Agronomic findings on stem density changes

A study of factors affecting crop yield cannot be effective unless all aspects are taken into consideration, and entomologists need to pay close attention to the biology of the crop itself. This need for an integrated approach drew considerable support from Trumble, Kolodny-Hirsch & Ting (1993).

In my 1988 paper, I noted that agronomic studies in Bangladesh showed that neither use of differing seed rates (30, 50, 70 and 90 kg/ha) nor use of transplanting to achieve a known high initial plant stand have significant effects on final yield. In 1981 moreover, a very precise transplanted spacing trial was laid down to study the effect of seedling number per hill with a constant seedling population of 200/m². The results showed that despite a very high variation in early tiller populations (626/m² with a 10 x 10 cm spacing and 419/m² with a 20 x 20 cm spacing) the final panicle numbers were similar (234 and 208/m², respectively). The wider spacing had fewer panicles but, nevertheless, gave a higher yield due directly to higher panicle weights (P. Francis, In Research work conducted by the ODA/BRRI Deepwater Rice Pest Management Project. Unpublished reports to the BRRI Internal Review meetings, 1981 and 1982).

A wide range of agronomic studies at Habiganj Deepwater Research Station also have shown that high seed rates may give high early tiller populations and high panicle numbers but often do not give consistently higher yields. For instance, a seed rate of 134 kg/ha gave 404 tillers/m² (at 30 days after sowing) and 194 panicles/m² with yield of 2.44 t/ha, but a seed rate of 68 kg/ha gave 211 tillers/m² and 147 panicles/m² with a yield of 2.52 t/ha (Agronomy Division reports to BRRI internal deepwater rice seminar, unpublished, 1983).

Francis & Griffin (1982) ascribed the high seed rate (92 kg/ha) habitually used by Bangladesh deepwater rice farmers to be due to allowance for considerable loss of the rice plants during raking and hand weeding and for considerable mortality of tillers during flooding. Similar studies in Thailand (Sittiyos et al., 1982) gave identical results. This drop in tiller numbers is not unique to deepwater rice but is characteristic of all rice crops; for example, Tanaka (1976) shows a 50% drop in conventional rice tiller numbers. The factors which affect crop establishment and tiller numbers are so numerous and complex in their interaction that stand reduction should not be ascribed to any single factor. For instance, Hutchings (1983) reviewed natural thinning in plant populations and showed that plant density commonly declines with increasing size of individual plants.

Since my 1988 paper, research by Akita on the effect of plant population density in rice came to light (published in Japanese but reviewed by Squire, 1990). The key and crucial fact is that total above-ground dry-matter, vegetative dry- matter and grain dry-matter all reached a maximum at around 30 plants/m², with grain dry-matter actually declining somewhat from the peak at that population density. As Squire wrote,

'At a higher population, when some mingling occurs, more of the resources are used, but each plant "competes" with its neighbours, and uses a smaller fraction of the resources.'

Other deepwater rice agronomists also have since examined aspects of the plant populations. In Bangladesh, Haque & Hossain (1988) presented various graphs of changes with time, including a decline from 10 tillers/hill at flooding to 4 tillers/hill at 90 days, then remaining unchanged to maturity. Nodal tillers increased from almost nil to 6/hill at 90 days, dropping to 4/hill at maturity but many did not produce panicles. Hoque & Nasiruddin (1988) examined the effect of population density on yield and agronomic traits of deepwater rice under field conditions in both pure and mixed stands. The initial densities were 50, 100, 150, 225 and 350 plants/m². A linear increase in tillering occurred at all densities up to the preflood stage. The increases were 6 tillers/plant (88% increase) for the low density, 4 tillers/plant for the medium density and 1-2 tillers per plant (11.42% increase) for the high density. Optimum initial density for DWR was 100-120 plants/m². Tiller number decreased at peak flooding by 13.6% at low, 21.4% at medium and 20.2% at high density, but they remarked that it was not clear from this study which tillers dred during flooding. They concluded that, 'Although the initial preflood and peak-flood population densities varied significantly, panicle densities showed insignificant differences and stabilised at around 200 in each density at maturity. Therefore, yield differences among densities were insignificant. The indication is that initial densities of 50-350 seedlings/m2 have no effect on yield.'

Jupp & Rahman (1988) studied the seasonal growth and yield of fertilized deepwater rice at two floodplain sites. They examined a wide range of growth components and factors in farmers' fields in 1986. Basal tiller density peaked at the end of June (being highest, >400 tiller/m², with both high levels of fertilizer, 90 kg N/Ha and 60 kg N/ha) but by mid-September (maximum flood) had dropped to around 180 tillers/m², and then there was no further decline, except for a small drop in the unfertilized plot. A late flood resurgence affected all yields and for all treatments these were around 1.7 t/ha. They commented that 'the normal decline in tiller density over the season was also apparent' (Vergara 1985).

In Thailand, Kupkanchanakul et al. (1988) carried out a study of growth patterns in different depths of water (150 and 20 cm). Initial tiller levels at 30 DE (days after emergence) were around 430/m² in both depths, declining steadily until at 170 DE tiller numbers were 167 and 134)m² respectively, and then remaining steady. Grain yield was 2.3 t/ha in the shallow water and 2.1 t/ha in the deep water. Deep water tended to reduce panicle numbers/m², because tiller number was reduced, but had no other effects on yield components. Kupkanchanakul, Vergara & Robles (1988) and Kupkanchanakul, Kupkanchanakul & Roontun (1990) studied the effect of time and frequency of leaf cutting for herbage on grain yield of deepwater rice. Leaves were cut at the collar of the last fully developed leaf at different times after emergence. Comparison with an uncut control showed there was no effect on yield when leaves were cut at 40 and 70 days after emergence. The cutting of leaves at 100 days after emergence, however, led to a significantly higher yield. This the authors attributed to an increased panicle number, possibly due to an observed reduction in plant height.

continue

href="\dwr\ysb12a.htm"