Weed Control by Soil Tillage and Living Mulch

Reduced soil tillage can be classified as minimum, sustainable, conservation, ploughless or zero tillage. For example, in the United Kingdom, such tillage systems are commonly referred to as non-inversion tillage. These different types of non-reversible soil tillage methods maintain at least 30% residue coverage on the soil surface [1]. In reduced soil tillage practices, residue coverage leads to lower moisture evapotranspiration, higher soil water content and soil structural stability, and more effective prevention of soil erosion [2-4]. Compared to conven‐ tional annual deep soil ploughing, reduced tillage may decrease technological production costs and improve the economic effectiveness of agricultural practices [5]. However, weed control in such soil tillage systems is more complex. Reduced soil tillage leads to different weed seed bank distributions in the soil and occasionally lower herbicide effectiveness, which delays the time of weed seed germination because of crop residue coverage [6] and other indices.


Introduction
Reduced soil tillage can be classified as minimum, sustainable, conservation, ploughless or zero tillage. For example, in the United Kingdom, such tillage systems are commonly referred to as non-inversion tillage. These different types of non-reversible soil tillage methods maintain at least 30% residue coverage on the soil surface [1]. In reduced soil tillage practices, residue coverage leads to lower moisture evapotranspiration, higher soil water content and soil structural stability, and more effective prevention of soil erosion [2][3][4]. Compared to conventional annual deep soil ploughing, reduced tillage may decrease technological production costs and improve the economic effectiveness of agricultural practices [5]. However, weed control in such soil tillage systems is more complex. Reduced soil tillage leads to different weed seed bank distributions in the soil and occasionally lower herbicide effectiveness, which delays the time of weed seed germination because of crop residue coverage [6] and other indices.
How much do different soil tillage systems influence the weed infestation of crops? First, weed stand density depends on the competition ability of the crop. Cereals generally have higher competitiveness than do cultivated crops (beet, maize, and potato). For example, Vakali et al. [7] showed that in deeply cultivated plots, the barley crop weed shoot biomass was 65-88% higher than that in reversibly tilled plots, but in rye no clear influence was found. Ozpinar and Ozpinar [8] established that shallow soil rototilling (compared with mouldboard ploughing) increased the total weed density by 72 and 58% in maize and vetch crops, while the differences in wheat were low. Similar results were found by Mashingaidze et al. [9]. In crop rotations with maize, the highest weed stand differences were obtained between mouldboard ploughing and no tillage technologies. Occasionally, no tillage resulted in up to 20 times more weed infestation [10]. The spread of perennial weeds was typically more evident [11]. However, Streit et al. [12] showed that for no tillage technologies without herbicides, the weed density was lower than that in conventional or minimum tillage. Different soil tillage intensities may slightly change the diversity of weed species in crops. In a 23-year experiment by Plaza et al. [13], in minimally tilled plots there were more weed species than in not tilled or traditionally ploughed plots. In a 14-year experiment by Carter and Ivany [14], the weed species diversity was slightly lower in ploughed soil than in shallowly or not tilled soil. In addition, high weed infestation resulted in substantial reductions in maize yield [15].
Worldwide experiments of reduced soil tillage have been widely and well documented, but investigations with maize crops (especially using the no-tillage system) are quite new in lands with low level of herbicide practice.
In chemicalless (ecological, organic, biological or similar) farming systems common problem is high risk of weed infestation. Weeds rival with crops for space, light, water and nutrients. Lazauskas [16] formulated the law of crop productivity, "...crop performance, expressed by the total mass of crops and weeds, is relatively constant and may be defined by the equation: Y = A − bx; Y -crop yield, A -maximum crop productivity, x -weed mass and b -yield depression coefficient". According to this law, the crop yield is inversely proportional to the crop weed mass. Rusu et al. [17] found similar results and concluded that maize green mass production losses could be considered equal to the mass of green weeds.
Living mulches (additional component of agrocenosis) can be useful for effective weed control [18]. According to the Lazauskas law, interseeded living mulch occupies part of total bioproduction and may decrease weed infestation. Nakamoto and Tsukamoto [19] found that "living mulches are cover crops that are maintained as a living ground cover throughout the growing season of the main crop". The winter rye (Secale cereale L.), ryegrasses (Lolium spp.) and subterranean clover (Trifolium subterraneum L.) might be used to suppres weeds in corn crop (Zea mays L.) [20]. However, living mulches can compete for nutrients and water with the main crop and yields of crop could decrease [21,22]. As a result, living mulch plants often must be mechanically or chemically controlled [23,24].

Impact of different primary soil tillage methods on weed infestation
Soil tillage is the main method to control weeds. The most valuable is primary soil tillage. For answer how effective primary soil tillage methods are, the long-term stationary field experiment is being conducted at Aleksandras Stulginskis University's (up to 2011 Lithuanian University of Agriculture) Experimental Station. The field experiment was set up in 1988 in the then Lithuanian Academy of Agriculture's Experimental Station. The soil of the experimental field is Endohypogleyic-Eutric Planosol -PLe-gln-w. The thickness of the soil ploughlayer is 23-27 cm. Soil texture -loam on heavy loam. The upper part of the ploughlayer (0-15 cm) contained: pH KCL -6.6-7.0, available phosphorus -131.1-206.7 mg kg -1 , available potassium -72.0-126.9 mg kg -1 . Primary tillage methods investigated: 1. Conventional ploughing at 23-25 cm depth (CP) (control treatment); 2. Shallow ploughing at 12-15 cm depth (SP); 3. Deep cultivation at 23-25 cm depth (DC); 4. Shallow cultivation at 12-15 cm depth (SC); 5. Not tilled soil (direct sowing) (NT). Crop rotation in the experiment: 1) spring rape; 2) winter wheat; 3) maize; 4) spring barley. The experiment involved 4 replications. Each crop was cultivated in 20 plots. The initial size of plots was 126 m 2 (14 x 9 m), and the size of record plots was 70 m 2 (10 x 7 m). The plots of the experimental treatments were laid out in a randomised order. The protection band of the plot was of 1 m width and that between replications of 9 m width. After crop harvesting, all experimental plots (except for treatment 5) were cultivated by a disc stubble cultivator Väderstad CARRIER 300 at the 12-15 cm depth. JOHN DEERE 6620 tractor was used in the experiment. According to the experimental design, primary tillage was performed in August-September (for winter wheat) or in October (for spring crops). The soil was ploughed with a conventional plough Gamega PP-3-43 with semi-helical mouldboards at the 23-25 cm depth (treatment 1) or at the 12-15 cm depth (treatment 2). Deep cultivation was carried out by a ploughlayer's cultivator (chisel) KRG-3.6 at the 23-25 cm depth (treatment 3). The plots of treatment 4 were additionally cultivated by a disc stubble cultivator Väderstad CARRIER 300 at the 12-15 cm depth. The plots of treatment 5 were not tilled.
In spring, after the soil had reached maturity stage, it was shallow-cultivated by a cultivator Laumetris KLG-3.6 (except for the plots of treatment 5), fertilizers were applied by a fertilizer spreader AMAZONE-ZA-M-1201. Pre-sowing, the soil was cultivated at a seed placement depth. The crops were sown by the following sowing machines: Väderstad Super Rapid Winter wheat. Cultivar 'Ada' in 2010-2012 was sown at a seed rate of 4.5-5 million seeds ha -1 at the 4-5 cm depth. Fertilizers were incorporated at the 6 cm depth. Sowing was performed by a continuous-row method with 12.5 cm wide inter-rows.
Maize. Hybrids 'Pioneer P 8000 (x6T584)' in 2010, 'Pioneer P 8000 (x027)' in 2011 and 'Es capris' in 2012 were sown at a seed rate of 100 thousand seeds ha -1 at the 6 cm depth. Fertilizers were incorporated at the 6.5 cm depth. Sowing was performed by a continuous band wide-row method with 50 cm wide inter-rows (between bands), 12.5 cm wide inter-rows between rows.
Spring barley. Cultivars 'Simba' in 2010 and 2012, 'Tokada' in 2011 were sown at a seed rate of 5-6 million seeds ha -1 at the 3.5 cm depth. Fertilizers were applied by placement method at the 4-4.5 cm depth. Sowing was performed by a continuous-row method with 12.5 cm wide inter-rows.
Weed seed bank in the soil was determined in treatments 0-5 (1 and 5 treatments) at the 0-15, 15-25 cm depths after primary tillage in 20 spots of a record plot in 2010 and 2012. The samples were taken with an auger, and a composite sample was formed. Sampling at the 0-5 cm depth was done to compare the weed seed bank in the upper ploughlayer of the conventionally tilled and not tilled plots. A 100 g dry soil sample was placed on a sieve with 0.25 mm mesh diameter and washed with running water until small soil particles washed out. Weed seeds and the remaining mineral soil fraction were separated from the organic soil fraction using saturated salt (or potash) solution [25].
Crop weed incidence was assessed by identifying weed species composition, weed number at the beginning of vegetation or at resumption of vegetation (winter wheat) during intensive weed growth. Dry weed weight was determined at the end of crops vegetation. Weed incidence was assessed in 10 spots of a record plot in 0.06 m 2 area. At the beginning of vegetation, weed seedlings were counted (weed seedlings m -2 ), and at the end of vegetation weed number (weeds m -2 ) and dry matter weight (g m -2 ) were established. The weeds were pulled out, dried to air-dry weight, and analysis of their botanical species composition was conducted [26].
The research data were statistically processed by the analysis of variance and correlationregression analysis methods. Software ANOVA was used when estimating the least significant difference LSD 05 and LSD 01 . The correlation-regression analysis of the research data was conducted using software STAT and SIGMA PLOT. In the case of significant difference between the specific treatment and the control (reference treatment), the probability level was marked as: * -differences significant at 95 % probability level; ** -differences significant at 99 % probability level.

Weed seed-bank in the soil
The effectiveness of weed control mainly depends on the ability to sweep out weed seed-bank and to prevent the addition with newer ones [27].
Analysis of the data on the effects of different primary tillage on weed seed bank in the soil revealed that nearly in all cases both in not tilled plots and conventionally deep-ploughed plots weed seed bank in the upper ploughlayer (0-5 cm depth) did not differ significantly (data are not presented). In deeper layer (0-15 cm depth) weed seed bank in reduced tillage treatments generally increased, except for not tilled plots, where weed seed bank was less abundant. Only single significant differences were established (Tables 1-4). In the samples taken from the 15-25 cm depth, the weed seed bank was generally smaller. The seeds of annual weeds prevailed in the soil. In many cases, having reduced tillage, the ploughlayer differentiated into upper layer characterized by more abundant weed seed bank (60.1 % of the total weed seed bank) and bottom layer characterized by less abundant weed seed bank (39.9 %). Weed seeds found in conventionally ploughed soil, at the 0-15 cm depth, in different crops accounted for 51.3 to 52.9 % of the total weed seed bank, and in the 15-25 cm depth -from 47.1 to 48.5 %, in shallowploughed soil -55.9-68. 6  Note: * -significant differences from control treatment (conventional ploughing) at 95 % probability level, ** -at 99 % probability level. Note: * -significant differences from control treatment (conventional ploughing) at 95 % probability level, ** -at 99 % probability level. Note: * -significant differences from control treatment (conventional ploughing) at 95 % probability level, ** -at 99 % probability level. Note: * -significant differences from control treatment (conventional ploughing) at 95 % probability level, ** -at 99 % probability level. According to the K. S. Torresen et al. [28] investigations, in top layer of minimally tilled soil there was found higher number of weed seeds than in 10-20 cm depth. Our investigations partlyš comfirms that findings, arable layer devited into upper one with higher number of weed seeds (60.1 % of total number) and deeper layer with less quantity of seeds (Table 5).

Weed spread
Analysis of the data on the effect of different primary tillage on the weed incidence in the crops at the beginning of vegetation revealed that almost in all the cases of reduced tillage or direct drilling into not tilled plots, the number of weeds increased; however, significant difference was estimated only for not tilled winter wheat plots (Tables 6-9). In conventional ploughing treatment, the spread of annual weeds was more intensive. Having replaced conventional ploughing by shallow ploughing, deep and shallow cultivation and direct drilling, the number of annual weeds tended to decrease, while that of perennial weeds tended to increase. Note: * -significant differences from control treatment (conventional ploughing) at 95 % probability level, ** -at 99 % probability level. Note: * -significant differences from control treatment (conventional ploughing) at 95 % probability level, ** -at 99 % probability level. Note: * -significant differences from control treatment (conventional ploughing) at 95 % probability level, ** -at 99 % probability level. Note: * -significant differences from control treatment (conventional ploughing) at 95 % probability level, ** -at 99 % probability level. Note: * -significant differences from control treatment (conventional ploughing) at 95 % probability level, ** -at 99 % probability level. Note: * -significant differences from control treatment (conventional ploughing) at 95 % probability level, ** -at 99 % probability level. Note: * -significant differences from control treatment (conventional ploughing) at 95 % probability level, ** -at 99 % probability level. Note: * -significant differences from control treatment (conventional ploughing) at 95 % probability level, ** -at 99 % probability level.  The correlation-regression analysis of the experimental data revealed that the spread of weeds partly depended on the soil structure and its stability, penetration resistance of deeper soil layers (35-50 cm), moisture content in the upper ploughlayer, soil phosphorus and potassium status, pH, crop stand density, amount of plant residues in the soil surface and weed seed bank in the ploughlayer.

Impact of living mulch on weed infestation
Numerous research and observations have been conducted aiming to establish weed spread methods and reasons and weed-crop competition peculiarities. Enhancement of the competi-tive ability of agricultural crops is one of the principal tools to increase the productivity of agricultural crops. Sowing living mulches between the rows of a main crop is a weed control method that does not employ herbicide application. Living mulches result in reduced field weed infestation and an increase in crop yield. A one-factor, stationary field experiment was conducted. Different living mulches inter-seeded in maize inter-rows were tested. In all experimental years, the same living mulches were inter-seeded in the inter-rows of maize monocrop. The plots of the control treatment were weeded out twice. The experiment was replicated four times. The plots were laid out in a randomised design. The total area of an experimental plot was 24 m 2 , and the area of a record plot was 20 m 2 . In 2009, black fallow preceded maize and in 2010-2011 maize was monocropped.
Maize monocrop inter-seeded with living mulches was grown without chemical pest control under arable agriculture conditions. In spring, when the soil had reached physical maturity, complex NPK 16:16:16 fertilizer at a rate of 300 kg ha -1 was applied, and later the soil was loosened at 4-5 cm depth. Maize was sown by a pneumatic-mechanical drill Köngskilde PRECI -SEM with 50 cm-wide inter-rows and 16-17 cm distance between seeds. Post-emergence of maize, inter-rows were loosened and living mulches were sown with a 7-row manuallyoperated greenhouse seeder. The marginal rows of the inter-seeded living mulches were at 1-2 cm distance from maize. In each experimental year, living mulches were inter-seeded in the Later the practice was abandoned since tractor-hitched implement would not be able to do this. Mulches were cut with a hand-operated brush cutter "Stihl" FS -550, using a designed and manufactured trolley, reducing the operator's load, with a protection hood, which evenly spreads the mulch in the inter-row and protects the crop from mechanical damage. Living mulches were cut after they had reached a height of up to 20-25 cm. Green mass of the living mulches was spread in maize inter-rows. At stem elongation stage (BBCH 31-32), the maize crop was additionally fertilized with nitrogen (N 60 ). When fertilizing at 250 kg N ha -1 rate, no significant differences were observed between maize cultivation systems. The objective of our experiment was to determine the competition among living mulches, maize and weeds; therefore the total nitrogen rate selected was as low as 108 kg N ha -1 . Maize samples for the determination of productivity were hand-cut at the end of September -middle of October (BBCH 87-88) at maize physiological maturity stage. After harvesting, the remaining plant residues were ploughed in by a reversible plough with semi-helical mouldboards at the 20-22 cm depth.
The first assessment of weed infestation in maize crop was made post emergence of crop and weeds. Weeds were counted in 5 randomly selected record plots 0.06 m 2 in size, analysis of weed botanical composition was done, the weeds were dried up to a dry weight and weighed [26]. Weed number was re-calculated into weeds m -2 , dry matter weight into g m -2 . Such assessment was conducted before each cut of living mulches and before maize harvesting. Soil contamination with weed seeds was estimated after maize harvesting. Soil samples were taken with a sampling auger in 10 places of the record plot from the 0-20 cm depth of the ploughlayer. The number of weed seeds found was re-calculated into thousand seeds m -2 [25]. The tests were done in 2009 and 2011.

Weed seed-bank in the soil
Weed seed bank in the ploughlayer was established at the beginning of the experiment in 2009 and at the end in 2011 ( Fig. 1-2). The seeds of Chenopodium album L. accounted for the largest share in the total weed seed bank. Analysis of the change in weed seed bank over the three experimental years suggested that living mulches reduced weed seed bank in the ploughlayer by 14.1 to 57.1 %. In 2011, compared with the control treatment, the lowest number of weeds was established when growing white mustard (8.0 %) and Persian clover (30.4 %) living mulches. Although weed suppressive capacity of Italian ryegrass was high, contrary to expectations, it gave only a small reduction in weed seed bank and the weeds were significantly, nearly twice as big as those in maize crop without living mulch.

The abundance of weeds and living mulch
At early development stages of maize, more intensive growth was exhibited by spring rape, barley and white mustard living mulches (Table 14). However, living mulches of Italian ryegrass, black medic, Persian and red clover up to the first cut (maize BBCH [15][16] were only at seedling stage and therefore competed weakly with weeds. An especially rapid growth rate was shown by white mustard and until the first cut its dry mass was the highest. However, spring rape, barley and white mustard intercrops were sensitive to mulching and their regrowth after cut was poor, and in the second half of the summer they completely rotted away, the cut mass rapidly decomposed, therefore at later development stages of maize weed number and mass increased. Irrespective of this, these living mulches served their major purposecompeted with weeds at the time when maize competitive ability was low. The Fabaceae living mulches grown in maize inter-rows developed slowly; however, in the second half of the summer their growth rate increased and after cutting continued until the end of maize growing season. Moreover, they produced the largest mass. Compared with other Fabaceae family plants, black medic exhibited a slower development rate. Its mass was lower than that of other Fabaceae plants and it suppressed weeds more poorly; however, better than spring rape, barley and white mustard living mulches that had rotted away by the end of the summer. Italian ryegrass living mulch also produced large mass and exhibited a good weed suppressive ability. Its vegetation also continued until maize harvesting. Notes: C -control treatment (without living mulch), SR -spring rape, WM -white mustard, SB -spring barley, IR -Italian ryegrass, BM -black medic, PC -Persian clover, RC -red clover; differences significant at: * -95 % probability level, ** -99 % probability level. Control -reference treatment when analyzing mass of living mulches -red clover living mulch. The correlation-regression analysis of the data from 2009 revealed statistically significant relationships between dry mass of living mulches and weed number and dry mass (Table 15).  In most cases, different tillage did not have significant impact on weed seed bank in the ploughlayer and weed abundance in the agricultural crops tested. The ploughlayer differentiated into the upper layer with a greater weed seed bank (60.1 % of the total weed seed bank) and bottom layer with a less abundant weed seed bank (39.9 %). In spring crops, weed mass in shallow-ploughed plots was by on average 28.6 %, in deep-cultivated plots by 41.5 %, in shallow-cultivated plots by 39.9 % and in not tilled crops by 16.1 % higher than that in conventionally ploughed plots, and in winter wheat crop by respectively 2.5; 2.3; 3.4 times higher, and in not tilled plots by 2.8 times lower.

2.
Non-regrowing living mulches (white mustard, spring barley and rape) competed with weeds at early development stages when maize competitive ability was poor. Living mulches whose vegetation was longer exhibited better weed suppressive ability and produced more biomass; however, they competed more for nutrients with maize. The correlation regression analysis of the experimental data indicated that at more advanced growth stages of maize, the number and mass of weeds mostly depended on the biomass of living mulches. Living mulches reduced weed seed bank in the ploughlayer by 14.1 to 57.1 %. The greatest change was established when growing Persian clover (57.1 %) and black medic (53.6 %).

3.
Most of the Poaceae and Brassicaceae living mulches competed more with weeds for space at the beginning of maize vegetation, while Fabaceae plants and Italian ryegrass -already after the first mulching of the inter-rows. In most cases, a strong correlation was determined between the surface area covered by weeds and living mulches.