tors and mechanisms involved in this regulation are the subject of
 active research (see Risch et al., 1983; Altieri and Letourneau,
 1982, and references cited therein). For example, paddy rice
 systems of southeast Asia are characterized by high genetic diver-
 sity, which confers at least partial resistance to pest attack.
 Farmers exchange seeds because they observe that any particular
 variety tends to accrue pest problems if grown on the same land
 for several years (King, 1927). In the Andes, farmers grow as
 many as 50 distinct varieties of potatoes in their fields. The
 maintenance of this wide genetic base is adaptive since it reduces
 the threat to crop loss due to pests and pathogens which are
 specific to particular strains of the crop (Brush, 1982). The clear-
 ing of comparatively small plots, typical of shifting cultivation, in
 a matrix of secondary forest vegetation permits easy migration to
 the crops of natural control agents from the surrounding jungle.
 Shade from forest fragments still standing in new fields reduces
 shade-intolerant weeds and provides alternate food and shelter
 for beneficial insects (Matteson et al., 1984).
 These built-in pest suppressive mechanisms are complemented
 by environmental manipulations conducted by farmers as part of
 their farming operations. Thus farmers, in addition to inter-
 cropping and use of resistant varieties, utilize cultural practices
 such as crop rotation, synchronous planting, increased seeding
 rates and changing time of planting (Litsinger et al., 1980). For
 example, Pangasinan farmers in the Philippines planting mung-
 bean after rice, often sow at increased densities and delay plan-
 ting for one or two months to avoid flea beetles during the early
 growth stages. Sowing of cowpeas into standing rice stubble in-
 terferes with host finding by bean flies, thrips and leafhoppers.
 Many farmers also place branches of plants (Glaricidia sepium
 and Cordia dichtoma) within or beside the fields as pest
 repellants (Litsinger et al., 1980).
 In China peasants utilize a variety of cultural practices to con-
 trol diseases in rice and wheat. Stripe rust of wheat is kept under
 control by utilizing local varieties, postponement of sowing
 winter wheat to reduce the chance of autumnal infection, increas-
 ed frequency of irrigation and eradication of wheat ratoons.
Fusarium is reduced in wheat by avoiding the use of fertilizers of
high nitrogen content and proper water management (Chiu and
Chang, 1982).
 In shifting cultivation systems, weeds can be controlled by
farmers provided that weed densities are relatively low and fallow
periods long. Long fallow periods effectively suppress annual
grasses and troublesome perennials. Burning can delay the need
for weeding up to five weeks after planting, while weeding is
recommended within two weeks after planting in unburned crop-
lands (Akobundu, 1980). In tropical Mexico, farmers utilize a
legume cover crop (Stizilobium sp.) in the off-season to smother
weeds (Gliessman, pers. comm.). The adoption of cropping pat-
terns which provide rapid canopy cover minimize weed competi-
tion. Intercropping short season crops such as maize and melon
with longer season crops such as corn and cassava can help prevent
buildup of weed species.
Improving Pest Control Systems
 Procedures for determining appropriate technologies for small
farmers through the farming systems approach (Shaner et al.,
1982) have been adapted to develop insect-control recommenda-
tions by IRRI's scientists (Altieri, 1984). The methodology in-
cludes:
 1. understanding farmers' current perceptions of pests, insect
 control practices and resources available for control;
 2. determining yield losses for each crop growth stage;
 3. matching key pests to measured yields;
 4. selecting appropriate insect-control technology;
 5. testing the technology on farmers' fields over several years;
 and


 6. evaluating the costs and returns of the technology.

48


 So far the methodology has centered around the quantification of
 yield losses for each growth stage of the crop by successively om-
 mitting insecticide protection during each stage, while providing
 control in the others. Results of these trials provide information
 on the correct timing of insecticide applications. It does not pro-
 vide an idea of insect dynamics and damage at various growth
 stages when using farmers' management, thus excluding those
 farmers who wish to maintain their traditional management and
 those who cannot afford purchase of inputs (Altieri, 1984).
 Researchers at IRRI recognize that no matter how strategies of
 chemical pest control are approached, farmers will have to spend
 more money (Litsinger et al., 1980). Given the economic cir-
 cumstances facing developing countries (i.e., external debt,
 transportation costs, international commodity price fluctuations,
 etc.) effective non-chemical means of pest control, both in-
 novative and traditional, should be thoroughly explored and
 preferred: resistant crop varieties, augmentation and conservation
 of natural enemies, cultural control, natural botanical insec-
 ticides, microbial pesticides, etc. (Matteson et al., 1984).


 Management Possibilities
 Polyculture management is basically the design of spatial and
 temporal combinations of crops in an area. There are many pos-
 sible crop arrangements and each can have different effects on
 insect, weed and pathogen populations. For insects, the attrac-
 tiveness of crop habitats in terms of size of field, nature of
 surrounding vegetation, plant densities, height, background col-
 or and texture, crop diversity, weediness, etc. are subject to
 manipulation.
 In intercrop systems, the choice of a tall or short, early- or late-
 maturing, flowering or non-flowering companion crop can
 magnify or decrease the effects on particular pests (Altieri and
 Letourneau, 1982). The inclusion of a crop that bears flowers dur-
 ing most of the growing season can condition the buildup of
 parasitoids, thus improving biological control. Similarly, the in-
 clusion of legumes or other plants supporting populations of
 aphids and other soft-bodied insects that serve as alternate
 pre/hosts can improve survival and reproduction of beneficial in-
 sects in agroecosystems. The presence of a tall associated crop such
 as corn and sorghum may serve as a physical barrier or trap to
 pests invading from outside the field. The inclusion of strongly
 aromatic plants (i.e., onion, garlic, tomato, etc.) can disturb
 mechanisms of orientation to host plants by several pests.
 The date of planting of component crops in relation to each
 other can also affect insect interactions in these systems. An
 associated crop can be planted so that it is at its most attractive
 growth stage at the time of pest immigration or dispersal, divert-
 ing pests from other more susceptible or valuable crops in the
 mixture. Planting of okra to divert flea beetles (Podagria spp.)
 from cotton in Nigeria is a good example (Perrin, 1980). Corn
 planted 30 and 20 days earlier than beans reduced leafhopper
 population on beans by 66% compared with simultaneous plant-
 ing. Fall armyworm damage on corn was reduced by 88% when
 beans were planted 20-40 days earlier than corn, when compared
 to the simultaneously planted intercrop (Altieri and Letourneau,
 1982).
 We still understand little of how spatial arrangements (i.e.,
row spacings) of crops affect pest abundance in intercrops. For ex-
ample, it has been noted that there is a greater reduction in
damage to cowpea flowers by Maruca testulalis in intra-row rather
than inter-row mixtures of maize and cowpea (Matteson et al.,
1984). Selection of proper crop varieties can also magnify insect
suppression effects. In Colombia, lower whorl damage by
Spodopterafrugiperda was observed in corn associated with bush


beans, than in corn mixed with climbing beans. In the same
trials, maize hybrid H-207 seemed to exhibit lower Spodoptera

 PROCEEDINGS of the CARIBBEAN FOOD CROPS SOCIETY-VOL. XX