However, in the midsixties, Chinese experts were still recommending densities in the range of 2.25-3.75 million seedlings/hectare for rice, citing studies documenting that the point of maximum yields lay within this range." Recent delegations of foreign agricultural scientists continue to be surprised to find densities far higher than are favored elsewhere.22 Since, in response to the Great Leap excesses, the question of optimal density has received careful and extensive study by Chinese agronomists, the difference in technique is likely to be purposeful.

Several reasons can be cited to explain the greater density levels promoted in China: First, in attempting to explain decreases in yield resulting from excessive density, Chinese scientists determined that the major problems were a reduction in the amount of light reaching plants and reduced air circulation which affected the microclimate. Optimum spacing, row arrangements, and fertilizer application timings were found to reduce the impact of these problems, thus permitting higher density levels. Second, Chinese research concluded, contrary to previous doctrine but in accordance with foreign findings, that density should be inversely proportionate to fertilizer use (because, with low fertilizer use, it was necessary to rely on the main stalk and neglect subordinate spikes). However, the levels of fertilizer use considered high enough to justify low density were on the order of 225 kilograms of nitrogen per crop hectare from all sources, far above average application levels anywhere in the world. Hence these findings do not really account for the difference between Chinese and foreign density levels.

The only convincing scientific justification is that dense planting is another corollary of the emphasis on fast crop maturation, which in turn is dictated by the emphasis on multiple-cropping and, in some northern areas, by the short length of the growing season. Under sparse planting, the main and subordinate heads develop at slightly different rates, and the ripening process is thus stretched out. Suppression of tillering and reliance on the main head, through dense planting, contributes to uniform rates of maturation, and reduces the length of the ripening stage.24

It is possible also that the process of seed selection and hybridization, carried out in China with high yield under dense planting as the major objective, has led to the development of seeds with a different response to high density than those developed elsewhere under broader objectives. Should seeds of foreign parentage come along which possess high yields, adequate earliness, and a high tillering rate, this technological "expansion path" may be quickly abandoned. This in fact seems to have happened in the case of 1 male-sterile rice, which is only advantageous because the density (and seed requirement) is one-tenth of the Chinese norm.

Ministry of Agriculture, Shui-tao tsai-p'ei (Peking, 1965), p. 30; T'ang P'ei-sung, "Raising Unit Yields Through the Efficiency of Plant Utilization of Light," JMJP, Nov. 12, 1963.

22 See for example National Academy of Sciences, Plant Studies Delegation, op. cit., p. 45.

23 Ministry of Agirculture, Nung-yeh ch'i-hsiang (Peking, 1963), pp. 24-25.

24 Kuo I-hsien, "On the Principles of High Yields in Paddy from the Perspective of Group Structure," JMJP, Apr. 9, 1963.

CONCLUSIONS

A common strategic objective underlies the entire program of technological change in Chinese agriculture, specifically the increase in the extent of multiple cropping. In comparable environmental circumstances, where other countries are growing a single crop per year, China seeks two; where others grow two, China seeks three. The impact of this goal on the forms and directions of technological change in Chinese agriculture cannot be exaggerated:

Multiple-cropping dictates extreme earliness as an overriding objective of Chinese seed breeding, at a cost of potential yields and ease of borrowing from foreign breeding programs. Multiple-cropping makes an available supply of organic and chemical fertilizers, which is now becoming adequate by the standards of a modern, single-cropping system, inadequate to satisfy the requirements of two or three crops, so that one or no crops can reach optimal yields. It also necessitates the absorption for the foreseeable future of large quantities of labor in low-productivity collection and processing of organic fertilizers, exacerbating the labor-productivity gap between agriculture and industry. Multiple-cropping increases the water requirement in Chinese agriculture, forcing further development of artificial irrigation in areas where rainfall or existing irrigation systems are adequate for only one or two crops, at a significant cost in capital, labor and land encroached on by the irrigation systems. Multiple-cropping creates the bottlenecks in labor and draft animal supply which make mechanization a prerequisite for further intensification, rather than a means of sustaining farm production with a decreased labor force as in other countries. It also forces the continued maintenance of a huge draft animal stock, which reduces grain available for human consumption or meat production. Multiple-cropping, through its requirement of earliness, creates the need for dense planting, whereas other countries have tended to reduce labor and seed production requirements through sparser planting with no loss in yields.

In view of the severe costs of increasing the multiple-cropping rate, one would hope that the benefits clearly outweighed those of alternative strategies. Unfortunately, I have seen no evidence that the Chinese have considered any alternatives at least since the 1950's, even though changing technologies may have made earlier appraisals of limited relevance. By now, the efficacy of multiple-cropping has become a matter of doctrine, at least at the official level.

The most plausible alternative would not have been single-cropping, except in the north, but rather a maintenance of the status quo ante, a system of double-cropping south of the Yangtse River (rice-rice in the warmer areas, winter wheat-rice otherwise), and intensification of production within that constraint. Based on the experience of other countries and the costs and difficulties experienced in changing this system in China, this would have been the natural course of development, in the absence of forceful state intervention. A rigorous comparison of the costs and benefits of the two alternatives, particularly in the light of technologies available outside China, would require quantitative data of a sort not readily obtainable, but the following argument against further increases in the extent of multiple-cropping can be made:

Under conditions of optimal management, geographical units as large as a county today can at best obtain annual yields as high as 10.5 tons per hectare (more commonly, 7.5 tons/ha) using a threecrop wheat-rice-rice system. While this is quite impressive, it represents yields per crop averaging only 3.5 tons per hectare, which is not unusual by the standards of advanced technology abroad (e.g., in Mexico or the Philippines) or in China. At the latter standards, it should be possible to commercially produce 3-4 (or even 5-6) tons per hectare of winter wheat and 7 or more tons per hectare of rice giving, in a double-cropping system, 10-11 or more tons per hectare annual yield. Moreover, the latter would presumably be obtained with equal or lower amounts of inputs, and without creating the severe time bottlenecks of the triple-cropping system. Instead of making a premature switch to triple-cropping, therefore, why not invest the extra labor and capital in better management within the double-cropping system (or single-cropping in the north), with the same potential gain in yields?

If one can imagine a Chinese response to this reasoning, it might be stressed that brigades, communes or counties which switch to a more intensified cropping system typically do so when the potential for further productivity growth under the previous system has been largely played out, and then do so voluntarily. For example, foreign observers have found Chinese officials and scientists enthusiastic about further extension of triple-cropped rice in Kwangtung, but found local cadres and peasants still unwilling to switch. Also, although the adoption of a new pattern may well be "sweetened" by provision of additional or subsidized inputs which insure a local gain in output over the old system, it is not necessary from the social point of view that the new system provide immediate economic advantages, merely that it opens a road to genuine long-run gains. A short-run social loss, if necessary, may be regarded as an investment permitting long-run social gains. In this regard, the case may be mentioned of a model farmer who in 1954 was getting 10.5 tons per hectare from a doublecropping system; today this yield is achieved on large areas in the same region, which has now switched to triple-cropping; model farmers today are getting well over 20 tons/ha on experimental plots under triple-cropping, yields significantly beyond the current commercial potential of the double-cropping system. Once the switch is made, therefore, the process of converting potential into actual gains is stimulated by the new bottlenecks encountered which account for the gap between experimental and commercial yields.

The Maoist optimism or blindness toward risk of this response, however tempered by gradualism, is a striking characteristic of the Chinese implementation of technological change in agriculture, and helps explain both the outstanding successes and failures of their technological policies. It ignores or denies the risk that experimental yields cannot ever be attained economically in commercial production, that the bottlenecks may never be removed, and that premature adoption may lead to serious economic losses before this is understood. However, the failure of the technological policies of the Great Leap was so devastating that ever since the risks have been kept within acceptable levels through the emphasis on experimentalism, gradualism and voluntarism.

Chairman Hua Kuo-feng has been more intimately involved in the promotion of technological change in Chinese agriculture than any other highlevel leader except Chen Yung-kuei, and this involvement has encompassed the failures of the Great Leap technological policies as well as successes in irrigation and seed development work in recent years, yet his optimism about further potential is as striking as the 4-5 percent growth target he recently announced. Although this paper has stressed historical developments, we have also assessed the innovations "in the pipeline" which, together with better exploitation of existing potential and further increases in industrial inputs, must provide the material basis for attaining that target. In seed development, for example, introduction of semi-dwarf wheat seeds incorporating the best foreign genetic stock has been so recent that the high and accelerating growth rate of wheat production over the last decade may continue for some time (wheat output grew 64 percent between 1969 and 1976).25

In the rice region, it now seems very likely that there will be a breakthrough in stagnant yield levels through use of F-1 male-sterile hybrids. Attention has also turned to low-yielding foodgrains such as soybeans which have imposed a drag on production increase, and to the breeding of entirely new crops such as triticale which are intended for low-productivity regions.26 Seed stock will surely benefit from China's new involvement with international seed breeding programs, especially through the introduction of pest-resistant genetic material. And through such techniques as anther culture, use of F-1 hybrids, and rapid stabilization, testing and distribution of new seeds, any potential can be realized within 3 to 4 years after its recognition. On the other hand, the rush to meet yield-increase targets is certain to increase genetic uniformity, and thus the hazard of major crop disasters due to the genetic susceptibilities of the seeds employed.

Fertilizer use will have to continue to grow rapidly to meet the nutrient requirements of the new seeds. It will be met not only from new chemical fertilizer production capacity, but also from the manure which is a byproduct of the planned expansion of pig raising. The replacement of draft animals through mechanization and the evident commitment to more grain-intensive feeding techniques should provide the basis for more rapid growth of pig numbers. The new emphasis on phosphate production should also be mentioned, since it is of benefit both to those red-soil areas in the South and Southwest which have insufficient sources of organic fertilizers and to soybeans in particular, which respond best to that nutrient.27

One may also look to major breakthroughs in irrigation in the next decade. The introduction of sprinkler irrigation on a large-scale trial basis in Szechuan is the first step, and can be expected eventually to narrow the gap between yields in the mountainous areas and those in the plains; those provinces which stand to benefit most are the lagging

25 FBIS Mar. 27, 1978, p. E18; Apr. 10, 1963; and Shantung-sheng Lai-yang nung-yeh hsueh-hsiao, Hsiaomai (Peking, 1975), pp. 1-3. 26 FBIS, Feb. 23, 1978, p. E17; Nov. 22, 1977, p. E6.

27 FBIS, Jan. 31, 1978, p. E19.

interior and southwestern rice producers. A more dramatic development will be the planned realization by 1985 of the dream of irrigating the north China plain with water from the Yangtse. The commitment to production of more earth-moving machinery for use in farmland improvement and irrigation construction will speed up the process of tapping existing potential-in China, as in many another LDC, surface irrigation facilities are chronically underutilized.28

In addition to the forms of mechanization implicit in the programs discussed above, the process of tractorization should be greatly speeded up, with as much emphasis on improved utilization of existing stock and production capacity as on new investment. Yields of foodgrains as a whole have been pulled down by the low yields of the second or third crops in multiple-cropping sequences, which is in turn due partly to delayed planting and poor management attributable to seasonal labor and draft animal shortages. By relaxing these bottlenencks, tractorization makes its most direct contribution to yields. However, the significant of tractors in transport, farmland construction, and in releasing the burden of draft animals on the feedstock should not be minimized.

As this incomplete summary should suggest, China's program for attaining its targeted growth rates is comprehensive and technically (and probably economically) sound; it responds simultaneously to the problem of accelerating the growth of the laggard regions and crops and that of enlarging potential at the technological frontiers. In its ambition and breadth of mobilization, it is comparable to the Great Leap, but, as with the more narrowly based programs of the last 15 years, the potential sources of growth have been correctly identified and are not illusory. The open question, then is not whether the program as stated is sufficient to meet the targets (even if we cannot quantify the potential contribution of each element of the program). Rather, it is whether China can mobilize and organize the resources to carry it

out.

28 FBIS, Feb. 16, 1978, p. E13; Feb. 10, 1978, pp. E8-10