Conclusions

The evidence obtained by my colleagues and I does not support the concept that Scirpophaga incertulas larvae feeding in the massive stem lumens of elongated deepwater rice cause significant damage and consequently seriously affect crop yield. At this stage, S. incertulas probably approaches a near perfect parasite-host relationship in many deepwater rice varieties. The large hollow stems provide a safe stable environment for the larvae which can feed on the parenchyma tissue without damaging vascular tissue.

It is clear that the deepwater rice crop reaches a peak plant population level, measured by tiller numbers, at around the time of the onset of flooding. The peak is reached, in mid-May to early June, as a combination of favourable levels of temperature and moisture (the pre-monsoon rains) and of relatively low plant heights, with little competition between tillers for light. The subsequent decline in plant population is primarily a natural phenomenon inherent in the biology of the crop itself and takes place irrespective of the numbers of S. incertulas larvae present.

Tiller numbers always decline because of normal competition between plants, and in deepwater rice the optimum at maturity seems to be around 150 tillers/m². When higher numbers of panicles are attained the panicle size is lower and no improvement in yield is achieved. The rotting of the submerged parts of elongated stems is a normal feature which does not affect nutrient uptake (nodal roots are a feature of deepwater rice) and does not affect crop anchorage (the stem is fibrous); additionally, even if the stems do break free from the lower parts, yields are little affected (Taylor, 1988).

There is evidence that when larval numbers of S. incertulas are high in the period of early flooding the decline in tiller numbers may be accelerated, but the emphasis should be on the term 'accelerated'. In view of all the foregoing, there seem to be good grounds for the argument that in the elongation phase the relation between S. incertulas and deepwater rice is that of a near perfect host-parasite relationship. Feeding activity of the borer larvae within elongated stems does not cause damage which can be translated into crop loss. This view gains support from Litsinger (1991), who reviewed most aspects of yield loss and crop loss in rice. He dealt mainly with conventional paddy rice and did not have much on either deepwater rice or yellow stem borer, but it is significant that he remarked 'the principal effect of early stem-borer attack is to kill tillers, but as these are produced in excess, considerable compensation by the crop is possible', and, also, 'there has been a tendency to equate crop injury with crop loss or to base crop loss assessments on untested assumptions, making no allowance for compensatory growth'.

The counter argument that 'larval feeding equals damage' stems from Catling (1979), who found from surveys of 320 farmers' fields in 1977 and 1978 that 60% of all fields had outbreaks at harvest. This was confirmed by Catling, Islam & Pattrasudhi (1987), who further stated - 'it is generally accepted that YSB is the major pest of DWR in the region because it consistently causes heavy stem damage every year (Catling, Islam and Pattrasudhi, 1984/5)' and concluded 'but despite many attempts, there is no control measure in sight for the severest pest of them all - YSB'. Catling & Islam (1995) note that their work in 1977- 1980 led to 'a series of further investigations over the last 15 years which were reviewed by Catling (1992)'. In that review, it was claimed that the best results in Bangladesh were obtained with insecticides timed to control Brood 5 moths and applied with a motorized knapsack sprayer from a boat. This was credited to Islam, Catling & Pojananuwong (1988) and it has since been stated that use was made of 'a tentative action threshold of 10% damaged stems at the booting to flowering stage' (Catling & Islam 1995).

However, Catling (1992) made no reference to the detailed and exhaustive crop loss studies in Bangladesh during 1981 and 1982 (Taylor, Alam & Razzaque 1982; Taylor & Islam 1984; Taylor 1984, 1988). In fact, it was the population dynamics model developed in late 1981 (Taylor, Alam & Razzaque, 1982) which underlay the innovative approach in the 1982 experiment. In that, the time of spraying was determined by an upsurge in the numbers of female S. incertulas detected by light-traps (i.e. the emergence of Brood 4 adults) and the experiment did not involve any so-called 'action threshold'.

To suggest that a threshold could be determined by stem dissection is misleading because most conventional insecticides have a contact action and thus are ineffective against internal parasites. Spraying, therefore, will not reduce any existing infestation. The level of stem infestation is a measure of past events and is not a reliable indicator of further attack, particularly as S. incertulas is a relatively strong flier and can be an invasive pest. Light traps can reveal that females are emerging or arriving and thus a spray directed at hatching larvae (from egg masses laid on the leaf surface) can be a technically effective control measure. This was demonstrated in 1982 and reported by Taylor & Islam (1984).

The results of the evaluations of the possibility for reducing crop loss in deepwater rice due to S. incertulas by use of insecticides clearly indicate the importance of the mid- to late-season brood(s) of the insect. Although the logistic difficulties of insecticide application to the deepwater rice fields at peak flooding, when the water is up to 3m or more deep, are considerable, they would not be insuperable. Selection of the insecticide would have to take into account the need to prevent loss of the fish populations.

Taking the combination of all the evaluated trials, detailed observations were made of 18 deepwater rice varieties and many others were examined in resistance studies and crop-cuts. The trials also encompassed three methods of crop establishment (broadcast dry seeding, line sowing and transplanting), variations in seed rate and spacing, several different forms, application rates and timings of fertilizer, and varying seed treatments. The four floodplains of Bangladesh have widely differing soil characteristics, largely due to major differences in the water quality and silt content of the rivers which flood over them. Thus, the crop environment is extremely complex, and consideration of any one crop factor in isolation of knowledge of the others is inadvisable.

Finally, if correct, the status of S. incertulas as a successful parasite of deepwater rice has a broader significance for students of evolutionary ecology, Chowdhury & Zaman (1970) considered the deep-water rice varieties of O. sativa as being descendants of the most prevalent wild rice species of Bangladesh, O. sativa var. fatua. The good capacity of deepwater rice for ratooning, that is continued generation of tillers after the harvest of the first crop, also indicates an affinity with the ancestral, perennial, rayada varieties of floodplain rice (this was shown in an investigation which I instigated, as an MSc project for N.B. Thompson, in 1983; the work was completed and reported by Bene 1988). Ratooning could also enable the plant population to withstand high grain sterility caused by the larvae of the final S. incertulas brood of the normal deepwater rice season. Thus, the grain loss may be a major problem for the farmer but it is not necessarily devastating for the deepwater rice. In a much earlier example of the pest problems which have faced the high-yield varieties of the 'Green Revolution', the historical selection of paddy rice varieties with shorter, narrower stems could have led to inherently greater damage to vascular tissues by larval feeding, and so led to deadhearts throughout the growth of the crop.

The following general conclusions, although derived primarily from work in Bangladesh, would seem to be applicable to all the deepwater rice areas of South and South-East Asia.

1. At the beginning of the season (April-May), deepwater rice can be regarded as being the same as dryland rice. Deadheart incidence will give a reasonable indication of stem-borer infestation and, if unusually severe, the loss of primary tillers could lead to crop loss. In Bangladesh at least, this is not a common problem.
2. If there are substantial pre-monsoon rains and moderate temperatures in late May to early July, S. incertulas larvae will survive in sufficient numbers to give a noticeable second brood before the onset of flooding.
3. The onset of flooding leads to rapid plant growth and then to stem elongation. A third brood of S. incertulas, or an extended second brood, may be observed during this period (mid-July to late August). At this time, S. incertulas may become numerous because of in-migration following emergence from conventional summer wetland rice (transplanted aman).
4. Light-trap results in the peak flood period, from August to mid-September, can be expected to show a potential S. incertulas population explosion in deepwater rice. The abundant larvae entering the maturing crop may well penetrate the terminal internodes, leading to the formation of whiteheads and total grain loss from the affected panicles. This is the major source of crop loss in deepwater rice due to S. incertulas. Reduction of this population by spraying at this time could give a significant saving in yield of some 10% or more.

END

href="\dwr\ysb14.htm"