Non-uniform data distribution for communication-efficient parallel clustering

Global communicationrequirements andloadimbalanceof someparalleldataminingalgorithms arethe major obstacles to exploitthe computational power of large-scale systems. This work investigates how non-uniform data distributions can be exploited to remove the global communication requirement and to reduce the communication costin parallel data mining algorithms and, in particular, in the k-means algorithm for cluster analysis. In the straightforward parallel formulation of the k-means algorithm, data and computation loads are uniformly distributed over the processing nodes. This approach has excellent load balancing characteristics that may suggest it could scale up to large and extreme-scale parallel computing systems. However, at each iteration step the algorithm requires a global reduction operationwhichhinders thescalabilityoftheapproach.Thisworkstudiesadifferentparallelformulation of the algorithm where the requirement of global communication is removed, while maintaining the same deterministic nature ofthe centralised algorithm. The proposed approach exploits a non-uniform data distribution which can be either found in real-world distributed applications or can be induced by means ofmulti-dimensional binary searchtrees. The approachcanalso be extended to accommodate an approximation error which allows a further reduction ofthe communication costs. The effectiveness of the exact and approximate methods has been tested in a parallel computing system with 64 processors and in simulations with 1024 processing element

Multi-level planning for semi-autonomous vehicles in traffic scenarios based on separation maximization

The planning of semi-autonomous vehicles in traffic scenarios is a relatively new problem that contributes towards the goal of making road travel by vehicles free of human drivers. An algorithm needs to ensure optimal real time planning of multiple vehicles (moving in either direction along a road), in the presence of a complex obstacle network. Unlike other approaches, here we assume that speed lanes are not present and that different lanes do not need to be maintained for inbound and outbound traffic. Our basic hypothesis is to carry forward the planning task to ensure that a sufficient distance is maintained by each vehicle from all other vehicles, obstacles and road boundaries. We present here a 4-layer planning algorithm that consists of road selection (for selecting the individual roads of traversal to reach the goal), pathway selection (a strategy to avoid and/or overtake obstacles, road diversions and other blockages), pathway distribution (to select the position of a vehicle at every instance of time in a pathway), and trajectory generation (for generating a curve, smooth enough, to allow for the maximum possible speed). Cooperation between vehicles is handled separately at the different levels, the aim being to maximize the separation between vehicles. Simulated results exhibit behaviours of smooth, efficient and safe driving of vehicles in multiple scenarios; along with typical vehicle behaviours including following and overtaking.

Electron trapping in shear Alfvén waves that power the aurora

Results from 1D Vlasov drift-kinetic plasma simulations reveal how and where auroral electrons are accelerated along Earth’s geomagnetic field. In the warm plasma sheet, electrons become trapped in shear Alfven waves, preventing immediate wave damping. As waves move to regions with larger vTe=vA, their parallel electric field decreases, and the trapped electrons escape their influence. The resulting electron distribution functions compare favorably with in situ observations, demonstrating for the first time a self-consistent link between Alfven waves and electrons that form aurora.

The global distribution of phytoplankton size spectrum and size classes from their light-absorption spectra derived from satellite data

The absorption spectra of phytoplankton in the visible domain hold implicit information on the phytoplankton community structure. Here we use this information to retrieve quantitative information on phytoplankton size structure by developing a novel method to compute the exponent of an assumed power-law for their particle-size spectrum. This quantity, in combination with total chlorophyll-a concentration, can be used to estimate the fractional concentration of chlorophyll in any arbitrarily-defined size class of phytoplankton. We further define and derive expressions for two distinct measures of cell size of mixed populations, namely, the average spherical diameter of a bio-optically equivalent homogeneous population of cells of equal size, and the average equivalent spherical diameter of a population of cells that follow a power-law particle-size distribution. The method relies on measurements of two quantities of a phytoplankton sample: the concentration of chlorophyll-a, which is an operational index of phytoplankton biomass, and the total absorption coefficient of phytoplankton in the red peak of visible spectrum at 676 nm. A sensitivity analysis confirms that the relative errors in the estimates of the exponent of particle size spectra are reasonably low. The exponents of phytoplankton size spectra, estimated for a large set of in situ data from a variety of oceanic environments (~ 2400 samples), are within a reasonable range; and the estimated fractions of chlorophyll in pico-, nano- and micro-phytoplankton are generally consistent with those obtained by an independent, indirect method based on diagnostic pigments determined using high-performance liquid chromatography. The estimates of cell size for in situ samples dominated by different phytoplankton types (diatoms, prymnesiophytes, Prochlorococcus, other cyanobacteria and green algae) yield nominal sizes consistent with the taxonomic classification. To estimate the same quantities from satellite-derived ocean-colour data, we combine our method with algorithms for obtaining inherent optical properties from remote sensing. The spatial distribution of the size-spectrum exponent and the chlorophyll fractions of pico-, nano- and micro-phytoplankton estimated from satellite remote sensing are in agreement with the current understanding of the biogeography of phytoplankton functional types in the global oceans. This study contributes to our understanding of the distribution and time evolution of phytoplankton size structure in the global oceans.

Are power calculations useful? A multicentre neuroimaging study

There are now many reports of imaging experiments with small cohorts of typical participants that precede large-scale, often multicentre studies of psychiatric and neurological disorders. Data from these calibration experiments are sufficient to make estimates of statistical power and predictions of sample size and minimum observable effect sizes. In this technical note, we suggest how previously reported voxel-based power calculations can support decision making in the design, execution and analysis of cross-sectional multicentre imaging studies. The choice of MRI acquisition sequence, distribution of recruitment across acquisition centres, and changes to the registration method applied during data analysis are considered as examples. The consequences of modification are explored in quantitative terms by assessing the impact on sample size for a fixed effect size and detectable effect size for a fixed sample size. The calibration experiment dataset used for illustration was a precursor to the now complete Medical Research Council Autism Imaging Multicentre Study (MRC-AIMS). Validation of the voxel-based power calculations is made by comparing the predicted values from the calibration experiment with those observed in MRC-AIMS. The effect of non-linear mappings during image registration to a standard stereotactic space on the prediction is explored with reference to the amount of local deformation. In summary, power calculations offer a validated, quantitative means of making informed choices on important factors that influence the outcome of studies that consume significant resources.

Dynamic distributed lanes: motion planning for multiple autonomous vehicles

Unorganized traffic is a generalized form of travel wherein vehicles do not adhere to any predefined lanes and can travel in-between lanes. Such travel is visible in a number of countries e.g. India, wherein it enables a higher traffic bandwidth, more overtaking and more efficient travel. These advantages are visible when the vehicles vary considerably in size and speed, in the absence of which the predefined lanes are near-optimal. Motion planning for multiple autonomous vehicles in unorganized traffic deals with deciding on the manner in which every vehicle travels, ensuring no collision either with each other or with static obstacles. In this paper the notion of predefined lanes is generalized to model unorganized travel for the purpose of planning vehicles travel. A uniform cost search is used for finding the optimal motion strategy of a vehicle, amidst the known travel plans of the other vehicles. The aim is to maximize the separation between the vehicles and static obstacles. The search is responsible for defining an optimal lane distribution among vehicles in the planning scenario. Clothoid curves are used for maintaining a lane or changing lanes. Experiments are performed by simulation over a set of challenging scenarios with a complex grid of obstacles. Additionally behaviours of overtaking, waiting for a vehicle to cross and following another vehicle are exhibited.

The power of perceptions: exploring the role of urban design in cycling behaviours and healthy ageing

Good urban design has the power to aid in the provision of inclusive journey environments, yet traditionally neglects the perspective of the cyclist. This paper starts from the premise that more can be done to understand and articulate cyclists’ experiences and perceptions of the urban environment in which they cycle, as part of a closer linking of urban design qualities with transport planning and infrastructure interventions. This approach is particularly applicable in relation to older cyclists, a group whose needs are often poorly understood and for whom perceptions can significantly influence mobile behaviours. Currently, knowledge regarding the relationship between the built environment and physical activity, including cycling, in older adults is limited. As European countries face up to the challenges associated with ageing populations, some metropolitan regions, such as Munich, Germany, are making inroads into widening cycling’s appeal across generations through a combination of urban design, policy and infrastructure initiatives. The paper provides a systematic understanding of the urban design qualities and built environment features that affect cycling participation and have the potential to contribute towards healthy ageing. Urban design features such as legibility, aesthetics, scale and open space have been shown to influence and affect other mobile behaviours (e.g. walking), but their role as a mediator in cycle behaviour remains under-explored. Many of these design ‘qualities’ are related to individual perceptions; capturing these can help build a picture of quality in the built environment that includes an individual’s relationship with their local neighbourhood and its influences on their mobility choices. Issues of accessibility, facilities, and safety in cycling remain crucial, and, when allied to these design ‘qualities‘, provides a more rounded reflection of everyday journeys and trips taken or desired. The paper sets out the role that urban design might play in mediating these critical mobility issues, and in particular, in better understanding the ‘quality of the journey’. It concludes by highlighting the need for designers, policy makers, planners and academics to consider the role that design can play in encouraging cycle participation, especially as part of a healthy ageing agenda.

Using reanalysis data to quantify extreme wind power generation statistics : a 33 year case study in Great Britain

With a rapidly increasing fraction of electricity generation being sourced from wind, extreme wind power generation events such as prolonged periods of low (or high) generation and ramps in generation, are a growing concern for the efficient and secure operation of national power systems. As extreme events occur infrequently, long and reliable meteorological records are required to accurately estimate their characteristics. Recent publications have begun to investigate the use of global meteorological “reanalysis” data sets for power system applications, many of which focus on long-term average statistics such as monthly-mean generation. Here we demonstrate that reanalysis data can also be used to estimate the frequency of relatively short-lived extreme events (including ramping on sub-daily time scales). Verification against 328 surface observation stations across the United Kingdom suggests that near-surface wind variability over spatiotemporal scales greater than around 300 km and 6 h can be faithfully reproduced using reanalysis, with no need for costly dynamical downscaling. A case study is presented in which a state-of-the-art, 33 year reanalysis data set (MERRA, from NASA-GMAO), is used to construct an hourly time series of nationally-aggregated wind power generation in Great Britain (GB), assuming a fixed, modern distribution of wind farms. The resultant generation estimates are highly correlated with recorded data from National Grid in the recent period, both for instantaneous hourly values and for variability over time intervals greater than around 6 h. This 33 year time series is then used to quantify the frequency with which different extreme GB-wide wind power generation events occur, as well as their seasonal and inter-annual variability. Several novel insights into the nature of extreme wind power generation events are described, including (i) that the number of prolonged low or high generation events is well approximated by a Poission-like random process, and (ii) whilst in general there is large seasonal variability, the magnitude of the most extreme ramps is similar in both summer and winter. An up-to-date version of the GB case study data as well as the underlying model are freely available for download from our website: http://www.met.reading.ac.uk/~energymet/data/Cannon2014/.

Scattered power from non-thermal, F-region plasma observed by EISCAT—evidence for coherent echoes?

Three rapid, poleward bursts of plasma flow, observed by the U.K.-POLAR EISCAT experiment, are studied in detail. In all three cases the large ion velocities (> 1 kms−1) are shown to drive the ion velocity distribution into a non-Maxwellian form, identified by the characteristic shape of the observed spectra and the fact that analysis of the spectra with the assumption of a Maxwellian distribution leads to excessive rises in apparent ion temperature, and an anticorrelation of apparent electron and ion temperatures. For all three periods the total scattered power is shown to rise with apparent ion temperature by up to 6 dB more than is expected for an isotropic Maxwellian plasma of constant density and by an even larger factor than that expected for non-thermal plasma. The anomalous increases in power are only observed at the lower altitudes (< 300 km). At greater altitudes the rise in power is roughly consistent with that simulated numerically for homogeneous, anisotropic, non-Maxwellian plasma of constant density, viewed using the U.K.-POLAR aspect angle. The spectra at times of anomalously high power are found to be asymmetric, showing an enhancement near the downward Doppler-shifted ion-acoustic frequency. Although it is not possible to eliminate completely rapid plasma density fluctuations as a cause of these power increases, such effects cannot explain the observed spectra and the correlation of power and apparent ion temperature without an unlikely set of coincidences. The observations are made along a beam direction which is as much as 16.5° from orthogonality with the geomagnetic field. Nevertheless, some form of coherent-like echo contamination of the incoherent scatter spectrum is the most satisfactory explanation of these data.

Financial viability appraisals for site-specific planning decisions in England

In England, appraisals of the financial viability of development schemes have become an integral part of planning policy-making, initially in determining the amount of planning obligations that might be obtained via legal agreements (known as Section 106 agreements) and latterly as a basis for establishing charging schedules for the Community Infrastructure Levy (CIL). Local planning authorities set these policies on an area-wide basis but ultimately development proposals require consent on a site-by-site basis. It is at this site-specific level that issues of viability are hotly contested. This paper examines case documents, proofs of evidence and decisions from a sample of planning disputes in order to address major issues within development viability, the application of the models and the distribution of the development gain between the developer, landowner and community. The results have specific application to viability assessment in England and should impact on future policy and practice guidance in this field. They also have relevance to other countries that incorporate assessments of economic viability in their planning systems.

Mapping power law distributions in digital health social networks: methods, interpretations, and practical implications

BACKGROUND: Social networks are common in digital health. A new stream of research is beginning to investigate the mechanisms of digital health social networks (DHSNs), how they are structured, how they function, and how their growth can be nurtured and managed. DHSNs increase in value when additional content is added, and the structure of networks may resemble the characteristics of power laws. Power laws are contrary to traditional Gaussian averages in that they demonstrate correlated phenomena. OBJECTIVES: The objective of this study is to investigate whether the distribution frequency in four DHSNs can be characterized as following a power law. A second objective is to describe the method used to determine the comparison. METHODS: Data from four DHSNs—Alcohol Help Center (AHC), Depression Center (DC), Panic Center (PC), and Stop Smoking Center (SSC)—were compared to power law distributions. To assist future researchers and managers, the 5-step methodology used to analyze and compare datasets is described. RESULTS: All four DHSNs were found to have right-skewed distributions, indicating the data were not normally distributed. When power trend lines were added to each frequency distribution, R(2) values indicated that, to a very high degree, the variance in post frequencies can be explained by actor rank (AHC .962, DC .975, PC .969, SSC .95). Spearman correlations provided further indication of the strength and statistical significance of the relationship (AHC .987. DC .967, PC .983, SSC .993, P<.001). CONCLUSIONS: This is the first study to investigate power distributions across multiple DHSNs, each addressing a unique condition. Results indicate that despite vast differences in theme, content, and length of existence, DHSNs follow properties of power laws. The structure of DHSNs is important as it gives insight to researchers and managers into the nature and mechanisms of network functionality. The 5-step process undertaken to compare actor contribution patterns can be replicated in networks that are managed by other organizations, and we conjecture that patterns observed in this study could be found in other DHSNs. Future research should analyze network growth over time and examine the characteristics and survival rates of superusers.

Testing projected wild bee distributions in agricultural habitats: predictive power depends on species traits and habitat type

Species distribution models (SDM) are increasingly used to understand the factors that regulate variation in biodiversity patterns and to help plan conservation strategies. However, these models are rarely validated with independently collected data and it is unclear whether SDM performance is maintained across distinct habitats and for species with different functional traits. Highly mobile species, such as bees, can be particularly challenging to model. Here, we use independent sets of occurrence data collected systematically in several agricultural habitats to test how the predictive performance of SDMs for wild bee species depends on species traits, habitat type, and sampling technique. We used a species distribution modeling approach parametrized for the Netherlands, with presence records from 1990 to 2010 for 193 Dutch wild bees. For each species, we built a Maxent model based on 13 climate and landscape variables. We tested the predictive performance of the SDMs with independent datasets collected from orchards and arable fields across the Netherlands from 2010 to 2013, using transect surveys or pan traps. Model predictive performance depended on species traits and habitat type. Occurrence of bee species specialized in habitat and diet was better predicted than generalist bees. Predictions of habitat suitability were also more precise for habitats that are temporally more stable (orchards) than for habitats that suffer regular alterations (arable), particularly for small, solitary bees. As a conservation tool, SDMs are best suited to modeling rarer, specialist species than more generalist and will work best in long-term stable habitats. The variability of complex, short-term habitats is difficult to capture in such models and historical land use generally has low thematic resolution. To improve SDMs’ usefulness, models require explanatory variables and collection data that include detailed landscape characteristics, for example, variability of crops and flower availability. Additionally, testing SDMs with field surveys should involve multiple collection techniques.

The calculative turn in land value capture: lessons from the English planning system

For over three decades, negotiated planning obligations have been the primary form of land value capture in England. Diffusing and evolving over the last decade, a significant policy innovation has been the use of financial calculations to estimate the extent to which policies on planning obligations for actual, proposed development projects and in plan making affect the financial viability of development. This paper assesses the extent to which the use of financial appraisals has provided a robust, just and practical procedure to support land value capture. It is concluded that development viability appraisals are saturated with intrinsic uncertainty and that land value capture that is based on such calculations is, to some extent, capricious. In addition, clear incentives for developers and land owners to bias viability calculations, the economic dependence of many viability consultants on developers and land owners, a lack of transparency, contested or ambiguous guidance and the opportunities created by input uncertainty for bias are further failings. It is argued that how viability calculations are applied has been, is being and will continue to be shaped by power relations.

A decision framework for considering climate change adaptation in biodiversity conservation planning

General principles of climate change adaptation for biodiversity have been formulated, but do not help prioritize actions. This is inhibiting their integration into conservation planning. We address this need with a decision framework that identifies and prioritizes actions to increase the adaptive capacity of species. The framework classifies species according to their current distribution and projected future climate space, as a basis for selecting appropriate decision trees. Decisions rely primarily on expert opinion, with additional information from quantitative models, where data are available. The framework considers in-situ management, followed by interventions at the landscape scale and finally translocation or ex-situ conservation. Synthesis and applications: From eight case studies, the key interventions identified for integrating climate change adaptation into conservation planning were local management and expansion of sites. We anticipate that, in combination with consideration of socio-economic and local factors, the decision framework will be a useful tool for conservation and natural resource managers to integrate adaptation measures into conservation plans.

CGIAR research program on grain legumes

EXECUTIVE SUMMARY Background and context The Grain Legumes CRP was established to bring all research and development work on grain legumes within the CGIAR system under one umbrella. It was set up to provide public goods outcomes to serve the needs of the sustainable production and consumption of grain legumes in the developing world, capitalising upon their properties that enhance the natural resource base upon which production so unequivocally depends. The choice of species and research foci were finalised following extensive consultation with all stakeholders (though perhaps fewer end users), and cover all disciplines that contribute to long-lasting solutions to the issues of developing country production and consumption. ICRISAT leads Grain Legumes and is partnered by the CGIAR centers ICARDA, IITA and CIAT and a number of other important partners, both public and private, and of course farmers in the developed and developing world. Originally in mid-2012 Grain Legumes was structured around eight Product Lines (PL) (i.e. technological innovations) intersecting five Strategic Components (SC) (i.e. arranged as components along the value chain). However, in 2015, it was restructured along a more R4D output model leading to Intermediate Development Outcomes (IDOs). Thus five Flagship Projects (FP) more closely reflecting a systematic pipeline of progression from fundamental science, implementation of interventions and the development of capacity and partnerships to promote and adopt impactful outcomes: FP1) Managing Productivity through crop interactions with biotic and abiotic constraints; FP2) Determination of traits that address production constraints and opportunities; FP3) Trait Deployment of those traits through breeding; FP4) Seed Systems, post-harvest processing and nutrition; FP5) Capacity-Building and Partnerships. Another three cross-cutting FPs analyse the broader environment surrounding the adoption of outputs, the capitalising of investments in genomics research, and a focus on the Management and Governance of Grain Legumes: FP6) Knowledge, impacts, priorities and gender organisation; FP7) Tools and platforms for high throughput genotyping and bioinformatics; and FP8) Management and Governance. Five FPs focus on R4D; FPs 5 and 6 are considered cross-cutting; FP 7 has a technical focus and FP 8 has an overarching objective. Over the three year period since its inception in July 1012, Grain Legumes has had a total budget of $140 million, with $62M originally to come from W1/W2 and the remaining $78M to come from W3/bilateral. In actuality only $45M came from W1/W2 but $106M from W3/bilateral corresponding to 106% of expectation. Purpose, scope and objectives of the external evaluation Principally, the evaluation of Grain Legumes is to ensure that the program is progressing in an effective manner towards addressing the system-level outcomes of the CGIAR as they relate to grain legumes. In essence, the evaluation aims to provide essential evaluative information for decision-making by Program Management and its funders on issues such as extension, expansion and structuring of the program and adjustments in relevant parts of the program. Subsequent to the formal signing of the agreed terms of reference, the evaluation team was also invited to comment upon the mooted options for merging and/or disaggregating of Grain Legumes. The audiences are therefore manifold, from the CGIAR Fund Council and Consortium, the Boards of Trustees of the four component CGIAR centres, the Grain Legumes Steering, Management and Independent Advisory Committees, to the researchers and others involved in the delivery of R4D outcomes and their partner organisations. The evaluation was not only summative in measuring results from Grain Legumes at arm’s length; it was also formative in promoting learning and improvements, and developmental in nurturing adaption to transformational change with time. The evaluation report was written in a manner that allows for engagement of key partners and funders in a dialogue as to how to increase ownership and a common understanding of how the goals are to be achieved. We reviewed research undertaken before the CRPs but leading to impacts during Grain Legumes, and research commenced over the past 2.5 years. For related activities pre- and post-commencement of Grain Legumes, we reviewed the relevance of activities and their relation to CGIAR and the Grain Legumes goals, whether they were likely to lead to the outcomes and impacts as documented in the Grain Legumes proposal, and the quality of the science underpinning the likelihood to deliver outcomes. Throughout, we were cognisant of the extent of the reach of CGIAR centres’ activities, and those of stakeholders upon which the impact of CGIAR R4D depends. Within our remit we evaluated the original and modified management and governance structures, and all the processes/responsibilities managed within those structures. Besides the evaluation of the technical and managerial issues of Grain Legumes, we addressed cross-cutting issues of gender sensitivity, capacity building and the creation and nurturing of partnerships. The evaluation also has the objective to provide information relating to the development of full proposals for the new CRP funding cycle. The evaluation addressed six overarching questions developed from the TOR questions (listed in the Inception Report, 2015 [http://1drv.ms/1POQSZh] and others including cross-cutting issues, phrasing them within the context of traditional evaluation criteria: 1. Relevance: Global development, urbanisation and technological innovation are progressing rapidly, are the aims and focus of Grain Legumes coherent, robust, fit for purpose and relevant to the global community? 2. Efficiency: Is the structure and effectiveness of leadership across Grain Legumes developing efficient partnership management and project management across PLs? 3. Quality of science: Is Grain Legumes utilising a wide range of technologies in a way that will increase our fundamental understanding of the biology that underpins several PLs; and are collected data used in the most effective way? 4. Effectiveness: Are Product Lines strategic contributors to the overarching aims and vision for Grain Legumes? 5. Impact: Are the impact pathways that underlie each PL well defined, measureable and achievable; and are they sufficiently defined in terms of beneficiaries? Does progress towards achieving outputs and outcomes from the major research areas indicate a lasting benefit for CGIAR and the communities it serves? 6. Sustainability: Is Grain Legumes managing the increasing level of restricted funding in terms of program quality and effectiveness, including attracting and retaining quality staff? Questions for the evaluation of governance and management focused on accountability, transparency, the effectiveness and success of program execution, change management processes and communication methods, taking account of the effects of CGIAR reform. The three crosscutting issues were considered as follows: i) gender balance in program delivery, e.g. whether each PL is able to contribute to the increased income, food security, nutrition, environmental and resource conservation for resource-poor women and men existing in rural livelihoods; ii) are internal and external capacity gaps identified/met, is capacity effectively developed within each product line, and are staff at all levels engaged in contributing ideas towards capacity building; and iii) is there effective involvement of partners in research and activity programming, what are the criteria for developing partnerships, how they are formalised and how is communication between partners and within Grain Legumes managed? It was not in remit to search for output, outcomes or impact, however as highlighted later, much of our time was spent on searching for information to support claims of impact, since Grain Legumes had no effective dedicated M&E in place at the time of undertaking the review. Approach and methodology The evaluation was conducted when Grain Legumes had been operational for approximately 3 years. The approach and methodology followed that outlined in the Inception Report [http://1drv.ms/1POQSZh]. The CCEE Team based its findings, conclusions and recommendations on data collection from several sources:  review of program documents, communications with the CO, minutes and presentations from all management and governance committee meetings  review of previous assessments and evaluations  sampling of Grain Legume projects in 7 countries1  more than 66 face to face interviews, a further 133 persons in groups and 4 phone/Skype conversations: ICRISAT, ICARDA, CIAT and IITA staff, partners and stakeholders. Meetings with one Independent Science and Partnership Council (ISPC) member.  meetings with over 100 people in 16 external groups, such as farmers’ groups  online survey completed by 126 (33.4%) scientists who contribute to Grain Legumes and a number of non-CGIAR partners and Management representatives  bibliometric review of 10 publications within each PL to qualitatively assess the design, conduct, analysis and presentation of results  quantitative and qualitative self-assessment of the contributions of each of the PLs to the six criteria and 3 cross-cutting issues of evaluation mentioned above completed by PLCs (see below). We reviewed the Logical Framework that underpins the desired Goals, or Impacts of Grain Legumes, and the links between the outputs and inputs as they related to the organisational units of Grain Legumes. The logical framework approach to planning and management of Grain Legumes activities implies a linear process, leading from activities, outputs, outcomes, to impacts, but within such an approach there may be room for a more systems dynamics approach allowing for feedback at every step and within every step, in order to refine and improve upon the respective activities as new results, ideas, and directions come to light. We then developed a matrix that summarised quantitatively and qualitatively the contributions of each of the PLs to the six criteria and 3 cross-cutting issues of evaluation mentioned above. Main findings and conclusions Grain legume production and consumption remain of great importance to the food security of not inconsiderable populations in the developing world, and merit sustained research investment. We conclude that Grain Legumes continues to contribute significant returns to research investments by the CGIAR, and such investment should continue. The global research community looks to the CGIAR for leadership in Grain legumes, but needs to be assured of value adding when bringing CGIAR centres under the expected umbrella of synergy. However, there is considerable scope for improving the efficiency with which outcomes are achieved. We note that an absence of an effective M&E has hampered the assessment of the effectiveness of proposed impact pathways. Likewise progress has been hampered by the limited numbers of research partnerships with Advanced Institutes and by budgetary constraints (lamented for their stifling effects on continuation of ongoing exciting research). The unworkable management structure constrains the CRP Director’s leadership role; responsibility without authority will never lead to effective outcomes. Good fortune is responsible for many of the successes of Grain Legumes, underpinned by a devoted work force across the participating CGIAR centres and partners. The quality of the science is not uniformly high, and we believe that mentoring of scientists should be given priority where quality is poor. Simplified yet informative reporting is an imperative to this. World class science underpins the identification of, and molecular basis for, traits important for yield improvement and this expertise should be extended to all grain legume species, capitalising upon the germplasm collections. The linking of Grain Legumes with regional research and development consortia has been very successful, with outcomes aligning with those of Grain Legumes. We see that with declining funding consolidation of research effort based on likely successes will be necessary, and welcome the move afoot to incorporate grain legumes into an agri-food system focused on successful value chains that deliver sustainable outcomes. Relevance and Strategy Grain Legumes has geographic and disciplinary relevance, addressing the major supply chain issues of variety development seed system and agronomy, with some attention to quality and postharvest marketing systems. The CRP has provided the opportunity to cut ongoing and to initiate new research. Research funded by the Gates Foundation (Anon, 2014) suggests that the need for improvement is greatest in Africa and advocates reducing the number of crop by country combinations when resources are sparse. The lesser research investment in Latin America, however, is not in line with the regions’ dependency on legumes. In spite of the fact that there is no evidence of strong inter-partner CGIAR centre or internal synergy, the program is still moving ahead on most fronts in line with the overall project logframe. This is in spite of continual pushing and pulling by in particular donors and the CO. However, to quantify real impact, we believe Grain Legumes must have access to reliable baseline data on production and consumption, and this is missing. Similarly, there is little evidence of the proposed ‘Inclusive Market Oriented Development’ (IMOD) framework being used to assist with priority setting. The product lines, eight of which cover most of the historical programmes in place in the partner CGIAR centres at the commencement of the Grain Legumes, do not cover all the constraints for formal constraints analysis was not undertaken at the inception of the Grain Legumes, and some of this additionally identified research is undertaken under the umbrella of the FPs; this needs to be rationalised. We found the PLs to be isolated in activity, even with minimally-integrated activities within each PL, with little evidence of synergy between PLs. Even though the SCs should ensure a systems approach, as with the new FPs, we did not get a feel that this is so. The underplaying of agronomy, and production practices may be one reason for this. We believe that treating legume crops as if they were horticultural crops will increase farmer returns from investment. The choice of Flagship Projects makes sense, with the flow of activity firstly around crop management and agronomy followed by the logical sequence of trait discovery, incorporation into improved varieties, dissemination of those varieties through appropriate seed chains leading to market impacts, and the capacity building required at all steps. One obvious omission, however, is the lack of a central and strategic policy on the role of transgenics in Grain Legumes. We found four notable comparative advantages for Grain Legumes: the access to germplasm of component species, the use of the phenotyping facility at ICRISAT, the approach for village level industry for IPM, and the emphasis on hybrid pigeonpea. Efficiency Each centre has strong control of, and emphasis on, their ‘species’ domains, and ownership of the same detracts from possible synergy. Without synergy or value add, the Grain Legumes brings with it no comparative advantage over each centre continuing their own pre-CRP research agendas. We found little evidence of integration of programmes between centres and almost no cross-centre authorship of publications, such as could have occurred with the integrated cross-centre approaches to stress tolerance including crop modelling: the one publication (Gaur et al., 2015) on heat tolerance by ICRISAT, CIAT and ICARDA does not provide any keys to inter-centre collaboration. The integration of each centre with NARS and university research programmes is good, but the cross-centre links with NARS are poor. A better coordinated integration with Grain Legumes, , rather than through the individual centres, may reduce transactions costs for NARS, Monitoring and evaluation is, as noted throughout our report, one area of Grain Legumes research management that has not been given the attention it should have received. If it had have received proper attention, some of the issues of poor efficiency might have been nipped in the bud. A strong monitoring and evaluation system would have provided the baseline data and set the milestones that would have allowed both efficiency and effectiveness to be better appraised. We found no attempt to define comparative advantages of the CGIAR centres and their R4D activities, although practice showed the better grasp of CIAT in developing innovative seed distribution systems. During field visits and interviews, the CCEE Team observed shortcomings in the communication processes within Grain Legumes and with the broader scientific community and the public. For example, the public face of the program on the internet is out of date. Survey findings, however, suggest that information is shared freely and routinely within the PL within which scientists work. Some external issues, such as those with funding, low W1/W2 and poor sustainability of funding (especially if funding is top heavy with a few agencies), undermine research investment and confidence of partners in the system (e.g. as voiced by researchers working on crops and countries not included in TL III and the cessation of ongoing competitively-funded projects especially in India), but other issues attributable to the governance and management of the Grain Legumes, such as opaque integration of W3/bilaterals with W1/W2 funding require attention. Offsetting this, the existence of the Grain Legumes did mobilise additional funding [that it would not have if Grain Legumes did not exist]. We were concerned that Grain Legumes is simply not recognised outside of the CRP, with a limited www presence and centres promote themselves, rather than Grain Legumes (with exception in IITA). This is not a good move if one wishes to increase investment in the Grain Legumes. Although funding agencies require cost:benefit ratios, for example for each PL we faced difficulty in determining comparative value for money between investment in different types of research, and in being able to clearly attribute research and development outcomes to financial investment. There was also a time CCEE frame issue too. There is poor interaction with the private sector, notably in areas where they have a comparative financial advantage. We questioned in particular the apparent lack of interaction with the major agro-chemical companies, with respect to the development of herbicide tolerant (HT) grain legumes and the lack of evidence that the regulatory and trade aspects related to herbicide tolerant crops had been considered. Quality of science The quality of the science is highly variable across Grain Legumes, with pockets of real excellence that are linked to good levels of productivity, whereas other PLs are struggling to deliver quality publications, and outputs and outcomes that are based on these. There is much evidence of gradualism in terms of research output and outcomes, i.e. essentially the same activities that were ongoing at the time of the launch of Grain Legumes are still in place. However, there are examples of game changers including those from valuable investments in genomics, phenotyping, and bio-control. We were pleased to see large proportions of collaboration on publications with non-CGIAR centres, reflecting cooperation with partners in developed and developing countries. The value of collaboration when ensuring quality of science cannot be stressed highly enough both within the CRP, and with other global and national partners. PLs should be given incentives to collaborate with other CRPs and external institutions. There is little cohesion between PLs and with other CRPs as evidenced by publications, although there are some exceptions. We suspect the reasons for this are driven by funding. Productivity from the different PLs is also highly variable and it is not clear what other activities staff are engaged in since, in some PLs, they do not appear to lead to quality publications. Effectiveness Grain Legumes has been very effective in addressing component issues of research, but not the continuum from variety development to legumes on someone’s dinner plate. Our overall assessment of the effectiveness of Grain Legumes in stimulating synergy, innovation and impact indicate that gradualism is more prevalent than innovation. It also shows, as do publications, that there is little integration of disciplines or a focus on ‘systems’. The absence of socio-economists from research teams is evident in the general lack of an end user focus. However, research on genomics, plant breeding and seed systems have made great strides forward, on the brink of delivering impact. Agronomy has been a poor sister, but some of the competitive grants within Grain Legumes have unearthed some potential game changers, such as objective use of transplanting as an agronomic practice. As mentioned earlier, the lack of effective M&E (however, this was part of some major projects such as TL II/TL III), and therefore the ability to monitor impact pathways and achievement of impact, implies no systematic management of data. This creates difficulty when attempting to evaluate the achievement of the Grain Legumes objectives. One might have expected at least one attempt to try to develop publications between centres arguing for similar biologies/research approaches, bringing species together under one umbrella, but we did not find any evidence for this. It is most unfortunate that, due to budgetary cuts, the new ‘schemes’, e.g. competitive grants and scholarships, were cut off before gaining a foothold. With 8 species addressed by Grain Legumes, it is not unexpected that there will be little evidence of shared protocols across centres/species. One rare example was that hosted by the United States Department of Agriculture (USDA) on shared methods for phenotyping of legume germplasm. Researchers from CIAT, IITA, ICRISAT and three USDA stations attended, focusing in simple canopy temperature and root morphology measurements. It is our belief that as a set of research centres, the CGIAR centres should be focusing on the research for which they have a comparative advantage. While imposing the restructure to FPs, which is fine for development objectives and outcomes (funded through W3/bilateral), it is less so for a research institute, and the structure should not detract from the more basic work expected of an international CGIAR centre (or set of centres as in a CRP). Impact It is well known that research does not always lead to scientific breakthroughs. Also, activities such as plant breeding are long term; making impacts difficult to assess. We believe that sufficient progress with genomics and associated research has been made to warrant impact, but we are unable to quantify the levels of impact, or the timeframe for the same. Work in Grain Legumes has enormous potential for real impact in scientific research, commercial, farming, smallholder and household communities, much of which is being realised. However, the PLs need to become more adept at providing convincing cases that are strongly evidenced for these impacts, as this is likely to be a key factor in leveraging future funding. Claimed gains must be referenced against baseline data, and these are not always readily available. The CCEE Team realises that such impact evaluation represents a significant drain on resources, and Grain Legumes should determine whether the balance of costs to benefits favours such investment. Interviews conducted by the CCEE during site visits showed that PLs are quantifying the area of adoption of varieties, but in most cases they are not measuring the impact on environment, health/nutrition. Since the health and nutritional benefits and the environmental gains from growing legumes are major arguments for supporting grain legume research, the community is currently missing substantial opportunities to strengthen its own case for continued support. Whilst there are some impressive examples of considering the whole value chain, e.g. white beans from production through to export; in the main, the pipeline to end user is somewhat piece-meal, with no clear definition of the end user nor differential responsibility of Grain Legumes and of partners. The lack of robust time-defined impact pathways is highlighted in Section 7.4, and even though developed for PL5, timeframes are essential for measuring progress against prediction. Sustainability In summary, there is general acknowledgement that future funding is likely to become more limited, specifically in W1&2 and there is understandable concern over the support for the staff and basic infrastructure that underpin the Grain Legumes programme. For example, it is reported that staffing in parts of CIAT has been dependent on W1&2 and that this is too unstable to re-establish a critical mass. The present system whereby W3 and bilateral projects do not pay a realistic level of overheads means that such projects are being effectively subsidised by W1&2. This position is not sustainable in the long term as there will be a progressive but definite loss of basic skills and resources in the core centres. The only obvious options to prevent this outcome include a severe reduction in the fixed costs of the centres and/or a refusal to accept W3 and bilateral funding with an inadequate overhead component. In the latter case, there is an obvious danger that funders will move their resources away from the CGIAR system towards other, perhaps less expensive, suppliers of research, and possibly more relevant development expertise. This issue must be addressed. As the Grain Legumes moves into the future, and if sustainable funding cannot be assured, decisions must be made concerning a reduction in activities, keeping some caretaker breeding maintenance, and focus (as has TL III) on fewer species and a reduced geographic focus. Cross cutting issues: Gender, capacity building and partnerships Gender is not mainstreamed, but there is some evidence that this is improving, especially with dedicated gender specialists and the slow integration of gender across CRPs. There is a need to approach gender through the vision of agriculture as a social practice, with recognition of what changes will be acceptable culturally and what not, and capitalising upon the perceived and actual features of production and processing that grain legumes are primarily women-based crops. Gender awareness may be high among Scientists, but it appears to be a predominantly passive attribute with few proactively seeking opportunities for gender equity. It is, however, a sound sensitivity base on which to build. Nevertheless, examples of notable gender initiatives were identified during field visits. For example, in Benin, the development of biocontrol technologies has enthusiastically integrated diversity, engaging with women farmers’ and youths while maintaining cultural norms. Women are gathering and processing, youths are taking the product to market. The implication is that several groups benefit, rather than domination by the majority group. In Malawi, innovative approaches have been developed to improving nutrition for children, such as incorporating nutrient enriched bean flour products into snacks. In India, scientists collaborating with gender scientists and socio-economists are identifying the impact of mechanical harvesting on agricultural labour and the potential displacement of female labourers. In Kenya, a novel initiative is improving the accessibility of certified seed for new varieties. Seed suppliers have introduced small packs of grain legume seed at low unit cost, which are being purchased by young people and women. Capacity building efforts for external partners are not clearly aligned with the research mandate and delivery of Grain Legumes. However, there are a number of training activities that are being undertaken by Grain Legumes, largely through the W3/bilateral project. Gender balance never reaches parity, but it appears that efforts are made to include female participants. Within the evaluation timeframe it was not possible to conduct external surveys to further validate or review external capacity building efforts in Grain Legumes. Training of scientists is significant, with >40 benefiting. Postgraduate training is varied across PLs, and there is some opportunity to increase the numbers being supervised. We consider that support for postgraduates at ICRISAT could be better coordinated, satisfying more of the students’ needs. It is important, however, to follow up investments in capacity building by monitoring effectiveness, career progressions and so on. Training activities appear to be rather centre-specific, not following a coordinated programme managed by, nor at the level of, the Grain Legumes. Numbers of persons trained and their gender are important, but a measure of the effectiveness of the training is more important. Although optimism is expressed by the great majority of Research Managers that partnerships were working well to leverage knowledge and research capacities, scientists have a less favourable view, particularly in terms of their incentives to participate. It seems likely that the activities taking place within Grain Legumes were, in the most part, continuations of previous collaborations. This is not surprising in light of the reduction in the emphasis on partnerships as Grain Legumes evolved to a funded project, and the consequent lack of opportunity and ambition for establishing novel partnerships. Where they exist, partnerships are good on the whole, especially with US. They could be expanded where comparative advantages exist (for example with Canada and Australia for machine harvestable legumes), but some earlier identified partnerships, e.g. with Turkey, have not been capitalised upon. Others experience problems of variety access (the embargo on exports of some sources of materials from India), yet others do have relevance e.g. imported Brazilian varieties in pre-release in Ethiopia (even though two of the three are from CIAT materials). Governance and Management The standard format of committee structure and responsibilities is common to other CRPs, as are the attendant problems. One of the major problems is that the Grain Legumes Director has responsibility but no authority; hence, even with the support of the RMC, the Director is unable to ‘direct’ in the literal sense of the work the activities of Grain Legumes. We also see the same sense of helplessness with the role of the PLCs. They have responsibility but no authority in managing the affairs of their PL, and they have no access to funds with which to promote intellectual collaboration and cooperation. Minutes from governance and management meetings do not reflect the compromised weak position of the Director and the associated difficulties in the management of Grain Legumes. Nor do the minutes reflect concerns about the amount of time spent by scientists in meetings for planning, integration, evaluation and reporting. Many scientists reported significant opportunity costs in participating in the ongoing imposed [by the CO] evolution of Grain Legumes and CRPs in general. The changes brought in by the CO have not helped promote any greater authority and capacity of the Grain Legumes Director to direct. Likewise, they do not address any of the issues with the conflict of interest in having the Lead Centre chair the Steering Committee. Indeed, we believe that the combining of the Steering Committee with the Independent Advisory Committee, besides becoming unwieldy in number, annuls any sense of independence in advice offered to the Grain Legumes management. We have concerns with the declining proportion of W1/W2 funds (as expressed in the section on Sustainability), and believe that when basic financial planning takes place, integration of W1/W2 and W3/bilateral sources must occur, and be linked to anticipated outcomes and impacts. This will ensure a close alignment of collaborators’ and partners’ objectives and contributions to that of the Grain Legumes. We also queried the process for, and the formality, or lack of, surrounding, the approval of annual budgets, and the level of priority setting when budgets are cut. Recommendations for Grain Legumes The CCEE Team makes the following recommendations, critical issues are highlighted in bold, and those that require action by an entity other than the Grain Legumes Research Management Committee or Project Management united are identified in a footnote. Relevance and Strategy Recommendation 1: A period of consistency is necessary to raise confidence, morale and trust across scientists, managers and partners to foster the assembly of enduring Grain Legumes outcomes2.  There needs to be a concerted effort to undertake baseline studies and to implement a robust M&E activity during this period. Without these data the foundation for integrated research in grain legumes is jeopardised.  There is a strong need to link more closely with the private sector, especially where there are financial and other comparative advantages to do so. Recommendation 2: The agronomic and physiological trait targets of Grain Legumes (tolerance to changing climate patterns, to the pests and diseases of today and of the future, incorporation of quality traits and adaptations to intensive production systems [machine-harvestability and herbicide tolerance], and short season high yielding characters) are all worthy of continued investment when selecting for improved varieties.  There needs to be a common strategy, implemented across centres and species, as to how to address these trait targets through conventional and modern breeding approaches, but only if adequate funding is assured and secured and if a consistency and unity of purpose can be achieved across a large-scale. This should take the form of cross-species coordinated research programmes to address these breeding targets that cooperate across centres and make efficient use of facilities and other resources.  The CRP should undertake a detailed strategic review of the role of transgenics across the range of targets in the mandate crops. Efficiency Recommendation 3: The lack of an effective M&E process is a significant omission, not least in terms of more efficient use of resources and the lack of baseline data with which to measure impact, and must be rectified.  Reinforcing Recommendation 1, an effective M&E system initially directed towards baseline studies must be implemented.  Transaction costs may be reduced through bilateral projects, which are seen as more cost effective than W1/W2 where transaction costs are disproportionately higher. Recommendation 4: To improve communication and coordination within the CRP, and with a broader audience:  There is a priority need for a central database containing, names of staff associated with Grain Legumes and their time commitments, their responsibilities, and involvement in CRP activities, their progress and achievements, their publications, plans of training, travel, and other opportunities for interaction.  Regular global meetings of staff involved in managing PLs, the entire CRP management staff and the IAC are essential for effective coordination of all activity within Grain Legumes.  The website must be given a complete overhaul and improvement and then regular maintenance must be provided to keep it current. Quality of Science Recommendation 5: It is essential to continue investment in good science and to institute a change from gradualism in research output and outcomes to an expectation of innovative and concrete achievements that can be attributed clearly to people, centres and core facilities.  A cost:benefit analysis and subsequent strategic planning must be undertaken to justify further investment in the genomics and phenotyping facilities at ICRISAT especially as such technologies advance rapidly. Strategic planning and coordination must also be implemented for capitalising on the investment in crop simulation modelling. (The phenotyping facility of ICRISAT needs to focus on delivering some outcomes, not only outputs.)  PLs should be given incentives to collaborate with other CRPs and external institutions. The CCEE recommends special recognition of high quality collaborative papers, thereby encouraging increased quality of the research programmes and widening the penetration of research impacts.  More importance should be placed on the quality of publication, rather than quantity of outputs and there should be recognition of other types of outputs from Grain Legumes. The CRP Director must be party to this.  If staff are engaged in activities that relate more to impact than publication then this needs to be monitored and recorded and a clearer understanding developed of what constitutes a pathway to impact and how success of such activities can be evaluated. A system must be devised and incorporated into the M&E to enable recognition of other types of outputs (non- publication based) from Grain Legumes, e.g. varieties for breeders. Effectiveness Recommendation 6: To develop greater synergy, Grain Legumes should review management processes and the direction of research activities. In particular, far more extensive integration of research and knowledge exchange should take place across both African and Asian continents so that the best aspects of both can be shared. A multidisciplinary approach is recommended that considers processing solutions, as well as breeding solutions, to capitalise upon the nutritional benefits of the grain legume crops. We recommend:  A better collaboration with social scientists at the design stage of experiments in order to improve the utility of the work carried out and to understand its reach.  Supporting3 the adoption of best practice electronic data collection, central storage and open access, particularly of genomic data, for public use.  Given the focus on the link between phenotyping and genotyping, we note that there is a lack of congruence between the populations that are being phenotyped and those being genotyped, and therefore these could be better aligned within each species.  Concentrating investment external to Grain Legumes on scaling up production of varieties with the most promising trait profiles to meet the basic seed requirement.  Developing a more holistic approach that coordinates an understanding of the disease pathology and epidemiology, and of new chemicals before they become commercially available, together with agronomic practice such that recommendations can be made for growers. Continuing work to establish whether agronomic factors hold true in different environments and to assess GxE effects within breeding programmes. Such rigorous trial practices should be used to inform the evaluation of breeding lines and to provide phenotype data to associate with markers for traits such as heat, drought and herbicide tolerance.  Considering grain legumes as if they were vegetable crops in terms of the strategy for intensification of production, both from the management perspective and for seed systems, will be a useful development objective into the future. This will bring about more rapid intensification and is likely to increase farmer returns from investment. Recommendation 7: The CGIAR centres should focus in on the research for which they have a comparative advantage. While imposing the restructure to FPs, which is fine for development objectives and outcomes (funded through W3/bilateral) it is less so for a research institute, and should not detract from the more basic work expected of an international CGIAR centre (or set of centres in a CRP).  Collaborative approaches should be explored within Grain Legumes, e.g. similar biologies/research approaches, bringing species together under one umbrella. Similarly better alignment is needed to address the lack of congruence between the populations that are being phenotyped and those being genotyped.  Despite positive impacts from research in genomics, plant breeding and seed systems, the lack of an effective M&E, already mentioned elsewhere, has reduced the ability to monitor impact pathways. This must be addressed.  The absence of socio-economists from research teams is evident in the general lack of an end user focus. Responsibilities of the different actors in the whole value chain must be considered and identified when developing impact targets, and the pathway leading to them, for individual projects. People with socio-economist skills must be part of the team from project inception so that appropriate frameworks are incorporated for measuring and influencing sociological and economic changes brought about by Grain Legumes research. Impact Recommendation 8: PLs need to become more adept at providing convincing cases in which impact is strongly evidenced, as this is likely to be a key factor in leveraging future funding.  Claimed gains must be referenced against baseline data, and these are not always readily available. The CCEE Team realises that such impact evaluation represents a significant drain on resources, and Grain Legumes should determine whether the balance of costs to benefits favours such investment.  It is essential that Grain Legumes provides training to staff on what constitutes impact and how it can be recorded.  Specific, rather than generalised, potential impacts arising from activity within Grain Legumes should be defined at the time of justifying the programme of work and a pathway to impact should form part of the documentation prepared ahead of a piece of research commencing. . In other words, centres should submit work plans to Grain Legumes before they are undertaken using W1/W2 funds Recommendation 9: The reporting activity must be streamlined to a single (brief) format that can be used to report to Grain Legumes, Centres and to donors for special project activities4. Sustainability Recommendation 10: As Grain Legumes moves into the future, and if sustainable funding cannot be assured, decisions must be made concerning a reduction in activities, keeping some caretaker breeding maintenance, and focus (as has TL III) on fewer species and a reduced geographic focus. Zeigler (Director General of IRRI) states “…time and effort would be better spent … making tough decisions about which programs deserve the precious support.”  The present system whereby W3 and bilateral projects do not pay a realistic level of overheads means that such projects are being effectively subsidised by W1&2 and there will be a progressive but definite loss of basic skills and resources in the core centres. To prevent this outcome it is necessary to significantly reduce the fixed costs of the centres and/or refuse to accept W3 and bilateral funding without an adequate overhead component.  In the absence of long term certainty, the scale of the budget allocated to each of the new CRPs should be very conservative, a feature that can only be achieved by restricting/reducing the scope, probably quite significantly. Cross cutting issues: Gender, capacity building and partnerships Recommendation 11: The challenge for Grain Legumes is to achieve pro-active gender mainstreaming, which facilitates opportunities for gender diversity within all activities, from employment processes through research to end users.  Strategic measurable gender indicators need to be embedded in research design, for instance, through specific IDOs for each of the flagships projects. Accurate baseline data are also required to facilitate M&E reviews of progress.  Implementation of the Gender Strategy is the responsibility of everyone, not solely the Gender Team. Thus, ownership could be encouraged by setting personal development for key personnel objectives with specific outcomes, e.g. employment practices or research outcomes.  Recognising the positive gender initiatives in progress or planned, feedback must be communicated and integrated into broader research planning to share opportunities, methods and outcomes.  In addition to promoting gender equity in research, Grain Legumes also needs to ensure that working environments are gender sensitive and that recruitment processes, including promotion opportunities are equitable. Gender imbalance in management should be actively examined to identify further opportunities for developing female leadership. Recommendation 12: It is recommended that a training plan be devised to ensure that capacity building efforts are more clearly aligned with the research mandate, delivery and timeframe of Grain Legumes. Moreover, we recommend that ICRISAT develop a strategy to treat their new cohort of researchers more equitably in the future. Recommendation 13: To develop a more coherent strategic programme designed to eliminate overlap and promote synergy between programmes with common aims, Grain Legumes should hold a meeting with a range of partners. Governance Recommendation 14: Governance processes should be re-assessed and the structure altered to ensure that the Grain Legumes Director has the authority and budget control to drive the execution of strategy.  The ISC should be truly independent and given the power to influence strategic decisions before they become final. We also recommend that PLCs are provided with the authority to manage the direction and finances of their PL; and that ring-fenced funds are provided for the promotion of collaboration, coordination and staff training5. The way ahead In our view, having seen the ineffectiveness of much of the attempts [or lack of attempts] to harness synergies between multiple centres, and of the strength in few or sole centre partnerships, we believe that there is little to justify a full retention of the 8 legume species and 4 CGIAR centres in a CRP. TL I and II and PABRA have shown to be reasonably good cross-centre and single centre integrated programmes, but even they suffer from incomplete value chain approaches to increasing rural incomes while increasing food and nutritional security; they both need multi-faceted solutions which are not immediately forthcoming from Grain Legumes. It is important to embed Grain Legumes research within the agri-food systems these crops serve. Figure ES1 broadly shows the perceived current and potential degrees of synergy between centres, PLs and species, and is discussed more in the text. It is clear that the value chains for individual species from trait determination to nutritional impact have more cohesion than do the individual activities (e.g. trait deployment) across species. For this reason we believe that the future for research in Grain Legumes is best addressed by focusing on each of the species separately, and within an ecosystem framework; any synergy for research across species can be effected through communication and not necessarily through obligatory cooperative research. The ecosystem framework will allow for strengthening of agronomy type systems research, the arguments for benefits of inclusion of grain legumes in cropping systems, which is notable by its absence in much of what Grain Legumes currently undertakes. Figure ES1. Current and potential degrees of synergy between centres, PLs and crop species We therefore agree with the innovation in agri-food systems approach of the CG, and believe that Grain Legumes rightly belongs in the Dryland Cereals and Legumes Agri-food Systems. We believe that the option of combining the crops of dryland cereals and legumes in the cereal-legume-livestock systems of subsistence farming communities for whole-farm productivity is closest to the best way forward. Indeed the inclusion of grain legumes may not warrant even a CRP alone, rather the legume components should fit in with the major crops that determine the production systems. Legumes will always be subservient to the major cereals, as necessary adjuncts to the whole production system, providing both nutritional diversity and environmental services, neither achievable from cereals alone. Figure ES2. Most suitable option for integration of Grain Legumes and Dryland Cereals into an Agri-Food Systems CRP Most suitable option for integration of Grain Legumes and Dryland Cereals into an Agri-Food Systems CRP, which  Incorporates ex-Dryland Systems, Dryland Cereals, Grain Legumes, some HumidTropics, some ex-Livestock &Fisheries into a new CRP  Will cover full agri-food system VC for all 8 legumes in all ecologies, but must interact (dock) with the relevant AFS-CRPs for the dominant cereal in the relevant ecology  Hence, will need to negotiate with other Agrifood Systems-CRPs on who does what for legumes  In addition, responsible for sorghum and millet in the mixed dryland crop-livestock agro-ecologies For major game changers to be effected, we believe that the game has to change, and there is little evidence of this. The direction of CRPs is the correct route, but the journey has not yet come to its destination. A major change of game [such as the adoption of a Flagship Project approach as exemplified by the Australian CSIRO – where flagships contract services from centres of research excellence] would be painful to implant. The CGIAR system is going down the right pathway but it has not gone far enough.