Summary 1.Agricultural intensification can cause a huge increase in productivity. However, associated costs in terms of reduced, self-regulation and increased reliance on external inputs for the control of pests, diseases and weeds are seldom taken into account or acknowledged. A pro-active approach in which ecosystems services are documented and potential effects of changes in agricultural practices evaluated may lead to more informed decisions prior to implementation. 2. We investigated the effects of management of cereal production in a semi-arid region on weed seed mortality caused by predators. Seed losses have a greater impact on weed population size than any other life cycle process and should therefore be of significance for natural weed control. We hypothesized that the conversion from rain-fed to irrigated production should lead to reduced and the adoption of no-till techniques to increased seed predation. 3. Seed removal and seed predator populations were monitored in irrigated ( N = 3) and rain-fed cereal fields ( N = 6) and field margins. Of the dryland fields half was conventionally tilled and the other half no-till. Seed removal (g g − 1 2-days − 1 ) was followed from April 2007 until June 2008, using Petri-dishes and exclosure cages. Populations of harvester ants were estimated by direct nest counts; rodent populations by Sherman live traps. 4. Seed removal in dryland cereals, mainly by harvester ants Messor barbarus was high from mid April to mid October, and should cause a strong weed suppressive effect. Seed removal in irrigated cereals, mainly by granivorous rodents Mus spretus , was low. 5. Seed removal was higher in no-till than in conventional fields and corresponded to differences in harvester ant nest densities. 6. Synthesis and applications . Our results show that tillage and irrigation in a semi-arid cereal production system results in a reduction and total annihilation of granivorous harvester ants, respectively. The concurrent decline in weed seed mortality could lead to increased herbicide use and dependency. In particular, in areas where economic margins are small or the environmental costs of tillage and irrigation high, the increased costs of chemical weed control may exceed the benefits. Here, preserving biodiversity to enhance natural weed control is a viable alternative to agricultural intensification.

This paper considers the relationships between the dispersal of seeds and the distribution pattern of an annual weed. A comparative study of seed dispersal by combine harvesters, with and without a straw chopper attached, was established using Lolium rigidum, a common weed in Mediterranean cereal crops. Seed dispersal distance was quantified and the relationships between dispersal and fine-scale seedling distribution evaluated. Primary dispersal of L. rigidum seeds occurs in a very limited space around the parent plants, but the density of seed is low because most seeds do not fall from spikes spontaneously. In contrast, many seeds are spread by combine harvesters. In this study the maximum dispersal exceeded 18 m from established stands in cereal fields, although the modal distance was close to the origin. In addition, the action of the combine harvesters tended to accumulate L. rigidum seeds predominantly under the straw swath, with some lateral movement. This action could explain the fine-scale banded pattern of L. rigidum in cereal fields. Although the treatment of straw by the standard and straw chopper combines differed, the resultant seed distribution showed few differences.

SummaryWeedy plants pose a major threat to food security, biodiversity, ecosystem services and consequently to human health and wellbeing. However, many currently used weed management approaches are increasingly unsustainable. To address this knowledge and practice gap, in June 2014, 35 weed and invasion ecologists, weed scientists, evolutionary biologists and social scientists convened a workshop to explore current and future perspectives and approaches in weed ecology and management. A horizon scanning exercise ranked a list of 124 pre‐submitted questions to identify a priority list of 30 questions. These questions are discussed under seven themed headings that represent areas for renewed and emerging focus for the disciplines of weed research and practice. The themed areas considered the need for transdisciplinarity, increased adoption of integrated weed management and agroecological approaches, better understanding of weed evolution, climate change, weed invasiveness and finally, disciplinary challenges for weed science. Almost all the challenges identified rested on the need for continued efforts to diversify and integrate agroecological, socio‐economic and technological approaches in weed management. These challenges are not newly conceived, though their continued prominence as research priorities highlights an ongoing intransigence that must be addressed through a more system‐oriented and transdisciplinary research agenda that seeks an embedded integration of public and private research approaches. This horizon scanning exercise thus set out the building blocks needed for future weed management research and practice; however, the challenge ahead is to identify effective ways in which sufficient research and implementation efforts can be directed towards these needs.

a b s t r a c tMany weed species are becoming rare due to intense agricultural management, which leads to a decrease of biodiversity in agroecosystems. Cultivating some of these species for their oilseed content may help preserve them while profiting agronomically. Thlaspi arvense is one of these species with potential as an industrial crop. The aim of this work was to develop a model to describe the emergence of this species, and that can help to make decisions for its management, whether for conservation or production purposes. The emergence of two accessions of T. arvense, one from Spain and the other from USA, sown in Spain (Almenar) and USA (Morris), over two seasons (2011-12 and 2012-13) and in Riga (Latvia) over one season (2012), was followed to compare patterns and extent, as well as to develop emergence prediction models based on hydrothermal time (HTT) and photohydrothermal time (PhHTT). For the USA accession, the percentage of seeds that emerged was significantly higher than that of the Spanish accession. Both accessions presented two emergence peaks (autumn-winter and spring) in both localities, but while these peaks could be considered as two different flushes in Almenar -for both accessions -, they appeared to be a single flush disrupted by low winter temperatures in Morris. An HTT-based model was applicable for both accessions with less precision than the PhHTT-based one, which was more accurate in most cases but failed in certain circumstances for the USA accession. The differences in emergence percentage among accessions suggest that some accessions might be more amenable to being used as a crop. The two models developed in this work predicted the emergence of both accessions of T. arvense quite accurately. The inclusion of photoperiod in the hydrothermal time equation, creating a new unit that we have called photohydrothermal time, offers a possibility to obtain more accurate models.

Summary A prerequisite for a precise estimate of the abundance of weed seeds in sou is prior knowledge of sampling variability, Based on an analysis of EWRS working group trials, it is shown that the sampling variance (s2) of seed counts increases with the sampling mean (S2) and can be predicted with the logarithmic form of Taylor's power law: log10(sz)=a+b log10(x̄). This relationship is constant over time for means greater than 0.1 seeds per core within each of the five different sites studied (temporal variability) but differs slightly among sites (geographical variability). An attempt is made to use a general s2 : x̄ relationship to predict the number of samples as a function of precision and density.