Optimization of foramsulfuron doses for post-emergence weed control in maize ( Zea mays L.)

Four field experiments were carried out from 2011 to 2014 in order to evaluate the effects of foramsulfuron, applied at the recommended (60.8 g a.i./ha) and reduced doses (1/3 and 2/3), on the efficacy against several of the most important weeds in maize. For each “year-weed” combination, dose-response curves were applied to estimate the dose of foramsulfuron required to obtain 90% and 95% weed control (ED 90 and ED 95 ). Foramsulfuron phytotoxicity on maize and crop yield were assessed. Foramsulfuron at 1/3 of the recommended dose (20.3 g a.i./ha) provided 95% efficacy against redroot pigweed ( Amaranthus retroflexus L.), green foxtail ( Setaria viridis (L.) Beauv.), wild mustard ( Sinapis arvensis L.) and black nightshade ( Solanum nigrum L.). Velvetleaf ( Abutilon theophrasti Medik.) , common lambsquarters ( Chenopodium album L.) and barnyardgrass ( Echinochloa crus-galli (L.) Beauv.) were satisfactorily controlled (95% weed efficacy) with ED 95 ranged from 20 to 50 g/ha of foramsulfuron (about from 1/3 to 5/6 of the recommended dose) depending on growth stage. The recommended dose was effective against pale smartweed ( Polygonum lapathifolium L.) at 2-4 true leaves (12-14 BBCH scale), but this dose did not kill plants larger than 2-4 true leaves. The ranking among weed species based on their susceptibility to foramsulfuron was: redroot pigweed = green foxtail = wild mustard = black nightshade > velvetleaf = common lambsquarters = barnyardgrass > pale smartweed. Dose of foramsulfuron can be reduced below recommended dose depending on weed species and growth stage. Foramsulfuron showed a good crop selectivity and had no negative effect on maize yield.


Introduction
Maize (Zea mays L.) is one of the most widely planted crop in the world and its production is increasing globally. Weed control has a major effect on the success of maize growth, because the competition ability of maize is relative low at early crop growth stages (Ghanizadeh et al., 2014). Weed control in maize largely depends on chemical methods. High input of herbicides results in environmental pollution, risks of residues carry-over and the development of weed resistance (Pannacci et al., 2006;Zhang et al., 2013). An effective way to reduce the side effects of the herbicides, according to the Integrated Weed Management 2 weed species" combinations (Dogan & Boz, 2005;Kudsk 2008;Pannacci et al., 2010). Furthermore, the knowledge of MDRE is one of the main factors in the implementation of the Decision Support Systems for Integrated Weed Management (Rydahl, 2004;Sønderskov et al., 2015), with the aim to decrease the dependence on herbicides that has become a distinct objective within the EU with the directive 2009/128/EC. The determination of MDRE requires dose-response studies for each "herbicide-weed species" under various environmental conditions (Kudsk & Kristensen, 1992;Pannacci & Covarelli, 2009).
Among post-emergence herbicides in maize, foramsulfuron is a sulfonylurea that exerts its herbicidal activity by inhibiting acetolactate synthase also known as acetohydroxy acid synthase and provides control of grass, perennial and some broadleaved weeds with a good selectivity to the maize (Bunting et al., 2004a(Bunting et al., , 2005. Furthermore, foramsulfuron can be applied in mixture with other herbicides increasing the control of some important key weeds, without risks of carry-over problems also in rotational vegetable crops (Bunting et al., 2005;Soltani et al., 2005).
In the past, several data were obtained on the efficacy of foramsulfuron used alone and in mixture with other herbicides against some problematic weeds (Bunting et al., 2004b;Nurse et al., 2007;Mc-Cullough et al., 2012), but few information is available about MDRE in the Mediterranean area conditions (Kir & Doğan, 2009). The objective of this study was to determine dose-response curves for foramsulfuron against several of the most important weeds in maize, and extrapolate the MDRE. Indeed, dose-response curve model applied in field herbicide efficacy trials was an important tool to estimate biologically meaningful parameters and to express weed control ability in terms of biological equivalent doses. Finally, the knowledge of MDRE for each "foramsulfuron-weed species" will allow to get the final goal to reduce foramsulfuron rates without losses of effectiveness against weeds.

Experimental site and design
Four field experiments in maize were carried out in 2011, 2012, 2013 and 2014 (Meier, 2001)] of maize. The growth stage of broadleaved weeds ranged from 2-4 to 6-8 true leaves stage (from 12-14 to 16-18 BBCH) and grass weeds ranged from first tiller visible to 3-4 tiller visible stage (from 21 to 23-24 BBCH), depending on species and years (Table 2). Foramsulfuron was sprayed at three doses (20.3, 40.5 and 60.8 g a.i./ha, corresponding to 1/3, 2/3, and the entire of the maximum recommended herbicide dose, respectively) using a backpack plot sprayer fitted with four flat fan nozzles (Albuz APG 110 -Yellow) and calibrated to deliver 300 L/ha spray liquid at 200 kPa. One untreated plot and one weed-free plot per block were added as controls. The experimental design was a randomized block with four replicates in separate blocks, randomised treatments within blocks and plot size of 17.5 m 2 (2.5 m width). Each plot was established from five rows, three central rows for meas-  Optimization of foramsulfuron doses in maize maize grain yield (adjusted to 15.5% of moisture content) was determined by hand-harvesting the central part of each plot (9 m 2 ). Meteorological data (daily maximum and minimum temperature and rainfall) were collected from a nearby weather station. The average decade of daily values were calculated and compared with multi-annual average values (from 1921) (Fig. 1). The times and quantity (mm) of irrigations were reported (Fig. 1).
urements and two border rows on the perimeter of each plot to reduce potential border effects.
In each trial, the efficacy of foramsulfuron towards each weed species was visually assessed on ground cover, 4 weeks after treatments (WAT), on a scale from 0 (no weed control in comparison with the untreated control) to 100 (total weed control). Phytotoxicity to the crop was evaluated visually, 2 WAT, on a 0-10 scale (0: no visible injury; 10: plant death). At harvest, the Table 2. Weed species occurring in the untreated control of the four herbicide field trials in maize crop: weed growth stages was referred at the treatment time; weed ground cover was evaluated 4 weeks after treatments (WAT).   I  I  I  I  II  II  II  II  II  III  III  III  III  III   mm I  I  I  I  I  II  II  II  II  II  III  III  III  III  III   30 I  I  I  I  I  II  II  II  II  II  III  III  III  III  III   30 I  I  I  I  I  II  II  II  II  II  III  III  III  III  III   30  Dose response curves for foramsulfuron against eight weed species showed always a good fit to experimental data (Fig. 2). ED 90 and ED 95 values for foramsulfuron and b values are reported in Table 3.

Weed species Common name
Velvetleaf was present in three field experiments and ED values for foramsulfuron were very similar in the three years (see Fig. 2 and Table 3). In more detail, this weed could be satisfactorily controlled (with an efficacy of 90% or 95%) using doses of foramsulfuron from 22.8 to 29.4 g a.i/ha; these doses correspond to 38-48% of the maximum labelled dose (60.75 g a.i./ha).
Redroot pigweed was present in all the field experiments all years and, also in this case, ED values appeared to be comparable across years ( Fig. 2 and Table 3). In particular, foramsulfuron showed a high efficacy against redroot pigweed, with ED 95 values ranging from 11.7 to 19.7 g a.i/ha depending of years.
Common lambsquarters showed EDs comparable in 2011 and 2012 with values ranging from 20.7 to 24.6 g a.i/ha, while in 2013 ED values were two times higher such as 40.1 and 52.6 g a.i/ha for ED 90 and ED 95 respectively (Table 3).
Barnyardgrass and pale smartweed were collected in all field experiments and their EDs showed a similar trend as for common lambsquarters, with lower values in 2011 and 2012 than in 2013 and 2014. In particular, in 2011 and 2012, barnyardgrass was satisfactorily controlled (with an efficacy of 90% or 95%) using doses of foramsulfuron ranging from 15.4. to 20.7 g a.i/ha; while in 2013 and 2014 ED values increased from 27.0 to 48.1 g a.i/ha ( Fig. 2 and Table 3). Similarly, pale smartweed was satisfactorily controlled with doses of foramsulfuron from 47% to 71% of the maximum labelled dose in 2011 and 2012; while in 2013 and 2014 this weed was not satisfactorily controlled, regardless of application dose, indicating a high tolerance to foramsulfuron (Table 3).
Wild mustard, green foxtail and black nightshade were only represented in one or two experiments. All these weed species showed low ED levels and thus high susceptibility to foramsulfuron (Table 3).

Phytotoxicity to the crop and crop yield
Phytotoxicity data showed that foramsulfuron could be considered safe to the crop with maize injury that appeared to be significantly correlated to

Statistical analysis
For each "year-weed" combination, data of herbicide efficacy was subjected to a non-linear regression analysis using the following logistic dose-response model : where Y is the percentage efficacy of the herbicide, x is the dose of herbicide, b is the slope of the curve around the inflection point, ED50 is the dose required to give 50% weed control.
The ED 50 parameter can be replaced by any ED level, so the selected model was used to estimate the dose of foramsulfuron required to obtain 90% and 95% weed control against each species, defined as MDRE (ED 90 and ED 95 values) (Copping et al., 1990;Seefeldt et al., 1995;Pannacci & Covarelli, 2009). When the upper asymptote did not reach 100%, it was included as a parameter in the model fitting .
The logistic dose-response model was directly fitted to the experimental data, by using the EXCEL ® Add-in macro BIOASSAY97 (Onofri & Pannacci, 2014). The goodness-of-fit was assessed by graphical analyses of residuals and F-test for lack-of-fit (Ritz & Streibig, 2005).
Data of phytotoxicity of maize were correlated to the three foramsulfuron doses in order to asses Pearson's r correlation coefficient (Kozak et al., 2012). Pearson's r correlation coefficient was performed by using EXCEL ® function.
Data of phytotoxicity and crop yield were subjected to analyses of variance and means were separated using Fisher's protected LSD at p = 0.05 level. The analyses of variance were performed with the EXCEL ® Add-in macro DSAASTAT (Onofri & Pannacci, 2014). Analysis of phytotoxicity data did not include untreated and weed-free plots, but these plots were included in grain yield data analysis.

ED levels for weed control
The four experiments were characterised by a different weed flora composition each year (Table 2). However, some weed species were ubiquitous and common in the four years, such as redroot pigweed, barnyardgrass and pale smartweed. Total weed ground cover was higher in 2013 and 2014 than in 2011 and 2012 (Table 2). Other sporadic weeds, not considered 5 Optimization of foramsulfuron doses in maize  (Table 4). Indeed, phytotoxicity symptoms, i.e. leaves slightly crinkled and distorted, were always transitory and dissipated within 3-4 WAT.
In general, crop yield in weed-free plots were at the same level in 2011, 2012 and 2013, while in 2014 was lower than in the previous years (Table 4).
The crop yield differences among untreated control and the increased foramsulfuron doses were lower in 2011 and 2012 than in 2013 and 2014 (Table 4), although the untreated controls showed a significant lower crop yield than the other treatments in all years, except in 2011. In particular, the highest losses in the untreated plots were observed in 2013 and 2014 with a yield reduction, compared to the weed-free control of 91% and 81% respectively; while in 2011 and 2012 the yield losses were of 38% and 33%, respectively (Table 4).  (Table 2). Indeed, in 2011 and 2012 the early sowing of maize at the end of April reduced the total weed flora, and due to low temperatures, barnyardgrass plants grew more slowly than in 2013 and 2014. It is well known that annual weed species are generally more susceptible to herbicides at early growth stages (Kudsk, 2008). Similarly, growth stage at application could explain the lower efficacy of However, reduced doses of foramsulfuron (20.3 and 40.5 g a.i./ha) gave crop yield levels not significantly different to that of weed-free plots in all the experiments, except in 2013 (Table 4).

Discussion
The four field experiments showed that weed susceptibility against foramsulfuron can be quantified by dose-response curves, and that weeds can be classified with respect to their susceptibility via ED 90 and ED 95 values. The ranking among weed species based on their sensitivity to foramsulfuron was: redroot pigweed = green foxtail = wild mustard = black nightshade > velvetleaf = common lambsquarters = barnyardgrass > pale smartweed. Baghestani et al. (2007) evaluated the efficacy of reduced rates of foramsulfuron in Iran and Table 3. Regression parameters (i.e. b, ED 90 and ED 95 ) for the relationship between the dose of foramsulfuron and its efficacy against weed species in , 2013 and ED 95 were assumed as "minimum doses requirement for satisfactory efficacy, MDRE". Standard errors are in parentheses.
Year and weed species b ED 90 (g/ha ) ED 95 (g/ha)  (Tracchi et al., 2002), that may frustrate the overall effectiveness of weed control (Pannacci & Covarelli, 2009). In such cases a good strategy is to use two or more active ingredients in mixture (i.e. sulcotrione or dicamba for foramsulfuron) in order to enlarge the spectrum of controlled weeds, and also avoiding the increase of herbicide resistant weeds (Green & Streibig, 1993;Pannacci et al., 2007;Beckie & Reboud, 2009). Foramsulfuron may be considered fairly safe to the crop, thanks to phytotoxicity symptoms always low and transitory as confirmed also by Tracchi et al. (2002), Nurse et al. (2007 and Zaremohazabieh & Ghadiri (2011) in other studies.
In general, lower crop yield in weed-free plots in 2014 than previous years can probably be due to the very high rainfall (113 mm in three consecutive days in the last decade of July) and low temperature (the daily average temperature decreased during the last decade of July until to 19 °C in these three days of rain) at flowering time (see end of July, Fig. 1d), that have affected pollination and growth of maize temporarily, reducing the potential yield of crop. The crop yield differences among untreated control and the increased foramsulfuron doses lower in 2011 and 2012 than in 2013 and 2014 were due to the high total weed control obtained also at low doses of foramsulfuron; indeed, the lower total weed pressure and younger growth stages at the treatment time increased their susceptibility to foramsulfuron (see Fig. 2). Otherwise, in 2013 and 2014, the low efficacy of foramsulfuron against barnyardgrass and pale smartweed (Fig. 2), together with their high ground covers, reduced crop yields values with no significant differences among the three doses (Table 4) (Table 2). Furthermore, in 2013 and 2014 the weather conditions, like water stress and high temperatures at the treatment time (21 June 2013 and 10 June 2014, see Figs. 1c and 1d, respectively) were more unfavourable to foliar-herbicides applications than in 2011 and 2012, especially against difficult-to-wet weed species, like barnyardgrass and common lambsquarters (Kudsk & Mathiassen 2004;Pannacci et al., 2010). This may be due to reduced herbicide availability caused by rapid drying of droplets to solid deposits in warm conditions (Ramsey et al., 2005). Furthermore, plants grown under high temperature, low humidity and soil water deficit tend to have leaves with thicker cuticles, more epicuticular wax and pubescence than unstressed plants, and this might retard interception, retention and penetration of the herbicide (Lundkvist, 1997). Indeed, common lambsquarters is characterised by very waxy leaves decreasing spray retention and uptake under hot and dry conditions (Steckel et al., 1997).
This study showed that foramsulfuron, also at reduced doses, is an effective herbicide to control several weed species (i.e. redroot pigweed, black nightshade, common lambsquarters and barnyardgrass) characterized by high competitiveness to the crop (Zanin et al., 1994;Pannacci & Tei, 2014), high seedbank densities (Graziani et al., 2012) and low susceptibility to several pre-emergence herbicides (i.e. velvetleaf). The use of herbicides at reduced doses seems to be possible; however, a limit could be the presence of non-susceptible weeds, i.e. Fallopia convolvulus (L.) Holub, Polygonum aviculare L., Chenopodium polyspermum L. and Stellaria media (L.) Vill. for foram- by Pannacci & Tei (2014) and by Pannacci & Onofri (2016), in the same area with similar weed infestations.
The results presented in this paper (i.e. ED 90 and ED 95 values) have shown that the post-emergence application dose of foramsulfuron can be reduced below recommended labelled rate depending on weed species and growth stage. In particular, foramsulfuron at 1/4 of the maximum labelled dose (≈ 20.3 g a.i./ha) provided 95% efficacy on redroot pigweed, green foxtail, wild mustard and black nightshade. The same efficacy was obtained against velvetleaf, common lambsquarters and barnyardgrass at 25-35 g/ha of foramsulfuron, corresponding to half of the maximum labelled dose. These results showed that foramsulfuron is an effective post-emergence herbicide to control the above mentioned weed species that are key weeds in maize and characterized by high competitiveness towards this crop. The MDRE's estimated in this study for foramsulfuron to optimize chemical weed control in maize, especially ED 95 values, can be implemented in Decision Support Systems (Kudsk, 2008). Finally, foramsulfuron showed a good selectivity to the crop, without negative effects to crop yield.