knapsack plot sprayer fitted with four flat fan nozzles (Albuz APG 110 e Yellow) and calibrated to deliver 300 L ha1 spray solution at 200 kPa pressure; applications were performed broadcast or in band on the row (50% of total surface). Hoeing, an inter-row mechanical control, was carried out with a 3 m-wide powered rotary hoe (Model CERES, Badalini, Italy, http:// www.badalini.it/home_en.php?azione¼scheda_prodotto_en&id¼ 50) at a cultivation depth of 50e60 mm, a driving speed of 4 km h1 and leaving 120-mm untilled strip in the crop rows. Hoeing-ridging was carried out with the same rotary hoe as mentioned above, but equipped with ridging implements to bury weeds along the row. Harrowing, a full surface mechanical control, was carried out with a 3 m-wide spring-tine harrow (Type SF-30, Faza, Italy, http://www.fazasrl.com/index_inglese.htm, equipped with 7 mm-diameter flexible tines) at a cultivation depth of 10e20 mm and a driving speed of 7 km h1. Split-hoeing was performed with an Asperg Gartnereibedarf split-hoe (Asperg, Germany, for more details see Tei et al., 2002) at a cultivation depth of 30e40 mm, a driving speed of 3 km h1 and leaving a 100-mm untilled strip in the crop rows. Split-hoe is an inter-row mechanical mean equipped with goosfoot tine cultivators in front and rotors with steel tine in rear moved by hydraulic power. The goosfoot tine cultivators penetrate and lift the earth, the rotors, turning in the direction of travel between the rows, intercept and crumble the soil and separate (split) earth and weeds. The weeds remain on the soil surface and die quickly. Metal crop shields (100 mm wide) protect crops from moving soil. Finger-weeding, an intra-row mechanical control, was carried out with a Kress finger-weeder (Kress Umweltschonende Landtechnik, Germany, http://www.kress-landtechnik.de/wEnglisch/ produkte/gemuesebau/hacktechnik/fingerhacke_start.shtml? navid¼12) at a cultivation depth of 10e30 mm and a driving speed of 5 km h1. Kress finger-weeder equipments were mounted on Kress Argus System (Kress Umweltschonende Landtechnik, Germany, http://www.kress-landtechnik.de/wEnglisch/produkte/ gemuesebau/hacktechnik/argus_start.shtml?navid¼19) equipped with special-flat share type “Holland” (340 mm wide, http:// www.kress-landtechnik.de/wEnglisch/produkte/gemuesebau/ hacktechnik/hackwerkzeug/hackwerkzeuge_start.shtml?navid¼ 31) that works between the rows. Rubber fingers grip from the

Table 1 Treatments in the field experiments with maize, sunflower and soyabean. Treatments, relative times and codes* 1st treatment Herbicides broadcast (HB) Herbicides on row (HR) þ Harrowing (HA) Harrowing (HA) þ Hoeing (HO) Hoeing-ridging (HOR) Harrowing (HA) þ Harrowing (HA) þ Split-hoeing (SH) Finger-weeding (FW) Split-h. (SH) þ Finger-w. (FW) Harrowing (HA) þ

Maize 2nd treatment Hoeing (HO) Harrowing (HA)

Hoeing (HO) Hoeing-ridging (HOR)

Split-h. (SH) þ Finger-w. (FW)

Sunflower

Soyabean

2002

2003

2004

2002

2003

2004

2004

2005

X X e e e X e X X X X e

X X e e e X e X X X X e

X X X e e X e X X X X X

X X e e X X e e X X X e

X X e e X X e e X X X e

X X X e X X X e X X X X

X e X X e e X e X X X e

X e X X e e X e X X X e

*Each treatment was applied only one time. First and second treatments were carried out in two different periods.

E. Pannacci, F. Tei / Crop Protection 64 (2014) 51e59

side around the plant and there they hoe the weeds. In this way, the area which no other mechanical hoe usually reaches will be weeded as well. Special-flat share cuts the weeds between the rows that remain on the soil surface and die. Preliminary tests were carried out in order to set the implements with the aim to obtain a level of cultivation intensity able to guarantee the highest efficacy against the weeds with the lowest crops damage. 2.1. Maize Maize, cv. DK440 (FAO class 300), was sown on 29 April 2002, 24 April 2003 and 28 April 2004 in 0.5 m-spaced rows, at a seeding rate of 14 seeds m2. After emergence, maize seedlings were thinned to a final density of 7 plants m2. In all years, 75 kg ha1 P2O5 were applied pre-sowing, while 150 kg ha1 N were applied at sowing time. A low-irrigation regime was adopted, with one irrigation in June and another in July (30 mm each). Pre-emergence broadcast and in band herbicide applications were carried out with metolachlor (1449 g a.i. ha1) þ terbuthylazine (725 g a.i. ha1), according to the most common weed control strategies with herbicides in central Italy (Pannacci and Covarelli, 2009). Mechanical treatments were performed with the crop, broadleaved weeds and grasses at the growth stage shown in Table 2. Maize was harvested on 07 October 2002, 12 September 2003 and 24 September 2004.

40 plants m2. A low-irrigation regime was adopted, with one irrigation in June and two in July (30 mm each). Mechanical treatments were performed with the crop, broadleaved weeds and grasses at the growth stage shown in Table 2. Post-emergence herbicides (bentazone 374.4 g a.i. ha1 þ fomesafen 104 g a.i. ha1 þ cycloxydim 200 g a.i. ha1) were sprayed in broadcast application, with the crop, broadleaved weeds and grasses at the growth stage shown in Table 2. Soyabean was harvested on 29 September 2004 and 27 October 2005. 2.4. Data collection and analysis About four weeks after mechanical treatments, weed ground cover (%) was rated visually on the central part of each plot (26 m2), using the BrauneBlanquet cover-abundance scale (Maarel, 1979). Afterwards, weeds on three squares (0.5  0.5 m each one 0.25 m2) per plot were collected, counted, weighed, dried in oven at 105  C to determine moisture content. The squares were posed on three rows (one on each row) of the central part of the plot (26 m2): the square position was random along the row but centred on the row, such as reported by Kurstjens and Bleeker (2000). Data on weed ground cover, weed density and weed dry weight were used to calculate weed reduction (WR) due to the different treatments, according to Chinnusamy et al. (2013):

WRð%Þ ¼

2.2. Sunflower Sunflower, cv. Sanbro, was sown on 15 April 2002, 16 April 2003 and 06 April 2004 in 0.5 m-spaced rows at a seeding rate of 10 seeds m2. After emergence, sunflower seedlings were thinned to a final density of 5 plants m2. In all years, 70 kg ha1 P2O5 were applied pre-sowing, while 100 kg ha1 N were applied at sowing time. Pre-emergence broadcast and in band herbicide applications were carried out with metobromuron (750 g a.i. ha1) þ pendimethalin (1250 g a.i. ha1), according to the most common weed control strategies with herbicides in central Italy (Pannacci et al., 2007b). Mechanical treatments were performed with the crop, broadleaved weeds and grasses at the growth stage shown in Table 2. Sunflower was harvested on 13 September 2002, 20 August 2003 and 06 September 2004. 2.3. Soyabean Soyabean, cv. Sapporo, was sown on 28 May 2004 and 09 May 2005 in 0.5 m-spaced rows to obtain a final density of

53

WU  WT  100 WU

where, WU: weed ground cover/density/dry weight in untreated plots. WT: weed ground cover/density/dry weight in treated plots. Weed seed rain was assessed in 2002 and 2003 in maize, and in 2005 in soyabean, using three seed traps for each plot. Each trap consisting of 138-mm-diam Petri-dishes without cap, filled with a 3 mm layer of quartz-sand. Traps were kept on the soil surface in the central part of each plot, from 29 July to 01 October 2002 and from 10 July to 12 September 2003 in maize and from 21 July from 10 October in soyabean. The content of each trap (sand þ seeds þ plant residues) was removed every twenty days and brought to the laboratory for further processing, while traps were refilled with new quartz-sand. Seeds were manually separated by means of a backlit magnifying glass (8x) and were identified and counted with the aid of an optical microscope (45x). At harvesting time, the crop density and grain yield (adjusted to 9%, 10% and 15.5% of moisture content for sunflower, soyabean and maize, respectively) was determined by hand-harvesting the

Table 2 Crops and weeds stage of growth at the two treatment periods. Treatments* 1st HB HR þ HA HA þ HO HOR HA þ HA þ SH FW SH þ FW HA þ

Maize 2nd

HO HA

HO HOR

SH þ FW

Sunflower

Soyabean

Crop

Broad leaved weeds

Grass weeds

Crop

Broad leaved weeds

Grass weeds

Crop

Broad leaved weeds

Grass weeds

pre pre þ 4e5 3 e e 4e5 e 3 þ 4e5 3e4 3e4 3e4 3 þ 3e4

pre pre þ 4 2 e e 4 e 2þ4 2e4 2e4 2e4 2 þ 2e4

pre pre þ 2e3 1e2 e e 2e3 e 1e2 þ 2e3 2 2 2 1e2 þ 2

pre pre þ 6 4 e 6 6 4þ6 e 6 6 6 4þ6

pre pre þ 6 2e4 e 6 6 2e4 þ 6 e 6 6 6 2e4 þ 6

pre pre þ 4e5 1e2 e 4e5 4e5 1e2 þ 4e5 e 4e5 4e5 4e5 1e2 þ 4e5

1e2 e 1 1 þ 1e2 e e 1þ2 e 1e2 1e2 1e2 e

6 e 2e4 2e4 þ 6 e e 2e4 þ 6e8 e 6 6 6 e

4e5 e 2 2 þ 4e5 e e 2þ5 e 4e5 4e5 4e5 e

(*) see treatment codes in Table 1; pre ¼ pre-emergence; numbers in the table indicate the number of true leaves or trifoliolate leaves in soyabean.

54

E. Pannacci, F. Tei / Crop Protection 64 (2014) 51e59

Table 3 Ground cover (%), density (n. m2) and dry weight biomass (g m2) of total weed species in the untreated checks of the experimental trials. Crops

Maize Sunflower Soyabean

2002

2003

2004

2005

Ground cover (%)

Density (n. m2)

Dry weight (g m2)

Ground cover (%)

Density (n. m2)

Dry weight (g m2)

Ground cover (%)

Density (n. m2)

Dry weight (g m2)

Ground cover (%)

Density (n. m2)

Dry weight (g m2)

223 (14.2) 122 (17.6)

69 (1.8) 57 (4.6)

334 (49.4) 62 (6.7)

158 (13.9) 130 (9.1)

65 (11.3) 85 (22.2)

306 (18.9) 55 (20.1)

183 (4.6) 153 (27.1) 263 (17.4)

59 (2.2) 141 (54.6) 89 (12.0)

265 (62.6) 211 (78.7) 245 (49.5)

132 (2.9)

48 (15.6)

113 (28.8)

Standard errors are in parentheses.

central part of each plot (26 m2). Crop density was used as indicator of treatments selectivity on the basis of uprooted crop plants, such as showed by Kurstjens (2000). Crop yield was related to the total weed dry matter by linear regression Y ¼ A þ BX, with total weed dry matter (independent variable X) and crop yield (dependent variable Y). Prior to ANOVA, data on weed percentage reduction in terms of ground cover, density and dry weight were arcsine-transformed and data on weed seed rain were log-transformed (Box and Cox, 1964). Means were separated by Fisher's protected LSD test at p ¼ 0.05. Analysis of variance was performed with the EXCEL® Addin macro DSAASTAT (Onofri, 2006).

showed a satisfactory efficacy only against weeds at early growth stages (2e4 true leaves or less), especially in the case of grasses. Indeed, FW gave worst results, particularly in 2002 and 2003, due to the presence of E. crus-galli. The combination of splithoeing þ finger-weeding (SH þ FW) did not remarkably improve the results obtained with SH alone. In 2004, harrowing (HA) showed a very low weed reduction both alone (from 30% to 52%) and followed by split-hoeing þ finger-weeding (HA þ SH þ FW). Indeed, in this latter case, weed reduction was never significantly higher than that of split-hoeing þ finger-weeding (SH þ FW). In 2002 and 2003, hoeing-ridging (HOR) showed a weed dry weight reduction significantly higher than split-hoeing þ fingerweeding (SH þ FW) without significant differences between them in terms of weed density reduction. This means that hoeing-ridging is more effective than split-hoeing þ finger-weeding in containing the growth of uncontrolled weeds, thanks to the “burial effect” of the ridging on weeds. In other words, surviving weeds showed less damage and more vigour after split-hoeing þ finger-weeding than after hoeing-ridging. Similar results were also observed from other authors (Rasmussen, 1993; Jones et al., 1995; Rasmussen and Rasmussen, 1995). Seed rain was higher in 2002 than in 2003 (Table 4), not only due to a different weed flora composition, but mainly due to an higher temperature and lower rainfall in July, August and September 2003 (Fig. 1b), which obstacle weed growth. However, in both years the lowest seed rain was observed with herbicide treatments (reduction from 95 to 99%, in comparison to the untreated check). Considering mechanical methods, hoeing-ridging and harrowing þ hoeing-ridging gave lowest values of seed rain in 2002. On the other hand, in 2003 differences among mechanical methods were not statistically significant. It is however to be pointed out as also a low number of weeds that survive to the mechanical treatment may produce an appreciable amount of seeds, especially if the method of control is not able to affect negatively their growth ability after treatment. Similar remarks

3. Results and discussion 3.1. Maize Total weed flora was quite constant in the three years with a ground cover on the untreated check ranging from 158% to 223% and a density ranging from 59 to 69 plants m2 (Table 3). The main weed species were: Amaranthus retroflexus (all the years); Chenopodium album (all the years); Portulaca oleracea (all the years); Echinochloa crus-galli (2002 and 2003); Polygonum persicaria (2002); Ammi majus (2002); Polygonum aviculare (2002) and Fallopia convolvulus (2003). In all years, chemical weed control alone (HB) or integrated with inter-row hoeing (HR þ HO) gave the highest efficacy (Table 4). Considering mechanical methods, the highest level of weed control was obtained by harrowing þ hoeing-ridging (HA þ HOR) with values of weed reduction from 88% to 97% (Table 4); hoeingridging (HOR) alone gave a slightly lower weed control level, even though the differences between them were not statistically significant. Split-hoeing (SH) also provided good control of weed with values of weed reduction from 71% to 90%. Finger-weeding (FW)

Table 4 Maize: effect of different weed control methods on ground cover, density, biomass and seed rain of total weeds. Treatments

Total seeds (n. m2)

Weed reduction (%) Ground cover

Density

Dry weight

2002

2003

2004

2002

2003

2004

2002

2003

2004

2002

2003

Untreated check Herbicide broadcast Herbicide on row þ Hoeing Harrowing Hoeing-ridging Harrowing þ Hoeing-ridging Split-hoeing Finger-weeding Split-hoeing þ Finger-weeding Harr. þ S.-hoeing þ F.-weeding

99 ab 100 a e 94 cd 97 bc 87 d 66 e 86 d e

99 a 100 a e 91 bc 95 b 71 d 57 e 87 c e

100 a 100 a 52 d 92 bc 95 b 86 bc 84 c 82 c 89 bc

100 a 100 a e 89 b 92 b 76 c 56 d 86 b e

99 ab 100 a e 92 bc 95 bc 90 cd 82 d 94 bcd e

100 a 100 a 44 e 92 b 89 bc 86 bcd 77 d 81 cd 84 bcd

100 a 100 a e 92 bc 94 ab 80 cd 49 e 69 de e

100 a 100 a e 95 a 96 a 71 bc 54 c 80 b e

100 a 100 a 30 c 88 b 88 b 77 b 68 b 79 b 82 b

57,399 a 223 e 974 d e 4,868 c 4,229 c 11,185 bc 24,675 ab 19,777 b e

18,105 a 438 c 1,546 c e 8,829 ab 6,719 b 6,065 b 13,326 ab 4,980 b e

S.E.M.

2.7

2.7

4.2

4.0

2.8

3.7

6.0

6.5

7.9

4,373.8

2,673.0

In each column, values followed by the same letter are not significantly different according to the Fisher's protected LSD test (P ¼ 0.05), performed on arcsine transformed data (weed reduction) and on log transformed data (total seeds).

Fig. 1. Average decade values of rainfall (mm; bold bar) and temperature ( C; solid line) recorded during the experimental trials in 2002 a), 2003 b), 2004 c) and 2005 d) compared to pluriennial (from 1921) averages (rainfall: mm, empty bar; temperature:  C, sketched line).

56

E. Pannacci, F. Tei / Crop Protection 64 (2014) 51e59

Table 5 Maize: effects of different weed control methods on crop yield. Crop yield (t ha1)

Treatments

2002

2003

2004

Untreated check Herbicide broadcast Herbicide on row þ Hoeing Harrowing Hoeing-ridging Harrowing þ Hoeing-ridging Split-hoeing Finger-weeding Split-hoeing þ Finger-weeding Harrowing þ Split-hoeing þ Finger-weeding

1.59 e 10.07 a 9.57 ab e 8.75 abc 8.52 bc 8.00 c 5.51 d 8.10 bc e

2.64 7.18 7.02 e 6.75 6.44 5.50 5.64 6.63 e

4.09 c 10.09 a 9.82 a 6.99 b 10.30 a 9.39 ab 8.51 ab 7.86 ab 8.95 ab 9.51 ab

S.E.M.

0.486

0.473

d a ab abc abc c bc abc

0.937

In each column, values followed by the same letter are not significantly different according to the Fisher's protected LSD test (P ¼ 0.05).

11 10 2002 2003 2004

Maize yield (t ha-1)

9 8 7 6 5 4 3 2

conditions (Fig. 1b). The untreated check showed a reduction in yield, with respect to herbicide broadcast (highest yield), of 8.48 t ha1, 4.54 t ha1 and 6.00 t ha1, respectively in 2002, 2003 and 2004. However, it is to be noted that the yield obtained with the most effective mechanical treatments (i.e. hoeing-ridging, harrowing þ hoeing-ridging and split-hoeing þ finger-weeding) was never significantly lower than that observed with chemical control. Split-hoeing and finger-weeding gave the lowest yield values, together with harrowing in 2004. The parameters of linear regression (see standard error in parenthesis) between crop yield and the total weed dry matter were: A ¼ 9.64(0.23), B ¼ 0.024(0.002), R2 ¼ 0.97 in 2002; A ¼ 7.05(0.17), B ¼ 0.014(0.001), R2 ¼ 0.95 in 2003; A ¼ 10.17(0.26), B ¼ 0.021(0.002), R2 ¼ 0.91 in 2004. These regressions showed that when total dry weight of uncontrolled weeds increased of 100 g m2, crop yield decreased of 2.4 t ha1 in 2002, 1.4 t ha1 in 2003 and 2.1 t ha1 in 2004 (Fig. 2); that correspond respectively to a crop yield loss of 24.9%, 19.9% and 20.6% with respect to highest crop yield obtained in absence of weeds. These results confirm the sensitiveness of maize to weed competition during critical period that may produce severe crop yield losses (Rajcan and Swanton, 2001) and showed as mechanical weed methods need to be effective not only to decrease weed density but also to contain the growth and competitive ability of uncontrolled weeds. Integrating inter-row cultivation with 50% reduced levels of herbicide by banding, maintained weed control and corn yield compared with the broadcast herbicide treatment and was an effective weed management option, as showed also by Buhler et al. (1995) and Pleasant et al. (1994).

1

3.2. Sunflower

0 0

50

100

150

200

250

300

350

Dry weight of weeds (g m-2) Fig. 2. The relationships between total dry weight of weeds and maize yield. Symbols show observed data, lines show linear regressions.

were made by Mertens and Jansen (2002) and Lutman (2002) in other experiments. In all years, weed control methods didn't show significant differences in crop density (data not shown), without uprooted crop plants after treatments. Average grain yield in 2003 was clearly lower than in the other two years (Table 5), due to the above mentioned environmental

Total weed flora was quite constant in 2002 and 2003 with a ground cover on the untreated check ranging from 122% to 130% and a density ranging from 57 to 85 plants m2 (Table 3). On the other hand, weed infestation was higher in 2004 than in the previous years, especially in terms of density and weight of weeds. The main weed species were: A. retroflexus (all the years); C. album (all the years); E. crus-galli (all the years); Polygonum lapathifolium (2004); F. convolvulus (2003) and Solanum nigrum (2002). Chemical weed control alone or integrated with inter-row hoeing gave always the best weed control and showed a weed reduction ranging from 91% to 100% (Table 6). Considering mechanical methods, the highest efficacy was obtained by hoeing-ridging (HOR) that was not significantly worse

Table 6 Sunflower: effect of different weed control methods on ground cover, density and biomass of total weeds. Treatments

Weed reduction (%) Ground cover

Density

Dry weight

2002

2003

2004

2002

2003

2004

2002

2003

2004

Herbicide broadcast Herbicide on row þ hoeing Harrowing Hoeing Harrowing þ hoeing Hoeing-ridging Split-hoeing Finger-weeding Split-hoeing þ Finger-weeding Harrowing þ S.-hoeing þ F.-weeding

99 a 100 a e 75 d e 96 b 88 c 56 e 94 b e

91 98 e 76 e 93 85 68 91 e

100 a 98 ab 47 g 52 fg 86 cd 94 bc 66 ef 62 efg 71 de 90 bc

100 a 100 a e 43 e e 95 b 73 d 46 e 89 c e

95 97 e 79 e 90 81 73 87 e

99 96 70 75 84 96 72 77 78 94

100 a 100 a e 46 d e 95 b 70 c 56 cd 90 b e

98 99 e 72 e 90 85 61 85 e

100 a 90 ab 33 d 63 bcd 73 bc 88 abc 64 bcd 35 d 56 cd 70 bcd

S.E.M.

1.6

3.0

6.1

3.2

3.8

6.3

7.8

b a cd ab bc d b

ab a cd bc cd d bc

6.4

a ab d d bcd ab d d cd abc

ab a cd bc cd d c

12.1

In each column, values followed by the same letter are not significantly different according to the Fisher's protected LSD test (P ¼ 0.05), performed on arcsine transformed data.

E. Pannacci, F. Tei / Crop Protection 64 (2014) 51e59

than chemical weed control. Especially in 2002, hoeing (HO) alone did not show a good weed control, due to the lower intra-row weed reduction. Split-hoeing provided a satisfactory control of weeds with weed reduction values ranging from 64% to 88%. Finger-weeding was effective only against weeds at early growth stages and its efficacy was rather low, especially in 2004 due to high infestation of E. crusgalli and P. lapathifolium. Likewise, the combination of fingerweeding with split-hoeing did not cause a significant weed control increase compared to split-hoeing alone, except in 2002. Harrowing (HA) alone gave the worst results (weed reduction from 33% to 70%) due to low efficacy against P. lapathifolium and E. crus-galli; when combined to hoeing (HA þ HO) or splithoeing þ finger-weeding (HA þ SH þ FW), it did not help achieve a better efficacy. All treatments didn't show significant differences in crop density (data not shown), without uprooted crop plants after treatments. Differences in crop yield were very small, also owing to high sunflower tolerance and competitiveness (Pannacci et al., 2007b). It should be noted that several mechanical methods gave yield levels comparable to herbicide treatments (Table 7). The linear regression between crop yield and the total weed dry matter was not statistically significant (data not shown). 3.3. Soyabean Total weed flora was twice in 2004 with respect to 2005 with a ground cover on the untreated check of 263% and 132% and a density of 89 and 48 plants m2 in 2004 and 2005, respectively (Table 3). In particular, the weed infestation lower in 2005 than in 2004, was due to sporadic presence of grass weeds and to low presence of broadleaved weeds. The main weed species were: A. retroflexus (both years); C. album (both years); E. crus-galli (both years); P. oleracea (both years); Sinapis arvensis (both years); Digitaria sanguinalis (2004) and S. nigrum (2004). Broadcast herbicides gave the highest reduction in term of weed ground cover and weed dry weight, but it was not as effective in reducing weed density, especially in 2004 (Table 8). Indeed, at the time of sampling, post-emergence herbicides had not completed their action and several weeds were still alive, even though they had already suffered severe damage with low ground cover and biomass weight. Considering mechanical methods, weed control was lower in 2004 than in 2005 due to high weed density with presence of grass Table 7 Sunflower: effects of different weed control methods on crop yield. Treatments

Crop yield (t ha1) 2002

2003

2004

Untreated check Herbicide broadcast Herbicide on row þ Hoeing Harrowing Hoeing Harrowing þ Hoeing Hoeing-ridging Split-hoeing Finger-weeding Split-hoeing þ Finger-weeding Harrowing þ Split-hoeing þ Finger-weeding

3.89 c 4.19 ab 4.39 a

3.81 c 3.91 bc 4.14 a

e

e

2.37 3.40 4.10 3.13 3.17 3.60 3.60 3.02 3.14 3.09 3.77

S.E.M.

0.090

0.073

0.321

e

e

4.11 bc

3.98 abc

e

e

4.25 4.26 4.21 4.44

ab ab ab a

3.95 3.77 3.87 4.04

abc c bc ab

c ab a bc abc ab ab bc bc bc ab

In each column, values followed by the same letter are not significantly different according to the Fisher's protected LSD test (P ¼ 0.05).

57

Table 8 Soyabean: effect of different weed control methods on ground cover, density, biomass and seed rain of total weeds. Treatments

Untreated check Herbicide broadcast Harrowing Harrowing þ Harrowing Harrowing þ Hoeing Split-hoeing Finger-weeding Split-hoeing þ Finger-weeding S.E.M.

Weed reduction (%) Ground cover

Density

Dry weight

2004

2005

2004

2005

2004

95 a

99 a

27 d

83 ab 63 a

39 e 52 de

46 f 60 ef

75 bc

Total seeds (n. m2)

2005

2005

98 a

80,838 a 1,903 d

30 cd 27 c 8c 30 cd 59 bc 0 c

24 c 32 c

89,048 a 52,979 ab

94 ab

82 a

92 ab

18,408 c

69 bc 59 cd 76 b

82 cd 68 de 89 bc

56 bc 74 ab 17 bc 57 abc 32,724 b 37 cd 66 ab 13 bc 44 bc 32,247 b 65 ab 85 a 42 ab 37 bc 35,533 b

5.5

4.9

8.4

79 ab 67 a

10.0

9.7

19.1

15,623.3

In each column, values followed by the same letter are not significantly different according to the Fisher's protected LSD test (P ¼ 0.05), performed on arcsine transformed data (weed reduction) and on log transformed data (total seeds).

(E. crus-galli and D. sanguinalis) less sensitive to uprooted action of mechanical tools (Table 8). Harrowing þ hoeing (HA þ HO) showed the highest weed control efficacy among mechanical treatments and seemed to be a good alternative to chemical control as observed by Gunsolus (1990). Split-hoeing and finger-weeding showed similar results with a slight increase of weed control when combined. Harrowing gave very low efficacy, regardless of the number of treatments as observed by Rasmussen et al. (2010). Seed rain showed to depend from weed reduction ability of the treatments, especially in terms of dry weight, as already observed in maize (Table 8). In particular, the lowest seed rain was observed with herbicide treatments (weed dry weight reduction of 98%), while among mechanical methods, lowest seed rain was obtained by harrowing þ hoeing (HA þ HO). Split-hoeing and fingerweeding showed seed rain values from 32,247 to 35,533 seeds m2 that were significant lower that untreated check; while harrowing (HA) alone and in twice application (HA þ HA) gave highest seed rain values statistically comparable to untreated control. In both years, mechanical and chemical treatments did not cause any significant reduction in crop density (data not shown), without uprooted crop plants after treatments. In general, crop yield was very low owing to low irrigation regime (only 90 mm in total) (Table 9). In particular, herbicide treatments gave highest yield values; however, it is to be noted that the yield obtained with

Table 9 Soyabean: effects of different weed control methods on crop yield. Treatments

Crop yield (t ha1) 2004

2005

Untreated check Herbicide broadcast Harrowing Harrowing þ Harrowing Harrowing þ Hoeing Split-hoeing Finger-weeding Split-hoeing þ Finger-weeding

0.61 1.80 0.63 0.79 1.60 1.47 1.04 0.71

0.73 1.64 0.74 0.47 1.43 1.31 1.31 1.47

S.E.M.

0.263

c a c bc a ab abc bc

bc a bc c a ab ab a

0.227

In each column, values followed by the same letter are not significantly different according to the Fisher's protected LSD test (P ¼ 0.05).

58

E. Pannacci, F. Tei / Crop Protection 64 (2014) 51e59

Table 10 Weed reduction (%, standard errors are in parenthesis) and efficacy ratings for different treatments as overall mean of all trial data (n.).

the best mechanical treatment (harrowing þ hoeing) was never significantly lower than that observed with chemical control, confirming as this mechanical treatment could be a good alternative to herbicide, especially in organic and low-input soyabean crop. Treatments based only by harrowing (HA and HA þ HA) gave the lowest yield values, together with untreated check. The linear regression between crop yield and the total weed dry matter was not statistically significant (data not shown). A summary of results were shown in Table 10 and Fig. 3. In particular, the global rating of different weed control methods gave useful information to assist farmers and technicians to choose the more appropriate weed control method (Table 10). Indeed, there are not so much experimental results on the weed efficacy of mechanical treatments that can be used with this scope: the choice to subdivide the different treatments according to an efficacy rating scale goes in this sense (Table 10). In particular, herbicides on row þ hoeing inter-row can be considered an excellent treatment both in the efficacy rating (on average 99% of weed

control) and crop yield (Table 10 and Fig. 3). Hoeing alone gave a fair weed control due to low efficacy against intra-row weeds while combining hoeing with harrowing the double mechanical treatment increased its efficacy, reaching a good level of weed control thank to increasing in the control of intra-row weeds (Table 10). Crop yield indices showed as there was a different response in terms of yield losses under different treatments among the three crops. In particular, it could be noted as the variability (as deviation from the overall mean of all the trials carried out for each crop) of crop yield indices for the different treatments was highest in soyabean, intermediate in maize and lowest in sunflower (Fig. 3). This means that the effects of treatments on crop yield have been more evident in soyabean and maize than in sunflower. Considering that the efficacy of treatments showed similar levels independently from the crops, the differences in crop yield indices could be due to the different competitive ability of the crops against uncontrolled weeds: highest in sunflower, intermediate in maize

Fig. 3. Crop yield indices (overall mean of trials carried out for each crop ¼ 100) of weed control methods (see Table 1 for corresponding code). Indices were grouped based on common treatment: 1) herbicides; 2) harrowing; 3) hoeing and hoeing-ridging; 4) split-hoeing and finger-weeding.

E. Pannacci, F. Tei / Crop Protection 64 (2014) 51e59

and lowest in soyabean. Whereby, seems to be evident as maize and soyabean need to be managed with more carefully, with respect to sunflower, in the choice of chemical and mechanical treatments in order to maximize the weed control reducing yield losses and weed seed rain. 4. Conclusion It is possible to halve the amount of herbicides with no loss in weed control efficacy and crop yield, by combining chemical weed control in the row with hoeing inter-row. If no herbicides have to be used (e.g. in organic farming), mechanical methods can ensure a good selectivity to the crops but they need to be carefully chosen, to avoid losses in weed control ability and crop yield. In our experiments, the best technique was hoeing-ridging, that gave an excellent control of both inter- and intra-row weeds and could hinder the development of uncontrolled weeds, reducing their competitive ability and seed production. This is an important issue, as few surviving weeds could produce an appreciable amount of seeds and increase potential weed infestations. Split-hoeing and finger-weeding showed some limitations. Split-hoeing gave a good inter-row weed control, showing an energetic action against both broadleaves and grasses also in relatively advanced developmental stages. However, this mechanical method did not effectively control weeds along the row. On the other hand, finger-weeding showed a better capability of getting close to the crop row, but it was not effective enough against weeds (especially grasses) with more than 2e4 true leaves. In the cases of scarce grass infestations, these two methods may be successfully combined to obtain a satisfactory inter- and intra-row weed control without appreciable losses in crops yield. Spring tine harrowing alone gave a poor weed control in all the crops determining severe losses on potential crops yield due to competition of uncontrolled weeds. However, when combined to other mechanical methods, spring tine harrow can help achieve a better efficacy. References Ascard, J., Fogelberg, F., 2002. Mechanical intra-row weed control in organic onion production. In: Proceedings 5th EWRS Workshop on Physical Weed Control, Pisa, Italy, p. 125. Balsari, P., Airoldi, G., Ferrero, A., 2002. Mechanical and physical weed control in maize. In: Proceedings 5th EWRS Workshop on Physical Weed Control, Pisa, Italy, pp. 18e31. Bonciarelli, F., Bonciarelli, U., 2001. Coltivazioni Erbacee. Edagricole e Edizioni Agricole. Italy, Bologna, p. 492. Box, G.E., Cox, D.R., 1964. An analysis of transformations. J. R. Stat. Soc. B 26, 244e252, 211e243, discussion. Buhler, D.D., 1999. Expanding the context of weed management. J. Crop Prod. 2, 1e7. Buhler, D.D., Doll, J.D., Proost, R.T., Visocky, M.R., 1995. Integrating mechanical weeding with reduced herbicide use in conservation tillage corn production systems. Agron. J. 87 (3), 507e512. Chinnusamy, N., Chinnagounder, C., Krishnan, P.N., 2013. Evaluation of weed control efficacy and seed cotton yield in glyphosate tolerant transgenic cotton. Am. J. Plant Sci. 4, 1159e1163. Cloutier, D.C., Van der Weide, R.Y., Peruzzi, A., Leblanc, M.L., 2007. Mechanical weed management. In: Upadhyaya, M.K., Re Blackshaw, R.E. (Eds.), Non-chemical Weed Management e Principles, Concepts and Technology, CABI Press, Oxfordshire, UK, pp. 111e134. Frondoni, U., B arberi, P., 2000. Attrezzature per le colture erbacee. Il Contoterzista, Suppl. Macch. Ecol., vol. 5, pp. 19e25. Gunsolus, J.L., 1990. Mechanical and cultural weed control in corn and soybeans. Am. J. Altern. Agric. 5 (03), 114e119. Hamill, A.S., Holt, J.S., Mallory-Smith, C.A., 2004. Contributions of weed science to weed control and management. Weed Technol. 18, 1563e1565. Harker, K.N., O'Donovan, J.T., 2013. Recent weed control, weed management, and integrated weed management. Weed Technol. 27, 1e11. Istat, 2013. Istituto Nazionale di Statistica. Agricoltura e zootecnica. Consultazione dati. Available on line at: http://agri.istat.it/sag_is_pdwout/jsp/Introduzione. jsp?id¼15A/18A.

59

Jones, P.A., Blair, A.M., Orson, J., 1995. The effects of different types of physical damage to four weed species. In: Proceedings Brighton Crop Protection Conference - Weeds, Brighton, UK, pp. 653e658. Kurstjens, D., 2000. Effect of weed & crop variability on selectivity of mechanical weeders. In: Proceedings 4th Workshop on Physical Weed Control, Elspeet, The Netherlands, 20-22 March 2000, pp. 28e29. Available on line at: http://www. ewrs.org/pwc/doc/2000_Elspeet.pdf. Kurstjens, D., Bleeker, P., 2000. Optimising torsion weeders and finger weeders. In: Proceedings 4th Workshop on Physical Weed Control, Elspeet, The Netherlands, 20-22 March 2000, pp. 30e32. Available on line at: http://www.ewrs.org/pwc/ doc/2000_Elspeet.pdf. Lutman, P.J.W., 2002. Estimation of seed production by Stellaria media, Sinapis arvensis and Tripleurospermum inodorum in arable crops. Weed Res. 42, 359e369. Maarel van der, E., 1979. Transformation of cover-abundance values in phytosociology and its effects on community similarity. Vegetatio 39 (2), 97e114. Available on-line at: http://download.springer.com/static/pdf/468/art%253A10. 1007%252FBF00052021.pdf?auth66¼1391703785_2251ddb6505b67827bb239 baf6f7a832&ext¼.pdf. Melander, B., Holst, N., Rasmussen, I.A., Hansen, P.K., 2012. Direct control of perennial weeds between crops - Implications for organic farming. Crop Prot. 40, 36e42. Melander, B., Rasmussen, G., . Effect of cultural methods and physical weed control on intrarow weed numbers, manual weeding and marketable yield in directsown leek and bulb onion. Weed Res. 41, 491e508. Melander, B., Rasmussen, I.A., Barberi, P., 2005. Integrating physical and cultural methods of weed control e examples from European research. Weed Sci. 53, 369e381. Mertens, S.K., Jansen, J.H., 2002. Weed seed production, crop planting pattern, and mechanical weeding in wheat. Weed Sci. 50, 748e756. Onofri, A., 2006. Enhancing excel capability to perform statistical analyses in agriculture applied research. In: Computational statistics and data analysis - Statistical Software Newsletters (Ed.), By International Association for Statistical Computing, 15/02/2006. www.csdassn.org/softlist.cfm. Pannacci, E., Covarelli, G., 2003. Valutazione dell’efficacia di mezzi meccanici tradizionali e innovativi nel controllo delle piante infestanti il girasole. In: Pro Italiana di Agronomia, Portici - Napoli, ceedings XXXV Convegno della Societa Italy, pp. 235e236. Pannacci, E., Covarelli, G., 2009. Efficacy of mesotrione used at reduced doses for post-emergence weed control in maize (Zea mays L.). Crop Prot. 28, 57e61. Pannacci, E., Graziani, F., Covarelli, G., 2007b. Use of herbicide mixtures for pre and post-emergence weed control in sunflower (Helianthus annuus). Crop Prot. 26, 1150e1157. Pannacci, E., Guiducci, M., Tei, F., 2007a. Mechanical weed control in organic onion seed production. In: Proceedings 7th Workshop of the EWRS Working Group: Physical and Cultural Weed Control, Salem (Mecklenburg-Vorpommern), Germany, 11-14 March 2007, pp. 119e120. Available on-line at: http://www.ewrs. org/pwc/doc/2007_Salem.pdf. Pleasant, Mt.J., Burt, R.F., Frisch, J.C., 1994. Integrating mechanical and chemical weed management in corn (Zea mays). Weed Technol. 8 (2), 217e223. Raffaelli, M., Peruzzi, A., Ginanni, M., Di Ciolo, S., 2002a. Mechanical weed control in sunflower and soyabean crops using spring-tine harrow: results of two-year trials. Agric. Mediterr. 132, 112e121. rberi, P., Ginanni, M., 2002b. Options for mechanical Raffaelli, M., Peruzzi, A., Ba weed control in grain maize - work parameters and crop yield. In: Proceedings 5th EWRS Workshop on Physical Weed Control, Pisa, Italy, pp. 153e158. Rajcan, I., Swanton, C.J., 2001. Understanding maizeeweed competition: resource competition, light quality and the whole plant. Field Crops Res. 71 (2), 139e150. Rasmussen, J., . The influence of harrowing used for post-emergence weed control on the interference between crop and weeds. In: Proceedings 8th European Weed Research Society Symposium, Braunschweig, pp. 153e158. Rasmussen, J., Griepentrog, H.W., Nielsen, J., Henriksen, C.B., 2012. Automated intelligent rotor tine cultivation and punch planting to improve the selectivity of mechanical intra-row weed control. Weed Res. 52, 327e337. Rasmussen, J., Mathiasen, H., Bibby, B.M., 2010. Timing of post-emergence weed harrowing. Weed Res. 50, 436e446. Rasmussen, J., Rasmussen, K., . A strategy for mechanical weed control in spring barley. In: Proceedings 8th European Weed Research Society Symposium, Budapest, pp. 557e564. Tei, F., Stagnari, F., Granier, A., 2002. Preliminary results on physical weed control in processing spinach. In: Proceedings 5th EWRS Workshop on Physical Weed Control, Pisa, Italy, 11-13 March 2002, pp. 164e171. Available on-line at: http:// www.ewrs.org/pwc/doc/2002_Pisa.pdf. Van der Schans, D., Bleeker, P., Molendijk, L., Lotz, L.A.P., Bauermeister, R., Total, R., Baumann, D.T., 2006. Practical Weed Control in Arable Farming and Outdoor Vegetable Cultivation Without Chemicals. PPO Publication 532. Applied Plant Research, Wageningen University, Lelystad, The Netherlands, p. 77. Van der Weide, R.Y., Bleeker, P.O., Achten, V.T.J.M., Lotz, L.A.P., Fogelberg, F., Melander, B., 2008. Innovation in mechanical weed control in crop rows. Weed Res. 48, 215e224. Vangessel, M.J., Schweizer, E.E., Lybecker, D.W., Westra, P., 1995. Compatibility and efficiency of in-row cultivation for weed management in corn (Zea mays). Weed Technol. 9 (4), 754e760. Wyse, D.L., 1992. Future of weed science research. Weed Technol. 6, 162e165.