A simple mathematical model of the life‐cycle of Avena fatua L. is described and used to consider the financial consequences to the farmer of controlling A. fatua infestations in winter wheat. The likely economic benefits of applying herbicides in different cultivation regimes are investigated. The sensitivity of the model to variations in herbicide performance, straw burning, initial weed infestation and the value of the wheat crop is tested. The model predicts that the highest long‐term benefits will be obtained when a herbicide is applied every time wild oats exceed a density of between 2 and 3 seedlings m−2. Whether or not this results in a significant financial saving over spraying every year depends on the initial level of infestation and on the cultivation method. Whether straw is burnt or not is likely to be of little economic significance.

Summary Two experiments have investigated the persistence of 16 arable, annual broad‐leaved weed species over 6 years in a silty loam and a clay soil. Small plastic beads were included as an ‘inert’ comparison. Seeds were broadcast in October on to plots at the start of the experiment, and these were either tine cultivated or ploughed annually thereafter. Plots were sown with either spring or winter wheat. As far as possible, weed seeding was prevented each year. For some species, the seed decline appeared to be slower on the ploughed plots than on the tined plots and in the winter wheat compared with the spring wheat. Seed decline also tended to be slower on the clay soil at Rothamsted than on the silty loam at Long Ashton. Some species declined rapidly (e.g. Brassica napus, Chrysanthemum segetum, Galium aparine, Galeopsis tetrahit), with annual decline rates in excess of 58%, whereas others declined very little (e.g. Papaver rhoeas decline rate 9%). Most of the other species had decline rates between 20% and 40%. The results are discussed in relation to the development of population dynamics models to predict long‐term consequences of alternative weed management strategies.

SummaryHorizontal movement of seeds by different cultivators was investigated in three experiments. Seeds of barley (Hordeum vulgare L.), field bean (Vicia faba L.) and oilseed rape (Brassica napus L.) were positioned on the soil surface or buried 0.1 m deep prior to cultivation, Seed movement was assessed by counting germinated seedlings. The type of tine implement used significantly affected mean seed movement, with surface‐sown seeds being moved significantly further than buried seeds. Primary cultivation with flexi‐ or spring‐tine machines moved seeds further than straight‐tine or power harrow implements. However, there was no overall difference between plough and flexi‐tine mean seed movement. A single pass with a tine attached to a seed drili moved seeds less than the two passes with tine plus seed drill, but the first pass had the greatest effect. Small oilseed rape seeds moved significantly further than larger barley or field bean seeds. More than 84% of seeds moved ≤ 1m from the source: no seeds were observed>4.8m in a forward direction or >0.2m backwards from the source, In a fourth experiment, plots containing seeds of Sinapis arvensis L. were cultivated in alternate directions in successive years, The results broadly validated the findings of the previous experiments Quantifying the horizontal movement of weed seeds is important to the understanding of the spatial dynamics of weed patches and thus in the prediction of future weed distributions.

within the five fields was visibly patchy, but the size and morphology of patches varied both within and between fields. Thus, the potential reduction in herbicide usage as a result of patch spraying varied with patch morphology and infestation level The effect of lowering the resolution from six (2 x 1) m^ cells to three (4 x 1) mĉ ells across the 12-m boom width showed little change in potential reduction in herbicide usage. SummaryThe spatial distribution of Efymus repens L. was mapped in five cereal fields during the 1994 season. Weed maps were created using a semiautomated system mounted on a vehicle that travelled up the tramlines, centred on a 12-mwide spray boom. Two operators detected weeds visually and recorded their presence and density by means of a simple push-button system that recorded data from (2x1) m^ areas, or cells, across the boom section on a portable computer. The position of the vehicle along the tramline was monitored by integrating output from a wheel sensor. Eight tramlines of one field were mapped three times consecutively, to assess detection and navigational error. Pair wise comparisons of the three runs gave approximately 85% repeatability on presence/ absence data, 80% on zero/low-or high-density data and 85% on low/high-density data when weeds were definitely present. Simultaneous comparisons of all three runs gave 78%, 69% and 75% respectively. Repeated runs of tramlines up to 550 m long recorded mean differences of approximately 2 m with a maximum error of 14 m (2,5%). The spatial distribution of E. repens

In this review the available information on the sexual reproduction of Cirsium arvense L. (Scop.) is summarized and discussed ia an ecological context. Certain aspects of its sexual reproduction system have given C. arvense a reputation for low efficiency: the dioecious mating system and dependence on insects for poilination and non-adaptive features for wind dispersal. The seeds are moderately persistent in soil and, for a temperate weed, have a high temperature requirement for germination. It is concluded that the contribution that sexual reproduction makes to the survival and spread of this species has heen underestimated, partly owing to an inadequate number of studies. Seedling recruitment via sexual reproduction may be an important mechanism for initiating continued genetic diversity at intra-and inter-population levels. efeienoe Havdeti, 1934 Bakker. lWO Denchdd & Sdwltz, l%0 Amm &. Harfis, 1974 Bostock & Bentco, I9T9 van Leeuwen. 1987 f UDstiatfed, without competitkiii, rather moist day loam, aerated lo a depth of 30-40 cm, maJe C. arvense within 20 m. Sd, light iiStoiisity in July 35-45% of full day light, riidier moist da> laiuii,aeTale