THRESHOLD DENSITY OF ECHINOCLOA CRUSGALLI (L) BEAW. IN RICE WEED COMPETITION

r Experiments were conducted to determine the threshold Echinocloa crusgaNi (L) Beauv. density in transplanted rice at the Regional Research Statinn, Angunukolapalessa, during the seasons Maha 1984185 and Yale 1985. Thesoil of the experimental site w s Ranna series of Reddish Brown Earths. The data indicated that the threshold E. crusgalli density at season long interspecific competition with rice was five plants per square meter. The decrease in yield at this weed density level was 8 - 17 per cent. The sensitive yield components of rice for the interspecific competi- tion with E. cnrsgalli were the panicle number and the spikelet number per panicle. Projecting the data for broadcast rice the threshold level of E. cnrsgalli density was 6% of the total density. Weed dry weight gave highly significant negative correlation with grain yield and paddy straw weight.


Introduction
One of the major problems that'reduces the rice yields world over is weed competition. Many weed species, vary from area to area and compete with rice in the field. Econo s and statistics of the United States, Department of Agriculture (U.S.D.A 7 eported that worldwide losses from weeds in rice was 33 m. metric tons annually valued at 12 billion dollars. ~a t t a~r e~o r t e d that annual and perennial weed infestation reduced rice grain yields 20-80% depending on the method of establishment and in transplanted rice 11% reduction was observed. In Sri Lanka, both annual and perennial weeds compete with lowland rice reducing the rice grairi yields drastically. Echinocloa crusgalli (1) Beauv. among large number of genera in Echinocloa spp., is one of the-most noxious weeds. Echinocloa colonurn is second in importance.5 -smith1 O reported that in the U.S.A. season long competition from E. crusgalli reduced grain yield by 25-70% and at a density of 11 plants/m2. --reduced main yields inversely to rice stands. chang2 indicated that weed species at d3ferent densities vary in their competitiveness with rice and E. crusgaNi at densities 100-200 ~l a n t s / r n~ reduced rice yields 86-91%, respectively.
The method of land preparation and fertiliser practice also encourages the growth of weed species. Arai and ~atsunakal reported that emergence of E. crusgalli in transplanted rice was greatly reduced, when the field was ploughed 15 -18 cm deep. Smith e t al, showed that pre-planting application of phosphorous stimulated the growth of E. crusgalli. .Kleining and ~o b l e~ indicated at moderate levels of N (140 kg/h) grain yield of rice was'. inversely proportional to the density of E. crusgalli. In the present study, threshold density of E. crusgalli in lowland transplanted rice was determined with recommended rates of fertiliser application.

Materials and Methods
The experiments were conducted at the Regional Research Station Yala experiment, an additional treatment with 25 E. crusgalli plants/m2 was used. Wet bed nurseries of both rices and E. crusgalli were raised and the rice was transplanted at the age of 21 days at a spacing of 20 cm x 20 cm with 3 plants/hill. The required density of E. crusgalli was also transplanted along , I with the rice and was well spaced in the plot. Fertiliser application and pest and disease control was carried out as per recommendations of the Department of ~gi-iculture.~ Plant height and flag leaf length at -harvest were measured from 10 plants selected at random. Tiller count/m2 at 30 days after transplanting (DAT) and at harvest, yield components viz; panicle number/m2 1000 seed weight, percentage filled grain, spikelets/panicle and panicle length from 16 panicles selected at random, were measured from each treatment. Paddy was harvested and Echinocloa straw was separated from paddy straw before threshing and weighed. After threshing and winnowing, grain weight and paddy straw weight were recorded. Any two means followed by the same lerter is not significantly different at 5% probability level. Table 1 shows the results of the Maha 84/85 season experiment. In this experiment, yield component data were not taken except the panicles/m2 and rice grain yields. Rice grain yield decreased with'an increase in density of Echinacloa and was significant when it exceeds 5 plants/m2. Tables 2 and 3. The plant height and flag leaf length showed no significant differences, whereas a significant drop in tiller count was observed at Echinocloa density of 25 plants/m2. Among the other yield components, panicles/m2 indicated a significant drop when the Echinocloa density exceeds 10 plants/m2 ( Table 2). Effect of competition on spikelet number/panicle shows that there were no significant differences among the treatments, but the treatments were significant compared to the control. Rest of the yield compopents do not show significant differences. Grain yield data indicate that weed density exceeding 5 plants/m2 results in a significant drop (Table 3). Further, the rice straw yield also decreased with increase in Echinocloa density.

Results of Yala season are presented in
Correlation analysis of weed straw weight with grain yield and rice straw yield of dry season data shows a highly significant negative correlation, -0.97 and -0.91, respectively. Any two means followed by the same letter is not significantly different a t 5% probability level.

Discussion and Conclusions
Plant height at harvest in both seasons and flag leaf length during dry season showed no significant differences among treatments even though there were slight differences in height with increased Echinocloa density.
~i l l e r s / m~ at 35 DAT indicated that there appears to be no effect ( Table 2) on increase in density of Echinocloa upto 20 plants/m2. However equal number of hills planted to rice and E. crusgalli (25 plants/m2) indicated a 33% drop in tiller number which was significant. On the other hand, panicle number at harvest, which is also the effective tiller number, showed significant differences among treatments and a decreasing trend with'the increase in E. crusgalli density/m2. This also indicates that increased interspecific competition with increased weed density rendered a greater number of tillers ineffective, thus bringin about significant differences in panicle number among treatments. NodaF also indicated that E. crusgalli competition during m-imum tillering reduces the panicle number, while during early ripening reduces grain yield and quality. Data further indicated that the spikelet number/panicle was highly sensitive to interspecific competition rendered by E. crusgalli and gave a significant drop even at a density of 1 weed/m2. Further significant decrease in spikelet number, was not observed even if the E. crusgalli density increased t o 25 plants/m2. The percentage filled grain and 1000 seed weight. did not show significant differences among treatments.
Grain yield in both seasons with season long interspecific weed competition decreased with the increase in E. crusgalli density and was significant when it exceeds 5 plants/m2. At this threshold level, the decrease in grain yield over the zero weed density was 8.33% and 16.9% during Yala and Maha seasons respectively. ~a t t a~ also indicated season long competition by mixed annual and perennial weeds reduced grain yield by 11% in transplanted rice. The threshold level as a percentage of total density (75 rice plants and 5 E. crusgalli plants) was 6.2%. Thus if we were to project this data for broadcast sown rice, we can safely conclude that a significant drop , in grain yield could be expected if the 8. crusgalli density exceeds 6%.
Fwther rice yield reduction, respectively. ~o d a~ also indicated that competition at early ripening stage reduced rice grain yield and quality. ~u b 2 a n a n d vega7 showed that E. crusgalli density of 20 plants/m2 reduced grain yield by 20%.
These data indicaie that E. crusgalli at very low densities renders a very high competition to rice. The highly competitive nature of this weed could be attributed to its physiological superiority, being a plant following C4 photosynthetic pathway. C4 plants normally have a very high growth rate and a high dry matter production, as reflected in these experiments. At the highest weed density level (25 plants/m2), while competing with 3 rice plants at the same hill, gave 5.65 mt/ha dry matter production whereas the control treatment (25 hills/m2 I with 3 rice plants per hill) with only intraspecific competition rice plants gave 4.50 mt/ha ( Table 3).
Correlation analysis showed a very high significant negative correlation between weed dry weight and grain yield. Significant negative correlation between weed dry weight and rice straw indicate an increase in weed dry matter with the increase in weed density and a corresponding decrease in rice straw weight due to increasing intensity of inter-specific competition ( Table 3). ~a t s u n a k a~ indicated that a linear relationship existed between weed density and yield loss, at E. crusgalli densities normally erlcountered in the field. At threshold population level decrease in rice straw weight due t o interspecific competition was 35.3% even though the decrease in grain yield was 8.33%.
In conclusion, the threshold E, crusgalli density with season long interspecific competition was 5 plants/m2 and t h e decrease in yield was 8.33 -16.9%. The major contributory yield components decreasing grain yield with the increase in weed population were panicles/m2 and the spikelet numberfpanicle. Significant negative correlation between weed dry weight with grain yield and rice straw weight was also observed.