«: AGROCHEMICALS: FATE IN FOOD AND THE ENVIRONMENT PROCEEDINGS OF A SYMPOSIUM, ROME, 7 - 1 1 JUNE 1982 JOINTLY ORGANIZED BY IAEA AND FAO l^J I N T E R ...»
that is, their fate in food and the environment. The objective of this paper is to set this specific area of concern within the wider context of the economic decisions on pest control facing farmers, governments and chemical suppliers; decisions which have to be taken against a background of rapidly increasing world populations, dominated in absolute number and in percentage growth rate by the developing nations.
* This paper was presented on behalf of Groupement international des associations nationales de fabricants de produits agrochimiques (GIFAP).
2. WORLD BACKGROUND Sixty seven per cent of the world's 4.5 bn population currently lives in that group of countries encompassed with the UN category of "developing countries" (footnote ). The average rate of population growth within these countries is 2.2% pa compared with 0.7% pa in the developed nations. The high proportion of the population in these countries at or below child bearing age ensures that high growth rates will continue, resulting in 79% of the world's projected 6.2 bn population living in the developing countries by the year 2000. In contrast, these nations produce only 48% of the world's total tonnage of cereal grains and a smaller proportion of its other more varied fruit and vegetable crops and animal products, resulting in their producing only 38% of the
•world's total agricultural production.
The future demands for food to maintain the living standards of the expanding population, and, in addition, allow improvements in quality of nutrition to reduce the gap between the current average consumption of 2 200 к cal per person per day in the developing nations compared with 3 300 к cal per head per day in the developed countries, and provide a more varied diet, will require large increases in agricultural production. These increases in production must, for political, foreign exchange and other economic and social reasons, occur within the developing nations themselves. Sustained growth rates of at least 2.8% pa are needed; more rapid rates if standards of nutrition are to improve markedly and food deficits be avoided.
A 2.8% pa average rate of production growth has been achieved in the developing world over the twenty years 1960 to 1980, largely due to the introduction of newer high-yielding varieties, increased fertilizer use and other improved management practices. This has been an historically unprecedented rate of growth, yet an even greater achievement is required of the future. FAO suggests a target rate of 3.7% Ref. .
Production increases will derive to a limited extent from expansion in area cropped although the potential for this is now very limited in many of the developing countries.
Table 1. PROJECTED CONTRIBUTION OF AREA EXPANSION AND CROP
INTENSIFICATION TO AGRICULTURAL PRODUCTION GROWTH INDEVELOPING COUNTRIES - 1975-2000 (Source: FAO) (% contribution to output growth)
The majority of the increase must come from increases in productivity per hectare ; productivity improvements resulting from increased use of irrigation, better cultivars, increased fertilizer use, improved husbandry practices and the increased use of crop protection.
In many parts of Asia, for example, expansion in yield per hectare became the main source of rice production growth from the late 1960s (accounting for more than 80% of extra production) and little growth came from area expansion.
FAO  project that yield increases and cropping intensity will account for 74% of production growth in developing countries over the period 1975-2000, with even greater dépendance on yield performance in Asia (Table 1).
By the year 2000 almost two-thirds of the developing countries population will live in countries with at least 95% of their total potential arable land under the plough.
3. GROWTH OF AGRICULTURAL CHEMICAL USEHistorically the use of agricultural chemicals has been a major part of the package of technical inputs which have brought about the growth of agricultural productivity in LEVER the developed world, and these trends are being repeated as less advanced economies grow- Pest pressures, diversity of insects, and intensity of weed growth are greater in tropical than temperate climates, heightening the need for pest control measures as part of modern production systems.
The overall growth of chemical usage in agriculture seems, therefore, an inexorable part of productivity growth to combat the losses caused by weeds, insects and diseases. These losses are estimated to represent, on average, 30-40% of potential production  and in many situations total crop devastation.
It is important that the overall pattern of pesticide use reflects an amalgam of good, rather than poor, decisions on product use by farmers and governments. An important part of that decision-making process is the assessment of costs, risks and benefits as they relate to each particular situation. It is not possible to generalise about the economic merits of chemical use, but it is possible to display a number of principles which should be considered.
Economic considerations embrace financial, social and a range of environmental issues.
4. FINANCIAL COSTS AND BENEFITS FROM PESTICIDE USEThe financial rewards from pesticide use may accrue directly from the reduction of losses caused by pest or disease attack, may be derived from savings in production cost (for example in reducing labour costs for weeding), or may result from a change in crop management system which would not have been possible without chemicals.
Numerous examples of direct financial reward from pesticide use, or losses through their lack of use, have been reported in the literature and a few are quoted here
(a) The average losses of grain in storage in Tanzania per year are estimated to be about 100 000 t (30% of the total stored quantity) worth sterling £4.5 m.
Cost of treatment to prevent this damage would be about £100 000/y, giving a benefit to cost ratio of 45:1 (ICI Estimates).
IAEA-SM-2229/40 (b) In Nigeria in the mid-1970s it was estimated that 50% of stored cowpeas were holed and 25% of their weight lost by Bruchid beetle attack .
(c) Fifty per cent of the protein of smoke-dried fish from Lake Chad was destroyed each year by leather beetles (Dermestes maculatus) during the 6 months between drying and sale in Nigerian cities .
(d) Fifty experiments over 14 years at IRRI showed that pesticide protected plots yielded 5.7 t/ha compared with 2.9 t/ha on unprotected plots (an additional
2.8 t/ha worth about US $630/ha at 1978 local rice prices to the grower). One hundred and thirty experiments in farmers' fields in the Philippines showed a 25% increase in yield where pesticides were used, and similar results have been generated in .
other Asian countries (e) Sugar-cane production in Pakistan was increased by 30% through improved insect control. A $77 000 expenditure on insecticides gave $7.2 m additional sugar; a benefit:cost ratio of с 100:1 .
(f) Yields of pyrethroid-treated cotton plots in 28 US University trials between 1976 and 1978 were 2.5 times greater than those on the untreated plots .
This corresponded to an increase in crop value of $1 780/ha for an expenditure of $87/ha; a benefit:cost ratio of 20:1.
In plantation crops, where the financial pressures on the business necessitate careful costing of inputs, and where labour management problems demand that available labour is employed on the most rewarding tasks, herbicide use has been demonstrated to result in large savings in cost. For example Yeoh et al.  report a reduction in weeding costs in a rubber nursery from M $1 006/ha using manual labour to M $156/ha using herbicides, with additional benefits from reducing injury to the roots and stems of seedlings caused by unskilled labour.
Such large benefits will not always result. Major estates continually review the relative costs of chemical and manual weed control, and the other labour management problems, in evolving their weed control programmes. The optimum system will often integrate chemical and mechanical techniques.
LEVER In arable crops the economics of herbicide use will depend upon labour costs, labour availability at the critical stages of crop growth, the physical problems connected with alternative methods of weed control, the levels of use of other inputs such as fertilizer, and the value of the crop per hectare.
In the lowest labour cost areas of South-East Asia or Africa, mechanical or manual weed control will often be the most economic approach. In many areas of transplanted paddy rice the workload for weeding will be contained within manageable proportions by good water management. However it has been estimated that, on average, the tropical farmer spends 60-70% of his working time controlling weeds . As labour costs rise, so there is pressure to substitute labour-saving herbicides.
There has been a noticeable transition from manual control through integrated weed control using few sprays of cheap herbicide, eg 2,4-D plus supplementary manual weeding; through wider use of cheap herbicides; finally moving to the more expensive,broader spectrum, herbicides used in, for example, Japan or USA.
The first pressure to use herbicides is likely to come to relieve peak weeding labour inputs. There will be strong inter-farm or even on-farm competition for labour at peak times and loss of yield results from poor timeliness in weed control.
Improved mechanical weed control methods can replace labour in many row crops (eg maize), but in densely planted crops such as rice, effective mechanical techniques are not available. For example, in upland rice, where weed growth is much more intense than in flooded paddies, labour inputs to provide one hand weeding can be as high as 350-600 man-hours/ha. A number of studies reported by De Datta et al. (see Ref. ) showed that early competition from weeds reduced upland rice yields by at least 50%, and totally uncontrolled weeds would decimate the crop. Broad-spectrum rice herbicides provide the only practicable solution.
The importance of good weed control increases with adoption of other improved management practices of high yielding (often shorter strawed) varieties and increased levels of fertilizer, which stimulates weed growth as well as crop growth. A number of studies in Philippine and Thai rice  show small increases in yield from additional fertilizer under conditions of poor weed IAEA-SM-263/41
Table 2. INTERACTION BETWEEN THE EFFECT OF FERTILIZER USE ANDWEED CONTROL ON YIELD AND NET INCOME (Laguna, Philippines, 1970 wet season and Don Chedi, Supan Buri, Thailand, 1971 wet season) (Source: Ref. )
Source : Rockwood and Lai  control, or, in some situations from added weed control without added fertilizer. The combination of increased fertilizer and better weed control can be very productive (Table 2).
High crop values also stimulate high input use. High levels of fertilizer are used because the marginal quintal of grain is sufficient to justify the marginal kg of fertilizer at the top of a crop's dose response curve.
High quality weed and pest control is required to protect the investment in yield.
Rewards may also be attained through beneficial changes in the whole production system. Diverse examples of beneficial changes are available. Rockwood and Lai , at the International Institute of Tropical Agriculture, Nigeria, investigated the use of herbicides in place of mechanical weed control as a means of establishing a surface mulch and minimising soil disturbance and erosion. They demonstrated a wide range of benefits from a no-tillage system including reduced soil temperatures IAEA-SM-263/40 (Table 3) (resulting in better seedling germination, emergence and vigour), better soil moisture conditions (Table 3) (through a dramatic reduction in rainfall runoff and evaporative losses), reduced soil erosion (Table
4) and decreased weed seed germination.
Further studies have led to the establishment of minimal or zero cultivation systems in many parts of Latin America. However, the economics of this herbicide-based change in husbandry systems vary significantly between locations. Initial International Institute of Tropical Agriculture (IITA) work  showed that on good soils with high fertility levels maize yields were similar for conventional and zero cultivation. The economic reward was to be gained through long-run benefits of soil conservation and reduced weeding costs rather than immediate yields benefits. Longer run trials at IITA , however, showed that, over a run of years, financial benefits began to appear as maize yields on conventionally cultivated plots began to fall in comparison with zero cultivated plots. On poor soils depleted by excessive cropping, there was some yield decrease with zero cultivation compared with conventional cultivation. Correct fertilizer usage is therefore important to the use of the technique and its usage can only be developed on suitable soils, with suitable attendant management practices, and where the economic benefits from erosion control are sufficiently great.
A major area of acceptance has been for large-scale soybean production in Brazil where potential benefits are perceived by government, farmers and chemical and machinery companies and each has been prepared to make the investment in a co-ordinated programme of research to establish a zero-cultivation programme. In southern Brazil the natural vegetation has been cleared to make way for soy bean and coffee. In Parana, the forest cover has been reduced from 84% of the land area to 9% over the last 30 years. The double cropping soybean/wheat rotation, has required two cultivations per year. The major soil types, latosols and podsols,are seriously prone to erosion, exacerbated by rainfall of 1 200-5 000 mm/year and intensities of 30-40 mm/hr. Even with terracing (which reduces soil loss from as much as 400 t/ha per year on a 6% slope to 97 t/ha per year) soil losses are too high to maintain soil fertility and make permanent land use possible .