Described as a chasm between scientists and practitioners by DeSimone (2013), the ‘yawning implementation gap’ and ‘The Great Divide’ in a Nature editorial (Anonymous. 2007); the value and relevance of a significant proportion of ecological research to practitioners on the ground often seem remote at best. There are a number of issues which contribute to the research–implementation gap, some of which Journal of Applied Ecology is increasingly making an effort to address (Hulme 2011). First, practitioners often cannot access relevant articles due to lack of funding for journal subscriptions. The idea of conferences on the ‘Think Global, Act Local’ theme sponsored by the Journal and the British Ecological Society (BES) is an example of a promising initiative (Memmott et al. 2010). Secondly, practitioners are busy people, coping with a wide range of issues from budgets to health and safety, as well as habitat and people management. They often do not have time to read multiple articles, let alone spending the time and money required to find them. Thirdly, some practitioners, who are not necessarily all ecologically trained, could have difficulty understanding the research results, particularly where complex statistical analyses and modelling techniques have been used. Therefore, a synopsis of all the relevant research on their habitat type, with advice on management options and their likely outcomes, would be far more useful. As Memmott et al. (2010) point out, there is inadequate research testing many management recommendations in the field and this results in the offered management solutions for particular circumstances falling far short of the ideal. However, there is also another fundamental difference in approach used by practitioners and researchers, which contributes to the research–implementation gap and is perhaps more difficult to address. To paraphrase Hill and Arnold (2012), practitioners have to make judgements and provide advice on the best available evidence, usually combined with relevant experience and professional judgement. In general, the available evidence is not particularly good – it may not be related to the geographical area, altitude, climate, soils or the precise mix of species on the site, it may result from small-scale experiments, whereas the situation demands large-scale approaches. The result leads to considerable uncertainty. In contrast, ecological scientists need compelling evidence from their studies to reach their conclusions, which in turn lead to the publication and sharing of that knowledge. It is worth exploring these differences more closely and I illustrate this issue using peatland restoration as an example, to find ways of closing the gap and gaining a more productive relationship between practitioners and researchers. Researchers partition the issues to focus on the feature of interest, undertake experiments that allow them to identify the factor/s affecting change and account for other contributory features or drivers. The practitioner, on the other hand, has to interpret the whole environment, juggle the significance of different factors in different locations and make proposals for change based on this assessment. The well-read practitioner depends on the researchers' published findings to assist in their interpretation and prognosis and to base their decisions on restoration or management, but the research will never cover the totality of the environment in which they are working and personal experience and judgement are always necessary. The process is more akin to the art of applying the science. This issue can be readily illustrated in the upland peatland environment in the UK and Ireland. Most of these upland peatland sites will have experienced a variety of effects related to grazing, drainage, gullying (water-eroded channels) and erosion of exposed peat, wildfire, air pollution, peat extraction and managed burning. There may also be other intrusions such as tracks or windfarms that affect the hydrology and integrity of the peat and its management. Some of these factors may have influenced the site over many decades, if not centuries (such as sulphur dioxide pollution close to the industrial centres around the South Pennines in England and in South Wales, both in the UK); each interacting with the other, leading to the current patterns and functions. Each of these factors may have been researched in various ways and in a mixture of places, perhaps far from the practitioner's site – geographically or altitudinally – but their interactions are rarely considered together in a scientific manner. The practitioner has to be aware of all the relevant research to best understand their site and the changes that they might be engineering. This might include issues of purple moor-grass Molinia caerulea invasion, the effect of air pollution on Sphagnum, the effect of drains on peat hydrology and associated vegetation and craneflies (especially, where these are an important food source of red grouse Lagopus lagopus scotica chicks, the production of which may be the economic rationale for the site's management). Some papers also lack descriptions of the sites and habitats or the integrity and condition of the peat, particularly where hydrological characteristics or features other than the site's ecology are being explored, thus making it difficult to extrapolate the findings to other sites. This situation could be eased if the electronic appendices now permitted with published papers were to incorporate such information, and authors should be encouraged to do this. The practitioner needs to understand the interactions between factors and identify the key drivers and how they came to their current interplay. Relevant research might provide indicators, but careful observation over multiple years and sites in order to interpret the patterns observed also come into play. Some of the more damaged blanket peat in the Southern Pennines in England illustrates this problem where eroding and gullying peat, probably caused by multiple wildfires, are being restored in large-scale projects. But knowledge is lacking on the process of erosion after a damaging wildfire which raises a number of questions? How quickly do rills (the initial stage of gullies resulting from erosion on an exposed slope) and then gullies form on different slopes and at different altitudes? Do they erode into vegetated ground or only into areas where vegetation has been lost? Where, When and How do peat pipes form (cavities of different sizes within the peat through which water flows) ? and Can the whole degenerative system be modelled to help inform restoration management and investment? There are many observations and inferences that can be drawn, but little research into these questions. Additionally, the amount of experimentation that is possible without destroying the resource itself is limited. Multiple wildfires cannot really be simulated at the catchment scale, and much of the land in the uplands lies in protected areas where the statutory nature conservation agencies, who have some control over site activities, would be unlikely to agree to anything that damages the peat. Another difference between the researcher and the practitioner's approach is that while a researcher has to generalize findings, which risks the loss of specific context, the practitioner needs to apply the research results to a particular place which could be quite different from where the research was undertaken. For example, how applicable is research carried out in Scotland to sites on Dartmoor (south-west England) or in the Pennines (northern England)? Milne et al. (2002) showed how significant the differences are between vegetation biomass potential across the UK in upland areas. An understanding of this research and the ability to apply it to a particular site will be needed to appreciate the rates of change resulting from grazing reductions, for example, on dwarf shrub heath vigour and health. Practitioners have observed that there can be a significant delay of a year or more, while severely suppressed dwarf shrubs (from heavy grazing usually) seem to gain sufficient energy to grow more vigorously again, but these observations and the mechanisms involved have not been researched. Another question a practitioner might ask is whether a new technique applied in particular circumstances will work in others. An example might be whether some work showing geojute (a mesh made of jute) successfully holding bare peat in place on fairly level ground (Anderson, Tallis & Yalden 1997) would be as effective on steep, bare gully sides. Although Anderson, Tallis & Yalden (1997) demonstrated that a 50% cover of geojute was as effective as a 100% at holding fairly level bare peat surface down long enough to revegetate it, further work is required to extrapolate the results of this study in small plots to sites involving restoration of many hectares of bare peat, much of which is on very steep gully sides. After much trial and error, the ongoing practical restoration work now reserves geojute for the steeper gully sides and other slopes rather than the gently sloping or flatter surfaces, for both practical and cost reasons. researchers' experimental scales tend to be small with multiple small plots and possibly a maximum of a few hectares for grazing plots (e.g. Fraser et al. 2009). In many instances, grazing stock in small numbers is used for short periods (weeks or even days) to reach a pre-set level of vegetation offtake or stocking rate. Although useful indicators, the results are difficult to apply directly to a site where there are mixtures of vegetation types which differ in their appeal to the stock in question (and this can vary with season as well), with additional complicating factors of topography, the tendency for some stock to move up a hill during the day to escape flies or to shelter from prevailing winds and rain on leeward slopes. Furthermore, the differences between breeds and stock types have been little explored and, judging from anecdotal descriptions, vary markedly across and between sites (Tolhurst & Oates 2001). Practitioners manage or restore whole sites or estates or multiple areas. These usually include not only the habitat that has been researched (e.g. effects of different grazing levels or animal dietary selectivity of species in dwarf shrub heath or blanket bog) but a number of other habitats such as flushes, springs, acid grasslands, scrub and woodland. Managing the grazing in complex sites requires knowledge of the research findings, but also an ability to judge how animals will behave with more choice of forage in a larger area and in relation to local environmental features, as well as their innate range behaviour (which will vary with species and breed). The effect of, and competition with, wild herbivores such as hares, rabbits and deer has also to be taken into account. Managing such sites comes mostly from experience and observation. Not everything can be, or is, researched. The multiple research projects examining stocking rates and vegetation effects, often with differing results, bear witness to the wide-scale variation and provide useful guidance, but they need to be interpreted by the practitioner on the ground. This could appear to be a daunting task unless the practitioner understands that the variability of the results demonstrates the variability of habitat and circumstances where the observations have been made. The diversity of recommended grazing rates could otherwise be confusing. Another limitation of research approach and methods in small plots is that the application of treatments tends to be at an equivalent scale (the geojute and spreading of heather brash or seed reported in Anderson, Tallis & Yalden (1997), or the larch brashings and bitumen applications described by Tallis & Yalden (1983) were all manually implemented). When applying the findings at a large scale, new techniques need to be trialled – there is no real scope for experimenting to find out if helicopter transport and spreading of lime, fertilizer, grass seed, dam materials, etc. will work because of the high costs involved. Someone has to try it and then pass on the knowledge to others – trial and error or refinement in successive years of works is the normal approach rather than any detailed and carefully monitored experimentation. A last major difference between researchers and practitioners lies in time-scales. Far too often in environments like the uplands, where change can be slow, the researcher is limited to three, perhaps four years to monitor the effects of his experiments. What the weather is like when the research takes place may be critical. A practitioner friend of mine has trialled cutting of dense Molinia caerulea to reduce its dominance and make space for other species to colonize or be added on some Peak District hills (England). He found that if cut in a dry year, the Molinia was likely to be reduced or die, but if cut in a wet year, it rebounded strongly. A PhD (which would normally be limited to three field seasons) undertaken over predominantly wet or dry seasons (as we have experienced in recent years in the UK) could have given totally different results – and extrapolating from this could have been very misleading without the practitioner's wider knowledge gained over many years. Three years monitoring of experiments is also too short in many habitats. The practice of combining studies into a series of PhDs to extend the time period is a useful approach to overcome this time-scale problem. The issue of short time-scales can also be circumvented in particular situations by comparing results of similar work that has been carried out over a longer period, such as the review of the effectiveness of grip blocking (grips are drainage ditches cut into peat with UK government support in the 1960s and 1970s with the aim of improving the blanket bog for sheep grazing and grouse management, and the process is now being reversed as the integrity of the peat is seriously affected) after a period of up to seven years by Armstrong et al. (2010). However, such reviews are based on a number of assumptions about the progress of individual sites up to the point of measurements being taken. This is clear from long-term monitoring studies such as for the Sustainable Catchment Management Programme (SCaMP) implemented by United Utilities (a water company who manages much upland catchment supplying their reservoirs) in the Peak District and Bowland (England). Here, a small, but statistically significant, reduction in dissolved organic carbon was found two years after peat grip blocking on two sites, but not until after seven years on another site, with considerable fluctuations in relation to weather in each year (http://corporate.unitedutilities.com/documents/SCaMP_Hydrological_Report_for_2011_(Aug_2012).pdf) – the results of which have not yet been submitted to any journal but which are available as part of the grey literature. This Practitioner's Perspective article has highlighted some of the gaps, but it should be acknowledged that there are also instances where the gap is being bridged effectively. The situation is much better than it was 30 or 40 years ago, thanks largely to the Internet and the increase in available information. It is easier to share and find information digitally – abstracts, for example, are available online for most papers in journals and some full papers are free to download. However, it would help if abstracts succinctly described all the salient features of the study, including the habitat, interventions used, results or conclusions and not just the objectives and methods. Nevertheless, frustration is still experienced by most practitioners without any institutional library access when they cannot access the full papers to identify whether the studies are relevant or not and cannot afford the fees for a large number of papers. Abstracts rarely tell the whole story and not everyone feels able to contact the author for a paper reprint. The Journal of Applied Ecology is taking a stronger lead to help bridge the gap identified, but more work needs to be done and both ‘sides’ need to contribute in order to develop other ways of bridging the gap. More interdisciplinary working would help – hydrologists, soil scientists and ecologists working together and reporting together on different experiments or trials. Each paper, or its online appendices, should include a description of the habitat and soils/peat on which they are working so that readers can extrapolate appropriately to other sites. Better access (which is already improving) to literature is vital – most practitioners do not belong to learned societies or have access to libraries (digital or university). Only time will tell how open access policy will work in practice, and there are many issues still to resolve. More value needs to be given to the grey literature. Much of the more recent material is available freely on the Internet, although some effort may need to be made to ensure that it is easily found and not buried in an obscure location. There may be a case for grey literature data repositories – something that the Chartered Institute for Ecology and Environmental Management might like to consider. Although it has not been peer-reviewed, some of the grey literature still has real value, and careful scrutiny by scientists will reveal whether the results conveyed are meaningful or useful. Some are well-designed experiments, but the authors fail to find time to submit the material for publication if it is not demanded as part of their job. Much of the grey literature can at least provide a wider context to, and multiple applications for, more formal research results. Older grey literature is often more difficult to find unless obtained at the time it was published. Higher regard for and use of practical experience and observation is also needed – practitioners and researchers need to work together and draw relevant conclusions based not just on results of experiments, but which are modified and qualified by collective practical experience and observations so that results are more realistic and widely applicable. Memmott et al. (2010) strongly advocate this approach as well as more interdisciplinary research. This would provide great opportunities for assessing and extending results of randomized controlled trials from the population studied to the wider context (Cartwright 2013). There is now a need to ensure research has impact (Milner-Gulland et al. 2012), which has the potential to have a significant effect on bridging some gaps between practitioners and researchers. A good example has been the recent seemingly academic work by Freeman and colleagues at Bangor University, Wales. This has unravelled the biochemistry of decomposition that occurs when drying peat becomes aerated, resulting in increased carbon dioxide release into the atmosphere and dissolved organic carbon in water draining from the catchment (e.g. Freeman, Ostle & Kang 2001). Practitioners' understanding of this has been critical in devising practical methods to reduce loss of carbon in streams (currently an issue for some water companies in the UK) through rewetting peat. This raises the water-table and reduces the aerated conditions in which the enzyme phenol oxidase can operate and regulates peatland hydrolases, which are responsible for peat decomposition. Publications in journals which are not read by, or are not accessible to, the target audience fail to reach the practitioner. Here, I mean target audiences that include not just other researchers but those that could be applying the findings on the ground. These are the people who will be able to show the impact of research. The number of times a paper has been cited by other researchers does not necessarily reflect on its practical application, or whether it has reached the practitioners who would benefit from its contents. Many projects also run workshops, training or presentation events, but these may miss most of the target audience and are a single event that does not support future researchers and practitioners. More needs to be carried out in this field. Some publications do assist bridging the gap – the Chartered Institute of Ecology and Environmental Management (CIEEM) produces a member's journal in which précis are given of numerous academic articles (including from this journal) and many members are consultants and practitioners; conservation evidence and many of the publications of Natural England and other country statutory agencies that are readily available on the internet also provide very useful literature-based reviews and advice on relevant subjects, and they are highly valued. Another avenue would be for more researchers to be involved in the fora set up around subject areas and to offer advice based on their research results through these channels – for example, nibblers (the Internet forum for the English Grazing Animals Project which hosts discussions related to conservation land management using stock grazing) and heath-net (for heathland managers) are good examples in the UK but there are many others. Many practitioners subscribe to these self-help Internet groups and would welcome expert input from researchers. My personal membership of several of these fora reveals the dearth of researchers involved who could make useful contributions to many of the questions as well as learning more about the Practitioners' concerns. Finally, there is an important role for consultants or advisors who regularly work with both researchers and practitioners. These are often the middle people who have better access to the academic literature (through membership of professional societies, for example); are perhaps well-placed to understand and interpret research; may have the experience to apply the appropriate findings to particular sites or area; and who work closely with practitioners (or are practitioners themselves) thus providing the link between academics and those on the ground. Many consultants or advisors belong to their professional institute, for example CIEEM, which has been working hard over the last few years to build capacity and increase levels of professionalism that will support the practitioner working to meet the researcher across the divide. This capacity building can be seen in the new competency framework to be used both for application of CIEEM to Chartered status and for applying for, and upgrading of, membership levels. The framework includes subthemes on scientific method, design and analysis, amongst a host of different topics allied to ecology and environmental management. The first tranche of Chartered Ecologists were awarded in October 2013, but amongst these, only three of the twenty-one applicants are academics or retired academics. All members are obliged to undertake an annual continuous professional development programme, and the Institute provides conferences and courses to assist in achieving this, some in collaboration with the BES and other professional bodies. These often provide opportunities for practitioners and researchers to share information, experience and skills. More researchers becoming members of the Institute would also help in this respect. Even if only some of the ideas outlined above are brought to fruition, but at a sufficient scale and breadth, then the gap between practitioner and researcher will surely close to the benefit of both parties. As a person with a foot in each camp, I would gladly welcome that. Penny Anderson established the widely respected ecological consultancy Penny Anderson Associates Ltd (PAA) in 1972, which now also employs soil and hydrology specialists to complement their extensive ecological capabilities. Penny has specialized in upland restoration and management as well as habitat creation. She has been involved in monitoring a number of long-running restoration and habitat creation projects and has prepared restoration plans for many thousands of hectares of blanket bog and dwarf shrub heath. PAA has been monitoring the SCaMP project since 2005. Penny regards herself as a practitioner and interpreter of scientific research to the wider practitioner community wherever possible, but likes to dabble in research where she can. She was President of IEEM from 2010 to 2012 just before it became a Chartered body.