1984, may have been as a result of construction of shelters at the project site in close proximity to experimental plots. The shelters housed numerous nests and the wasps were always present within experimental plots. During the 1981 trials, there were no such shelters and extremely high levels of fruitworm infestation oc- curred. This relationship is to be further investigated. Spodoptera eridania is basically a defoliator but in high density situations attacks fruit, making them totally unmarketable. The black form predominates under such conditions and often ap- pears lodged within cavities in the fruit. As much as 4.9% fruits were destroyed by this pest in untreated plots during the dry season of 1982. Plots treated with B. thuringiensis (16,000 IUP per mg) in the form of Dipel (0.55kg AI/ha), demonstrated 39% less fruit damage. The same treatment reduced infestation levels by 59% during the wet season of that year and when application rates had been increased to 1.1 kg AI/ha in the dry season trials of 1983 and 1984, fruit infestation levels were reduced by 68% and 94% respectively. Like Heliothis, Spodoptera is also preyed upon by P. canadensis and early instars are trapped in silk spun by a variety of spiders. Parasites of Spodoptera include a number of hymenopterans. Keiferia lycopersicella has been observed to infest as many as 99% fruits in Southern California (Elmore, 1937). This seems somewhat astronomic when compared to fruit infestation levels in Trinidad, which in untreated plots were 0.62, 1.89 and 0.81% for the dry seasons of 1982, 1983 and 1984 respectively (Table 3). A wet season trial of 1982 revealed a fruit infestation level as low as 0.22% The pin worm is a pest of great economic importance in many tomato growing regions of the world and as a result of farmers' complaints has been the focus of research in Trinidad for some time. Experiments conducted at St. Augustine (Katwaru, 1978), Chaguaramas (Katwaru, 1979), Macoya (Barrow et al., 1979), Piarco, Aranguez and Santa Cruz (Jones, 1981), all revealed low infestation levels of this pest, damage being confined mainly to the leaves. Though subsequent investigations revealed higher levels of infestation, observations were by no means alarming. Pest populations are definitely much higher under dry conditions, but even during the very dry spell of 1984 (4.35mm/wk) untreated plots revealed just 0.81% fruit damage (Table 3). Leaf damage is considerably higher in Trinidad than is fruit damage (Jones, 1983) but does not appear to affect yields significantly. Thus, pin worms may be of economic significance only when fruits are be- ing penetrated. In 1984, fruit infestation was reduced by 79% when weekly sprays of B. thuringiensis were applied and the same treatment (1.1kg AI/ha) resulted in a 50% reduction of leaf damage. Natural enemies of the pin worm in Trinidad include Euderus sp., Bracon sp., Microchelonus sp. (Barrow et al., 1979), Chelonus sp., Apanteles sp., Agathis sp., (Des Vignes, 1974), an unidentified fungus, Polistes canadensis (Jones, 1981) and Odynerus sp. (Jones, 1984). This rich natural enemy population tends to suggest that in Trinidad, the pin worm ought not to be a serious pest in a well managed tomato ecosystem. During the dry season of 1982, the army worm Agrotis repleta almost totally hindered establishment of vegetable crops, in- cluding tomato, in the Moruga/Penal area where it seems to have developed resistance to all traditional soil pesticides. As many as 12 caterpillars/sq ft land surface were counted at peak infesta- tion. This pest has not been observed to cause serious damage to tomatoes in other areas of Trinidad during the last five years despite its wide distribution. Population densities at Moruga/Penal have been considerably reduced in 1983 and 1984. Manduca quinquemaculata is the most voraciously feeding cater- pillar in the system, one larva demonstrating a feeding capacity of one large fruit or an entire two week-old plant in less than 24 h. Fortunately, population levels are quite low year round, though slightly higher in the wet season. Manduca is mainly a defoliator, but unlike other defoliating caterpillars and beetles, it habitually consumes stems. Should this pest acquire major status, its effects VOL. XX-PROCEEDINGS of the CARIBBEAN FOOD CROPS SOCIETY on yield would be far more drastic than those ot any other pest. Prophylactic measures should include preservation of the wasp Microptilis sp. which parasitizes the larva. White egrets and brown doves are suspect predators of larval stages. Fruit-piercing bugs Nezara viridula and Phthia picta cause fruits to ripen prematurely with corky blotches. Fruits are again of poor quality for the fresh market. In 1983 and 1984, 1.17% and 0.05% fruits in combined treated and untreated plots were damaged by such bugs against which Dipel application is ineffective. The eggs of Nezara are parasitized by a wasp Trisolcus bassalis. A number of Chrysomelid beetles including Epitrix spp., Ceratoma spp. and Diabrotica spp. feed on tomato leaves from the seedling stage. They scrape away layers of leaf tissue creating tiny shot holes, but loss of chlorophyllic tissue is negligible and these insects present no serious threat to tomatoes in Trinidad. Leaf miners, Liriomyza sativae, infest tomato leaves by burrow- ing through mesophyllic tissue. Population density fluctuates, considerably, but is usually higher under dry conditions. Presence of the pest did not necessarily affect tomato yields in California (Johnson et al., 1980) and does not appear to do so in Trinidad either. In commercial areas, particularly in Aranguez, leaf miner populations have been rising since 1982. It has not been established whether this is due to pesticide induction as in the case of methomyl in California, but that probability cannot be ruled out (Johnson et al, 1980). Leaf miners have been observed to be parasitized in Trinidad by yet unidentified wasps. The mole cricket S. vicinus is a seedling pest and has been observed to do serious damage in experimental plots at Piarco under conditions of poor weed control within the surrounding area (1982). Less than 1% seedling damage occurred when good weed control had been practised in 1983 and 1984. In commercial areas, crickets appear to be a serious seedling pest because of poor field sanitation and developing resistance to such insecticides as Malathion, Methamidophos, Aldrin, Chlordane, Cartap Dimethoate, Carbaryl, Decamethrin, and Carbofuran. Spider and russet mites have become serious problems in some commercial and home gardens since 1983. They now persist throughout dry and wet seasons. Feeding on the under sides of leaves, young stems, flowers and fruits, they give rise to chlorosis, crimping, flower and fruit shed, wilting, russeting and other pathological symptoms with resultant low yields. In addition, fruits may be deformed and bronzed, and thus unattractive for the fresh market. Prior to 1983, mites were not reported to be a serious pest of tomatoes and occurrence seemed to be quite sporadic and limited to the dry season, the first showers of rain being sufficient to significantly reduce population levels. Documentation on russet mites as pests of tomatoes in Trinidad is unavailable. Outbreaks have occurred in controlled environment -systems and in open fields since the advent of pyrethroid insec- ticides and the Speedling method of producing seedlings, but mites have not yet been observed in experimental plots. Pest composition of the tomato ecosystem in Trinidad appears to be rather complex and observations suggest that pest outbreaks are induced through improper pest management practices. Dif- ficulty in reducing mole cricket populations may be related to poor field sanitation and pesticide abuse, while pin worm out- breaks in commercial plots seem related to improper application of pesticides. In view of the number of potentially dangerous pests associated with the crop, it is imperative that the system be managed in a manner such that secondary pests are not induced to cause major damage. This necessitates that the natural enemy population be protected and facilitated in functioning optimally. Management Strategies Cultural Practices In controlled environment systems, mites seem to be the only pest of significance. They do damage that is as serious as that done by the entire pest complex in open fields. The hot, dry conditions which prevail in most controlled environ- 177