The establishment of weeds through contaminated rice seed has long been overlooked by farmers. During harvest, weed seeds similar in maturity and size are harvested and processed with rice seeds, thereby contaminating them. As a result, several major rice weeds are inadvertently disseminated.
A 1985-1986 study conducted by the International Rice Research Institute in Guimba, Philippines, found that samples of farmers' processed rice seed were contaminated with various weed seeds. The present study was conducted in the same area to examine farmers' practices regarding seed selection and processing.
Despite these efforts, 97% of the samples were contaminated with weed seeds of Echinochloa glabrescens (79% of the total samples), Ischaemem rugosum (71% of total samples), Echinochloa crus-galli (L.) P. Beauv. (30% of samples) and F. miliacea (26% of samples). Among samples from irrigated vs rainfed systems, there was no difference in weed seed content except for sedge seeds which were found in significantly higher amounts in the rainfed systems. Of four criteria used to select farmers, only seed source (owned vs. exchanged) was highly significant.
Sixty to seventy-five percent of the farmers thought that the Echinochloa species was disseminated via contaminated seed rather than through the soil. Most (91%) thought that I. rugosum was soil- rather than seed-borne. Farmers' lack of awareness that I. rugosum is a seed contaminant may have contributed to farmers being less careful in eliminating it from their rice seed.
Farmers' rice seed samples were cleaner than those obtained in 1985-86, but were similar in terms of level of contamination by E. glabrescens and I. rugosum. The 1991 study sampled grain that was selected, processed and stored by farmers for use as seed. Farmers' seed selection and processing practices appear to be relatively effective in producing seed which is as clean as that in marketing stations.
The mean seed count for Echinochloa spp. of 87 seed/kg translates into 7,000 plants/ha (assuming 100 kg rice seed/ha and 80% germination). This does not necessarily represent a substantial problem in transplanted fields which normally have 250,000 hills/ha, but transplanted Echinochloa spp. outgrows rice seedlings, and appears on rice hills where removal is difficult and competition is high. The transplanted weeds' chances of reproduction are enhanced, which can exacerbate problems in subsequent crops.
Weed seed contamination clearly needs to be reduced throughout the rice-growing world. Needed steps include:
Similar studies need to be conducted in countries without well-established seed certification systems if technologies to reduce weed seed contamination are to be developed. Not only do farmers, seed growers and seed merchants need to address the problem of weed seed contamination in rice seed, scientists and national policy makers need to be aware of the problem to direct wider efforts toward solutions.
Social Sciences and Agronomy-Plant Physiology Divisions
P.O. Box 933, Manila, PHILIPPINES