This paper places the participatory approaches to research presented in the six following papers into the broader debate on priorities in agricultural research and on the role of specific approaches to research. It reviews the principal features of the approaches outlined, the extent to which they have been incorporated into existing institutions such as national agricultural research services, and their implications for support services. Key issues which future participatory research should address are highlighted.

A framework is proposed for the classification of farm systems, which are defined as decisionmaking units comprising farm household, cropping and livestock systems that transform land, capital and labour into products for consumption and sale. Two general principles underlying the classification are outlined. First, since farm systems are embedded in a hierarchical structure, the classification is based on the characteristics of the underlying systems and their interactions. Secondly, ecological factors, i.e. physical and biological parameters, are the primary determinants of farm systems. Changes in farm systems, at least in the foreseeable future, depend on the development of socio-economic variables. The classification is summarized in a set of comprehensive tables.

L. O. Fresco y E. Westphal: Una clasificación jerárquica de sistemas agrícolas

The principal contribution that the farming systems perspective brings to agricultural research is a farmer and problem-solving orientation. Development of this orientation helps strengthen linkages in national research systems between commodity programmes and between disciplines for both applied and adaptive research. It can also strengthen linkages between research and extension and between research and policy analysis. Issues in strengthening each of these linkages are briefly reviewed with particular emphasis on the socio-economic dimensions and the potential role of social scientists. It is argued that the farming systems perspective in agricultural research should be seen as performing an integrative role in research systems. This can often be developed without significant institutional reorganization.

Four methods of measuring crop production are reviewed in the context of different survey objectives. A popular technique, crop cutting, tends to overestimate yields and does not produce good estimates of individual plots. For high accuracy, harvest of the whole plot is the best method. If statistics of regional production are required output can be sampled after harvest, or, if the farmer harvests in consistent units, his own estimate can be taken. Limited evidence shows that farmers' estimates may be no more biased than crop cutting, but require fewer resources and supervision. There is no best method. The method used must be chosen for the purpose of the study. Whichever method is chosen a distinction should be made between biological and economic yield and correction must be made for threshing and moisture content.

The study of livestock in farming systems research presents several challenges. Factors such as the small number of observations, high statistical variability in measurements, and the difficulty in valuing non-market inputs and outputs make it difficult to evaluate on-farm research. Regular Research Field Hearings (RRFH) can improve communication between scientists and farmers, thereby increasing mutual understanding, willingness to collaborate in evaluation, potential technologies, and also improving test results. Of three groups of sheep and goat farmers involved in on-farm testing in northeast Brazil, the group which participated in RRFH experienced a significantly higher level of success with the technology as measured according to performance criteria such as weight gains and farmers' perceptions of their animals' performance. The results confirm the importance of farmers' participation in agricultural research.

This is the first of several papers that will present ideas on statistical method in agricultural research, especially in relation to the planning and analysis of experiments, based upon experience but not found explicitly discussed in standard textbooks. Indeed, some of the contents of these papers may appear to conflict with what can be found in books: closer examination is likely to disclose lack of real disagreement with the better texts. The first paper outlines a plan for the series. Thereafter, it concentrates on relations between agricultural scientists and statisticians and on some of the issues that are too easily overlooked when one of the former, often for good reason, chooses to be his own statistician.

As the art of plant breeding became a professional and increasingly expensive activity, legal protection was gradually provided in member states of UPOV (International Union for the Protection of Plant Varieties), at present 17 in number. The requirements and scope for protection by Plant Breeders' Rights (PBR) are reviewed and are considered to be favourable for plant breeders, growers and national agricultural productivity. Advantages and disadvantages of PBR and patent protection in connection with biotechnology, the rôle of multinationals as plant breeders and aspects of genetic erosion are discussed.

That experimental design is fundamentally more important than the formal analysis of variance of results is emphasized. Attention is drawn to the many weaknesses of split plot designs, and the case for adopting other types of confounding whenever possible is stated. Split plots are sometimes necessary but rarely desirable. The need for randomization is stressed, with a warning against abandoning this principle before fully considering the dangers. The error in analysing repeated observations as though they are from split plots is discussed at length. After brief comment on variance heterogeneity, a final section looks at how measures of the precision of treatment means should be presented. The two main points made are that standard errors of treatment means should always be included in tables of results, and that existing multiple range significance test procedures have little or no practical utility.

In India, sorghum is grown during the post-rainy season on Vertisols with stored soil moisture. As water stress towards the end of the growing season is believed to be a major cause of poor yields in these crops, an experiment was carried out involving five sorghum genotypes and four sowing dates with and without irrigation.

Supplementary irrigation with 40 mm of water in the earliest sown crop increased grain yields by 24 and 26% in 1979 and 1982, respectively, while irrigation with a total of 120mm water in the latest sown crop increased yields by 87 and 60% compared with the corresponding dryland crops. Early sowings always resulted in better stover and grain yields irrespective of irrigation treatment. Grain yields of hybrids were similar to those of the cultivars but their stover yields were less. The post-rainy season cultivar SPY 86 gave the best grain yield across sowing dates. The widely grown post-rainy season cultivar M 35–1 yielded less but produced more stover and its yields were more stable. The importance of the stability of grain and stover yields and of evaluating the materials for more than one date of sowing in selecting sorghums for the post-rainy season is discussed.

Six sugarcane varieties grown alone and in mixtures containing equal proportions of two or three varieties were evaluated for the economically important traits of the first crop and subsequent ratoon crops over a three year period. An increase in the yield of cane and sugar of up to 25% more than the average for the pure components was observed in some mixtures. The number of millable canes, sucrose content and cane yield in the mixtures was more stable than in the pure varieties. The advantages of growing varietal mixtures of sugarcane are discussed.

The Adaptive Research Planning Team in Luapula Province (ARPT-LP) is one of nine provincial teams in Zambia conducting farming systems research within the Research Branch of the Ministry of Agriculture and Water Development. ARPT-LP has been relatively successful in developing a partnership with farmers, involving them in both the planning and implementation of the team's research programme. This paper examines the formal and informal occasions on which farmers have been involved in the team's activities, with examples to illustrate how farmers' opinions have influenced decisions made by ARPT-LP. The factors which have influenced the frequency and quality of the scientists' interaction with farmers are discussed.

The line-source sprinkler irrigation system provides a continuously variable water application rate, which depends on distance from the line-source. The system is simple to set up and minimizes the amount of land required for experimental work. As the irrigation treatments are allocated systematically, the assumptions of analysis of variance are not satisfied. It is proposed that the effects of irrigation treatments be assessed using analysis of covariance, with distance from the sprinkler line as covariate, thus adjusting for a linear fertility trend. This method of analysis provides an approximate residual mean square for the fitting of response curves, but could be vulnerable to a quadratic fertility trend.

Sixteen years' yield and growth data from an oil palm planting experiment, comparing 56, 110, 148 and 186 palms ha−1, and additional records from a progeny experiment, were used to study the effect of palm age and planting density on carbohydrates incorporated in total above-ground dry matter production per palm (TDMc) and its components: fruit bunch yield (Yc) and vegetative growth (VDMc). The canopy efficiency (e), the amount of carbohydrate incorporated in dry matter production per unit of absorbed radiation per hectare, decreased from the fifth to eighth year then levelled off, but increased once crown expansion was completed. The decrease in e paralleled the expansion of the intercepting leaf surface, and might therefore be linked to an increase in maintenance respiration losses. These losses were apparently not compensated for by an increase in photosynthetic production. The subsequent increase in e in older palms appears to have been due to improved light distribution as a result of an increase in light penetration. The effects of changing levels of light interception and distribution on e were more pronounced at higher density, and resulted in marked differences in changes of TDMc with age between densities. Corresponding differences occurred with Yc, so that the optimum density for current yield decreased until 12 to 13 years from planting, then markedly increased. These findings are discussed in relation to the measures which might be used to increase yield per unit area.

A participatory method that promotes farmer involvement in systems-problem research was developed in response to the failure to adopt improved cropping patterns among upland farmers. Techniques to identify systems problems involved group and individual farmer meetings. Problems concerning the cultivation of marginal cogonal (Imperata cylindrica) uplands were identified. An informal random sample survey, guided by topics of inquiry and biological measurements that employed systems analytical tools, was used to obtain a farmers' perception of ‘systems-problems’. Systems diagrams also provided a framework for searching and screening solutions. A sequence of vining legumes was tested in rehabilitating the marginal uplands. For this experiment farmers elaborated hypotheses on control of Imperata, recovery of soil fertility, and reduced labour costs in re-cultivation. Extensive research activity among the farmers indicates the value of this participatory method.

Shoot and root growth and soil water depletion were studied in mungbean and soyabean grown at three plant populations after non-flooded rice (Oryza sativa) during the post-monsoon dry season in the Philippines. The site had a shallow fluctuating water table (1-2 m) but rooting depth was restricted to 0.8 m by a volcanic tuff layer. Soyabean had a longer duration (89 days) than mungbean (68 days) and intercepted more solar radiation, but from 61 days after sowing was severely stressed and accumulated little dry matter. Mungbean avoided severe water stress due to its short duration. Despite this stress, grain yield (1.0 t ha-1) was similar for the two species and soyabean yielded more grain nitrogen but less straw nitrogen than mungbean. Higher plant population achieved by narrower row spacing increased cumulative light interception and both grain and forage yields in both crops.

K. D. Shepherd, P. J. Gregory, T. Woodhead, R. K. Pandey y E. C Magbujos: Crecimiento de la soja (Glycine max) y del frijol mungo ("Vigna radiata) en la estacidn post-monz6n despuis del arroz de secano.

The present structure of plant breeding and seed multiplication in India is highly centralized. Furthermore, only a small number of new varieties is officially released each year. The system therefore appears inappropriate for the requirements of the large proportion of Indian farmers located in risk-prone and highly diverse environments. An alternative strategy is described whose central feature is close matching of the characteristics of farmers' traditional rice varieties with those of advanced breeders' lines. A selection from these lines is then distributed in small quantities for on-farm trials managed by farmers themselves. If the success of these initial efforts is to be sustained, a more decentralized approach to breeding and multiplication will be necessary.

The mineral status of soil, native forage and cattle tissues in the Swaziland middleveld was investigated during the dry and wet seasons of 1985/86. In relation to beef cattle requirements, forage Mg, P, Na, Cu, Zn and crude protein were deficient during both seasons and K was low during the dry season. Cattle serum and liver samples were generally low or deficient in P, Zn, Cu and Mn, while their Fe content was in excess of normal requirements. The forage, liver and serum analyses indicate that cattle in the middleveld are likely to require supplementary Na, K, Cu, Zn, Mn and crude protein, but Ca, Mg and Fe levels appear to be adequate.

Breeding for high and stable yields is a major objective for cassava breeding programmes. Owing to the simultaneous formation of steins and leaves (source) and storage roots (sink), competition for assimilates implies that there is an optimum leaf area index for yield. This has led to the use of the harvest index as a practical criterion for selection. The optimal harvest index and leaf area index are discussed for selection for stressful environments. A successful strategy for developing high yielding varieties for favourable conditions is presented. Recent information has shown that selection for a slightly larger than optimal leaf area index, and hence greater biomass, can lead to stable yield in both favourable and stressful environments. Moreover, yield levels can be increased by selecting for moderate harvest indices (e.g. 0.5-0.65) and increased biomass. Longer leaf life is an integral part of this selection strategy. These concepts are discussed in detail.

James H. Cock y Mabrouk A. El-Sharkawy: Caracteristicas fisiológicas para la selección de la mandioca.

Yield and growth records from an oil palm planting density experiment, comparing 56, 110, 148 and 186 palms ha−1, and a progeny experiment, planted at 115 and 143 palms ha−1, were used to estimate the partitioning of assimilates into those used for structural dry matter (DM) production, and those used for growth and maintenance respiration.

Gross photosynthetic assimilation (A) for closed canopies was estimated from absorbed photosynthetically active radiation (PAR), derived from actual sunshine hours, and the assimilation-light response curve, to be 128 t CH2O ha−1 year−1. A for non-closed canopies was calculated by correcting for the degree of light transmission, which in turn was estimated from recorded leaf area index values (L), i.e. the total leaf area per unit ground area.

Forty-eight percent of gross assimilation was used for DM production, about half of this being lost in growth respiration. The remaining 52% was lost in maintenance respiration. These losses appeared to level off before crown expansion was completed, and since trunk biomass continued to increase, maintenance respiration per unit biomass (R) decreased with age.

An increase in planting density reduced the assimilates available for bunch DM, had little effect on those for vegetative growth, but strongly reduced maintenance respiration and, since biomass was little affected, reduced R. Assimilates for bunch DM ha−1 reached a maximum at L = 5.6.

The observed trends in R as a function of palm age and planting density merit further study.