Discussion Paper No. 1 "is a practical, methodological introduction to setting up a participatory monitoring process for sustainable agricultural initiatives." The contents are based, in part, on the authors involvement in action research on methods of monitoring and impact assessment suitable to the needs of small-scale producers, rural workers unions, and NGOs in Brazil. The impact of this field based experience is evident in the frequent (and appropriate) cautions and warnings given to the reader about the application of various principles, methods and procedures. This research is also a practical example of the author learning by "monitoring the monitoring", one of the processes recommended in the paper. The second source of experience is the existing literature on participatory monitoring and evaluation. This is put to good use by appropriate quotations of key concepts and examples of practice, and helps give the paper breadth, complimenting the depth of experience available from the action research.

The paper has two main sections: the main body of the paper and a large annex. The paper itself starts by context setting, explaining in clear terms the relevance of participatory approaches from the perspective of different actors This is followed by a useful clarification of key terms, and then four chapters on "how-to-do-it". The text manages to avoid jargon and mystification without being too simplistic or normative in its approach. As with much of the literature on M&E there is a strong focus on indicators and methods of data collection. The absence of any reference to "logical frameworks", has probably simplified the authors task, but nevertheless the exposition on indicators can still be easily related to the "logical framework", by those who use it.

The main annex, making up half the text as a whole, describes 21 different participatory monitoring methods. These are supported by one or more graphical examples of the method, or its documentation. Given the diversity of participatory methods of inquiry that has developed under the rubric of PRA, and its descendants, this is an ambitious task. In the process of being summarised some of the method descriptions suffer from their enforced brevity, leaving some questions unanswered . Nevertheless, there were still interesting examples of relatively unknown methods (to the reviewer), and the commentary on those that were known often added value in terms of new ways of using those methods.

The author has already noted some important areas of omission. These include "how to construct a sample in a participatory manner, whether or not to use a control group, or how to negotiate indicators with many different groups and participants" The latter limitation is also noted by the same author in respect to the wider literature on participatory M&E, in Discussion Paper No. 2, reviewed below. More than once in Discussion Paper No. 1 the reader is sensibly warned to avoid having too many indicators and to be realistic about what data can be collected let alone analysed. At the same time it is correctly emphasised that participatory approaches by necessity mean that there will be a diversity of views on what indicators are appropriate, and that these will change overtime. Various references to the need to negotiate indicators seem to imply that, like Marsden and Oakley (1994), resolution must come through some form of joint agreement by the different stakeholders. However, the analysis in Discussion Paper No. 2 shows that the authors see the need for solutions that are both more realistic and more sophisticated. Ideally, the author's useful list of indicators of how well a monitoring process is fulfilling its objectives (page 45), would include the presence of different criteria of value (gendered and otherwise) being used in impact assessment analyses and documentation.

Another area touched upon repeatedly in the paper is the task of analysis, as distinct from data collection, and how it might be made participatory. It is these latter tasks, as well as the dissemination of results and analyses, which are typically neglected, relative to the planning stage of monitoring systems (and in doing so, mirroring many projects themselves). An explicit section on this aspect of practice would be of value, in any future texts of this type.

Discussion Paper 2 differs from the first in two respects. It focuses more specifically on monitoring "environmental changes and the impact of natural resource management interventions" rather than the more inclusive "sustainable agricultural initiatives". Although still very practical in its contents is also more of a review than a guidebook, placing more emphasis on the analysis of published literature. While the text in both Discussion Papers is well supported by relevant examples, giving real life details, difficulties and responses, in this second paper paper there is smaller and more detailed set of documented methods in the annex. These are representative of what the authors argue are three main differences in approaches (PRA based methods, oral testimony methods, and ecology based methods). The merits of each of the 10 examples are also usefully summarised using a form of matrix ranking, containing criterion of suitability developed in the main text.

The paper starts with an introductory explanation of what monitoring is. In the process, the authors identify the areas of overlap, as well as difference, between monitoring and evaluation, and between what are called scientific and participatory approaches to monitoring. The authors are not recommending another set of polarities or reversals, but an appreciation of context and the practical implications for the choice of method. One of these, for example, is scale: the spatial coverage and frequency of iteration that is needed. One of the central messages, underlying their whole analysis is the importance of knowing who the end user of the information will be and what their needs are.

The largest section of the paper is analysis of participatory monitoring. The first part summarises points made in the first paper, the forces pushing for participatory monitoring, and an outline of how participatory monitoring can be undertaken. Three main sets of issues are then examined: the potential and actual benefits of participatory monitoring; the different forms of participation in monitoring, and the thorny subject of developing indicators. As in the first paper, the focus is on monitoring approaches that involve partnerships of multiple stakeholders, and especially the question of what the role is for different stakeholders in each of stage of monitoring. The final part examines the tension between demands for rigour and participation, and the variety of ways of viewing rigour, and other competing criteria, for evaluating monitoring methods. The issues are dealt with competently, without idealisations of practice or corrosive scepticism. Problems are identified but there is a sense of innovation taking place in the face of real life difficulties.

The paper ends with a summary of "key findings and knowledge gaps" identified in the text. The impact of participatory monitoring is left as an open question, but one which needs examining. The question of the appropriate roles for different stakeholders is highlighted, and attention is drawn to the fact that most approaches used to date still treat rural communities or "local people" as one stakeholder, rather than as multiple stakeholders, having significant differences in interests (within and between households). Costs in money, staff time, and most importantly "local peoples" time, is recognised as a major issue that must temper expectations about the scale and intensity of participation. The design of participatory monitoring is very much about trade-offs, between different expectations, and capabilities, a point that is well made throughout the paper. The issue of baselines, recognised as important, is dealt with realistically with suggestions at that at a minimum what is needed is comparable sets of observations over time, though their contents may change as understandings about objectives and information needs evolve. The final question is about how monitoring based analyses are fed back into and effect development activities. Monitoring systems assume this takes place but rarely is evidence sought for this sort of impact. Nevertheless, there is on both Discussion Papers plenty of useful analysis, suggestions, examples of practice and leads to other useful sources, for those organisations who do want their monitoring activities to have an impact.

In contrast to the Discussion Papers reviewed above, this book focuses more on evaluation than monitoring. It is concerned with bounded and identifiable "development projects" rather than wider issues of environmental change. The author's main aim is to build an inclusive model of evaluation practice, within the field of development aid projects. It is not intended as a guide to practice or a literature review.

In the first chapter sets the context for the rest of the book. Key terms are introduced (e.g. strategic / operational planning, programmes / projects). The project matrix and basic information requirements of project documents are covered briefly but well. Distinctions are then made between appraisal, monitoring and evaluation, the purpose of evaluation and the "underlying rationality". These distinctions form the basis of the evaluation frameworks developed in the rest of the book. His introduction also manages to ground what otherwise might be seen as an almost normative account of evaluation, in some historical context (changing views of evaluations purposes and methods) and in some cultural context (how different types of rationality effect the concerns of evaluations and evaluators). A brief reference to the development of variants forms of the logical framework might also have been of value.

In the second chapter the author develops, step by step, seven different "development frameworks". These are summarised in diagrams (similar to flow charts) which enable the author to incorporate parallel and circular processes, in addition to simple linear sequences. The first framework links what he calls the "core variables of evaluation" (efficiency, effectiveness, relevance, impacts, sustainability) to different stages of the cause and effect sequence in the Logical Framework. The subsequent frameworks incorporate feedback loops via monitoring and evaluation, and then a list of mediating categories of organisational structure and processes. The fourth version inserts "institution building" activities into the main sequence of events. Contentiously, he states that impact here must still be assessed in terms of the impact on the end clients' lives, not changes that take place in the assisted organisation. The fifth framework focuses on the impact stage. The author argues for beneficiary based judgements as the best means of obtaining holistic assessment of impact, one that takes due account of wider non-project related forms of social change. At this point the framework diagram reaches the limits of its usefulness, the potential inter-linkages with these events are too numerous and variable to be represented.

The sixth framework refers specifically to the evaluation of "general funds", very briefly. The author acknowledges the difficulties here, associated with the various ways in such funds can operate, but does not link his analysis back to similar problems noted earlier with programme evaluations. The seventh framework incorporates community development: attempts to work with systems of locally related institutions. Here the community basis of judgements is not only relevant to assessing final impacts, but also to the choice of goals and methods of the various organisations. The originally linear structure of the first framework diagram is now looking more like a network of relationships. Here any evaluation will have to take more account of the different "underlying rationalities", ethnographic and participatory approaches become more relevant. The original precise terminology in the first framework has been modified to allow more ambiguity. Reflection processes rather than planning exercises are emphasised . We are at the opposite end of continuum from the Logical Framework. An area where methodological innovation is needed.

The third chapter on managing evaluations is more mixed and incomplete, and arguably out of sequence. An abstract analysis of types of evaluation processes is followed by some other observed distinctions between evaluations in practice which don't go beyond taxonomy. The outline of desirable contents for evaluation reports that follows may be of value to new practitioners, at best. This chapter ends with the surprisingly uncritical reflection that while "evaluations are rarely operationalised and integrated into development programmes..." this fact "is an indicator of deficiency in the respective organisations rather than the evaluation system" (p.103)

The final chapter is an overview of methods of evaluation, which appears to be straying away from the earlier objective of developing frameworks. The main topics include sources of information, qualitative and quantitative inquiry, indicators and sampling. The emphasis on the distinction between qualitative and quantitative methods is questionable, since there is no clear relationship to the different frameworks outlined in chapter two. This all-too-familiar polarisation may do more harm than good, separating that which needs in practice to be integrated. Participatory approaches have a wide range of applicability, and should not necessarily exclude the use of quantitative as well as qualitative data. This final chapter ends with residual but lengthy notes on indicators and sampling, suggesting a loss of direction.

This book does raise some important issues that need to be pursued with more focus in the future: how to represent and evaluate complex development initiatives. In particular, programmes of projects whose contents and objectives have evolved over time, and community development projects with contending actors and objectives. Recognising the networked nature of interests in these events is one thing. Constructing a single evaluation document that makes sense, and is of value to most of those interests, is another.

Rick Davies (Dr), Social Development Consultant and Research Fellow, Centre for Development Studies, University of Wales-Swansea, rick@shimbir.demon.co.uk , www.swan.ac.uk/cds/rd/rd1.htm