College of Agriculture, California State University, Chico, California, 95929, USA

Background and objectives
Pesticide use by chilli farmers in Sri Lanka is within recommended limits for frequency and dose published by the Department of Agriculture [1], but use is excessive when measured by pesticide efficacy on insect pests and diseases and on yield. Only when sulfur was applied on 3-7 occasions or when imidacloprid was applied on 2-4 occasions was there a significant reduction in aphid and thrip activity. Mite, whitefly and armyworm populations were unaffected by pesticides, as were intensifies of anthracnose, Cercospora leaf spot, and Choanephora blight. CMV incidence was not affected by reduced aphid intensity [2]. Farmers choose from an array of insecticides to treat chilli. Chlorpyriphos and sulfur are the most commonly used, but some farmers also use monocrotophos, profenofos, pirimiphos-methyl, endosulfan, fenthion, oxydemeton-methyl, carbaryl, prothiofos and dicofol. Fungicides are used less frequently than insecticides. Mancozeb is the fungicide of choice, but use also is made of benomyl, captan, propineb and thiophanate-methyl. Chilli is first treated with pesticides 3-4 weeks after transplanting; that is, chilli is not commonly treated while in the nursery. Over two seasons and 162 fields, mean use frequency was 2.65 treatments/farm and mean quantity applied was 4.67 kg/ha [2]. The present study tested the hypothesis that limited but timely use of pesticides would improve pesticide efficacy and chilli yield in contrast to efficacy and yield obtained by conventional pesticide practices.

Materials and methods
Seventeen farmers agreed to follow my recommendations and were assigned to Group 1. They treated nursery chilli with imidacloprid 1 week before transplanting, made one or two additional foliar applications of imidacloprid during the growing season based on observations made jointly by the farmer and myself, and refrained from using other insecticides. Fungicides were applied only if recommended after determining disease intensity. Twelve farmers followed conventional practices of pesticide use and were assigned to Group 2. From all fields note was taken of: date chilli nursery planted, date chilli transplanted, fertilizers used and dose, pesticide use frequency and dose, irrigation schedule, field size, insect and disease intensities taken every 10-14 days from three plants in each of 10 plots per field, and weight of green chilli from eight picks (standard practice).

Results and conclusions
Group 1 farmers made the same number of pesticide applications (mean 3.17) as Group 2 farmers (mean 3.16), but Group 1 farmers applied 43% less pesticide than Group 2 (mean 3.24 kg/ha Group 1; mean 5.71 kg/ha Group 2). Aphid severity was 88% less among Group 1 farms using imidacloprid on nursery chill and after, in contrast to Group 2 farms that did not treat nursery chilli and used eight insecticides, but not imidacloprid. Interestingly, CMV severity based on symptoms was 42% less on Group 1 than on Group 2 farms. Group 1 farms showed significantly less severity than Group 2 farms for mites, thrips and whitefly. AUDPC values based on anthracnose, and Cercospora leaf spot severities from Group 1 were 11 and 14% less, respectively, than values from Group 2 farms. Group 1 farms used biteranol and chlorothalonil as well as mancozeb, the pesticide of choice by Group 2 farms. There were no group differences for armyworm and Choanephora blight, and intensities were minimal. Mean green chilli yields from Groups 1 and 2 were 8684 and 3769 kg/ha, respectively.

1. Anon., 1994. Agrochemical Recommendations. Ministry of Agricultural Development and Research, Gatambe, Peradeniya, Sri Lanka, Annex 5, p. 12.
2. Burleigh JR et al., 1998. Agriculture, Ecosystems and Environment (in press).