ABSTRACT

Herbicide tolerant oilseed rape is amongst the first of the GM (genetically modified) crops to be nearing commercialisation in Europe. The FACTT (Familiarisation and Acceptance of Crops incorporating Transgenic Technology) project is partly funded by the EU and was commissioned to consider the introduction and appropriate management of GM oilseed rape. This project uses hybrid oilseed rape from PGS which is tolerant to the herbicide glufosinate ammonium as the example GM type. One of the major aims of the project was to investigate the effectiveness of herbicide tolerance and the impact on herbicide management in oilseed rape production. Results of trials with winter rape are reported here for the 4 regions of the FACTT study; the UK, Germany, France/Belgium and Denmark/Sweden for 2 seasons of the project. Application of glufosinate ammonium was compared with 2 conventional herbicide treatments and a control without herbicide. A significant response in yield to herbicide application was shown in 6 of the 11 trials in 1996/97 and in 8 of the 12 trials in 1997/98. At these sites in both seasons, the untreated plots had a lower yield than herbicide treatments. Glufosinate ammonium phytotoxicity only occurred at very low levels when seed lots were slightly impure. Although the trials have shown that herbicide application does not always give a yield response in winter oilseed rape, herbicide tolerance may have particular benefits in certain circumstances, particularly for controlling problem weeds such as Brassica species. It may also be useful where conditions make use of conventional herbicides difficult, or to establish a flexible management system based on actual weed occurrence.

INTRODUCTION

Several GM crops have undergone trialling programmes in Europe and are nearing commercialisation. With regard to oilseed rape, an application for marketing consent has been made for varieties tolerant to glufosinate ammonium (Liberty) (Anon, 1998), and glyphosate (Roundup) tolerant types are also close to the market. FACTT is a 4 year project which was initiated in 1996 to consider the introduction and appropriate management of GM oilseed rape hybrids and is partly funded by the European Union. The project includes 21 partners in 6 member states and encompasses the main oilseed rape growing regions in the EU. For the purposes of FACTT these were divided into 4 regions; Germany, the UK, France/Belgium and Denmark/Sweden. Experimental programmes are undertaken for both winter and spring sown trials.

The ‘example’ GM crop used in the trials was GM hybrid oilseed rape, created with a hybrid system based on genetic transformations developed by Plant Genetic Systems, Belgium. Additionally it is tolerant to the herbicide glufosinate ammomium (Liberty). The specific objectives of FACTT are to evaluate herbicide tolerant GM oilseed rape under conventional agronomy, to evaluate its responses under changes in agronomy and also to establish demonstration sites and disseminate information regarding GM crops. Utilisation of a different herbicide regime afforded by the herbicide tolerance aspect of the particular GM type involved in the project provides a major new option for crop management. The aim of the trials reported in this paper was to investigate the effectiveness of the herbicide tolerance system compared to conventional alternatives, for winter sown varieties. The FACTT project is still ongoing hence this paper gives a preliminary view of the herbicide trial findings on the basis of 2 years data.

METHODS

For the herbicide experimental programme, trials were undertaken in all 4 regions of the FACTT study. Trials were conducted at 11 sites in 1996/97 and at 12 sites in 1997/98. The herbicide trials consisted of 4 treatments which were repeated in the 2 seasons. The application of Liberty (600 g ai/ha) in the autumn at 2 - 4 leaves of largest weeds was compared with a control treatment where no herbicide was applied and 2 different conventional herbicide treatments selected by partners as appropriate for the individual sites. These are detailed in Table 1. Two glufosinate tolerant varieties were used in each season of the trials; PGSW1 for both 1996/97 and 1997/98, and PGSW2 in 1996/97 which was replaced by PGSW3 in 1997/98. Analysis of Variance was carried out on yield data on an individual site basis. Further statistical analysis was undertaken for the 2 individual varieties for each year with site data grouped according to high, medium and low yields. In 1996/97, yields were regarded as high at > 4500 kg/ha, medium at 3500 - 4500 kg/ha and low at <3500 kg/ha. Yield categories for 1997/98 were >4000 kg/ha, 3000 - 4000 kg/ha and <3000 kg/ha for high, medium and low yielding sites respectively.

Table 1 Conventional herbicide treatments used by partners, and yield category achieved for both varieties.

NB H = high yield, M = medium yield and L = low yield, refer to text for yield levels.

RESULTS

In the 1996/97 season, 6 out of 11 sites analysed on an individual basis showed that the application of herbicide had a significant effect of yield. When sites were split into yield groupings for the 2 varieties (Table 2), it can be seen that a significant effect of herbicide treatment was evident for PGS W1 for the medium and low yielding site groups. For the medium yielding group, the untreated control was associated with a significantly lower yield than all the other treatments. The Liberty treatment tended to produce a higher yield than the first conventional treatment for this variety. For PGS W1 at the low yielding sites, the first conventional herbicide produced a significantly lower yield than the Liberty treatment. The yield for the untreated control tended to be higher than the first conventional treatment, but not significantly so. For the high yield group for PGSW1, there was a tendency for the untreated control to be lower yielding than the conventional herbicide treatments and for the Liberty treatment to be associated with a higher yield, although differences were not significant.. Yields for variety PGS W2 tended to be lower for the untreated control at both high and medium yield group sites, but not significantly so and there was little difference in yield between herbicide treatments. For PGS W2 at the low yielding sites, Liberty tended to produce a higher yield than the conventional herbicides and the untreated control, although differences were not significant.

Table 2 Effect of herbicide treatment on yield for the 1996/97 season.

In the 1997/98 season, for 8 out of the 12 trials, the application of herbicide was associated with a significant effect on yield. When viewed in yield groups, herbicide did not have a significant effect on yield, however certain trends could be seen (Table 3). For PGS W1 at both high and medium yield levels, the untreated control tended to be associated with lower yields than all other treatments, which produced similar yields. For the low yield group for PGS W1, the second conventional treatment tended to produce the highest yield, with little difference for others. For PGS W3 at the high yielding sites, again a lower yield was seen for the untreated control whilst little difference in yield was noted between herbicide treatments. For variety PGS W3 at the medium yielding sites, higher yields tended to be associated with Liberty application whilst there was little difference in yield between the untreated control and conventional herbicides. For PGS W3 in the lowest yield group, the untreated control tended to have the lowest yield, with Liberty tending to produce the highest yield. A significant interaction between site and herbicide treatment was noted for this group, with the untreated control at one site giving a higher yield than all herbicide treatments.

Table 3 Effect of herbicide treatment on yield for the 1997/98 season.

In the 1996/97 season it was noted that a small number of plants treated with Liberty, showed a phytotoxic effect and died. This was due to slight seed impurity. No phytotoxic effects were noted in the second season of trialling.

CONCLUSIONS

The results indicate a variable yield response to herbicide application, but in the majority of the yield groupings considered, there was a trend for the application of herbicide to result in a positive yield response. In some cases there was an indication that Liberty may give yield benefits above the conventional herbicide treatments.

The financial margin over cost of herbicide will have a major bearing in a grower’s decision to select a herbicide tolerant oilseed rape variety. As can be seen from the conventional herbicide programmes used by partners in the present trials, there is frequently a need to control both grass weeds and broad-leaved weeds resulting in considerable spray costs. For instance a typical conventional herbicide programme of pre-emergence Butisan (metazachlor) followed by a post-emergence graminicide, such as Fusilade will cost approximately £78/ha, whereas the cost of a non-selective herbicide such as Liberty, which has the advantage of providing a single treatment option for weed control, will be approximately £25/ha. Therefore achieving a similar yield response for Liberty to conventional herbicides will, potentially, give financial benefits. It will also be necessary to include in the gross margin an extra cost for seed of herbicide tolerant varieties, and possibly for technology fees, but since these varieties are not currently commercially available in Europe the level of these extra costs is, as yet, unclear.

Although the trials have shown that a yield benefit from herbicide application does not always occur, herbicide benefits may have particular benefits in certain circumstances. Use of herbicide tolerant oilseed rape will facilitate control of Brassica weeds, such as charlock, runch and hedge mustard, or for herbicide resistant blackgrass. In addition, use of herbicide tolerant oilseed rape may be used to control volunteers of speciality oilseed rape types and help to maintain quality. Herbicide tolerance may also be useful for certain circumstances, where conditions make use of conventional herbicides difficult, such as conditions of low soil moisture or high organic matter where glufosinate is not affected. It may also be used to establish a flexible management system based on actual weed occurrence, unlike many conventional systems which are based on pre-emergence herbicides applied before the actual weed problem can be ascertained. This could potentially allow more targeted use of herbicides. Conversely, inclusion of a herbicide tolerant crop in the rotation will also involve some additional management as discussed by Orson and Oldfield (1999). It will be necessary to ensure control of herbicide tolerant volunteers by another herbicide and undertake careful record keeping. It may also be beneficial to maintain isolation distances from other oilseed rape crops to minimise potential cross pollination.

The variable response to herbicide application seen in the present trials has been noted previously for winter oilseed rape, for example by Walker et al. (1990). A range of factors may lead to this variation, including levels of weed competition and factors influencing the competitive ability of the oilseed rape crop, including sowing date and over-wintering conditions. Further reports from the FACTT project will discuss the relationship between these factors and the effectiveness of the herbicide tolerance system in oilseed rape.

ACKNOWLEDGEMENTS

Part of the funding for this work was provided by the European Union.

REFERENCES

1. Anon, 1998 Advisory Committee on Releases to the Environment, Annual Report No 4: 1996/97, Department of the Environment, Transport and the Regions, March 1998.

2. Orson J H and Oldfield J F 1999 Gene flow and the practical management of genetically modified crops in the UK. 1999 BCPC Symposium Proceedings No 72: Gene Flow and Agriculture: Relevance for Transgenic Crops, 247-252.

3. Walker K C, Whytock G P and Davies D H K (1990) Evaluation of yield response and financial benefits from weed control in oilseed rape in Scotland. Proceedings Crop Protection in Northern Britain 1990, 301-306.