Biodiversity is described as the variety of living organism that supports the food chain in the natural habitat, and is considered at three levels: genetics, species and ecosystems. It is argued that once lost biodiversity cannot be renewed without preserving the parental organism. The quantum jump in the yield potential through genetic improvement has changed the balance of productivity and profitability of different varieties of rice vis-à-vis their other characteristics such as insect and disease resistance, drought and submergence tolerance and the adaptability to climatic stresses and problem soils. Depending on the value that the farmer attaches to the diverse traits, he/she decides on which variety to grow or drop, and that decision affects the maintenance of cultivar diversity. Since increasing land productivity is a major concern for farmers in land scarce environments, the high-yielding MVs have an edge over traditional land races, which may go out of circulation in the in-situ conservation in farmers' fields. In Sri Lanka, for example, it is reported that the number of rice varieties grown by farmers has dropped from 2000 in 1958 to less than 100 in recent years.

Concerns and cultivar diversity

In Bangladesh, the land base for agricultural production has been shrinking due to diversion of prime agricultural land for housing, expansion of cities and towns, non-crop economic activities (such as brick field and fisheries) and the development of transport infrastructure. According to agricultural censuses, the net cropped area has declined from 8.16 million ha in 1983-84 to 7.19 million ha in 1996, and the average size of farm holding has declined from 0.80 to 0.60 ha over this period. It is expected therefore that the farmers will abandon the low-yielding TVs (traditional variety) in favour of the high-yielding MVs, in order to meet the growing demand for food in the face of population growth. This is a general trend for all land scarce countries. However, it is perhaps pertinent to present available information to assess the extent to which the erosion of cultivar diversity has become a serious concern for Bangladesh.

Official statistical documents do not report area under specific rice varieties; they report area under all modern and traditional varieties lumped together by season. So it is not possible to assess the impact of the spread of MVs on the reduction in the number of varieties grown by farmers. The Department of Agriculture Extension (DAE) conducted a nation-wide sample survey in 1996 and 1997 to estimate area and yield of specific rice varieties grown by the farmers. The sample survey found that the farmers grew 67 MVs covering 61 percent of the total rice area, and about 700 TVs covering remaining 37 percent. The varietal diversity index (VDI) for a particular region of the country (or for the village and the farmer) is defined as one minus the squared sum of the proportional area planted by each variety. The total number of rice varieties grown by farmers must be higher than this number, which is quite large. The VDI is 0.96 at the national level and ranges from 0.73 to 0.96 for the 64 districts. The number of MVs grown by the farmers was greater than the varieties released by BRRI because farmers grow many Indian varieties obtained through informal farmer to farmer exchange of seeds over time. The popular Indian varieties found by the survey were Swarna, Parijat, Ratna, Minikit, CNM25, Saket-4, Jaya and Tusar. IR8, IR20, IR50 and IR5 are among the IRRI bred varieties that still cover significant rice area in Bangladesh.

The concentration of land under a few rice varieties was not found exceptionally high. While five varieties accounted for 38 percent of the total area, it took about 10 and 40 varieties respectively to cover 48 and 70 percent of the land. The ten most popular varieties were found to be BR11 (16.3), BR14 (8.1), BR3 (6.4), Pajam (3.7), Purbachi (3.4), BR10 (3.1), BR1 (2.5), Hashikalmi (1.7), BR16 (1.6), and IR8 (1.2); the figures within parentheses are the percent covered. Traditional rice varieties that still cover more than 40,000 ha are Hashikalmi, Nizershail, Sadamota, Shaita, Digha, Lalmota, Dharial, Depa, Rajashail, Kataktara, Laksmidigha, Tepeboro, Parangi, Kaliboro, Beroi, Laksmilata, Kajalshail, Latishail, and Panishail.

Empiricism on diversity

To assess the impact of the diffusion of modern rice varieties on cultivar diversity, Mahabub Hossain, Manik Lal Bose and others published a research paper under the aegis of IRRI/PETRRA. I presume that this work has been the first ever study on biodiversity in rice crops based on field level information. The authors compared the number of varieties grown both at household and village level in Bangladesh. (The forthcoming book: Rural Livelihoods in Bangladesh: Changes and Challenges by Mahabub Hossain and others contains further details).

The study picked up the concentration of area under the varieties for the district with the highest proportion of area under MVs (Bogra, mostly irrigated in the northwest region) and the district with the least coverage (Bagerhat, mostly rainfed in the coastal area). The findings seem to be revealing. In Bogra 29 MVs cover 84 percent of the total rice area, and 25 TVs cover the remaining 16 percent. In Bagerhat 14 MVs cover 24 percent of the total rice area, while 21 TVs cover the remaining 76 percent. The number of rice varieties is thus higher for the technologically progressive district than for the technologically backward one. In Borga the most popular MV, BR11, covered 20 percent of the rice area, while in Bagerhat the most popular TV, Sadamota covered 22 percent of the area. In Bogra the top 10 varieties (among them are two TVs, Khatohadal and Bhatraj) cover 84 percent of the total rice area, while in Bagerhat the top 10 varieties cover 81 percent (62% by TVs) of the total area. The VDI is estimated at 0.90 for Bogra and 0.89 for Bagerhat. The numbers do not support the hypothesis that the diffusion of modern rice varieties has led to a reduction in the number of rice varieties. The comparison does not indicate significant negative impact of technological progress on cultivar diversity either.

The study also assessed cultivar diversity at the farm and at the village level through in-depth household surveys for six villages representing different agro ecological conditions. The results are reported to be interesting too. The MVs have spread most in the well-drained medium highland (Darikamari, Bogra) and the least in the deeply flooded very lowland area (Dopmajgram, Pabna). The coverage by MVs was also somewhat lower in the upland areas (Ratnaiore, Dinajpur). The number of varieties grown at the village level varies from 13 to 19, but there is no systematic relationship between the coverage of MVs and the number of rice varieties grown. A large proportion of households grew more than three rice varieties, indicating considerable diversity. In one of the villages where MVs cover nearly 90 percent of area (Teghor), 64 percent of the households grew only one variety, but similar concentration was also found in the village that grew mostly TVs (Dopmajgram). In that village 56 percent of the households grew only one variety. The VDI varied from 0.43 to 0.90 across the villages. The index was not systematically related with the extent of coverage of modern rice varieties. In the six villages under study, it showed 47 rice varieties with the most popular variety (BR11) cover 25 percent of the rice area, and the five most popular varieties covering 60 percent. The VDI was estimated at 0.89.

Determinants of diversity

To analyse the factors influencing varietal diversity of rice the authors of the research paper ran a multivariant regression model with the household level data, relating the VDI to select socioeconomic and agro ecological factors. It is found that the diversity was higher on households with larger farm size and better-educated manager. But the diversity was lower in households with higher incidence of tenancy. The argument seems tenable. If you have more parcels of land, you can afford experiment with large number of varieties. On the other hand, tenants have to maximise earnings on the fixed amount of land and hence cannot afford to use different varieties. Also the higher the intensity of cropping for the farm, and the larger the proportion of low-lying parcels in the land portfolio of the farmer, the greater was the cultivar diversity. The cultivar diversity was found positively related with the extent of coverage of irrigation, contrary to the popular notion that expansion of irrigation reduces cultivar diversity by enabling farmers to grow modern rices. The regression coefficient for the variable representing the coverage of MVs was not found statistically significant, suggesting no relationship of cultivar diversity with the adoption of modern varieties of rice.

The major factor why farmers grow a number of rice varieties is the fragmentation and scattered landholdings and the diversity of the parcels with regard to agro ecological conditions. A single variety, although highest yielding compared to others, may not be suitable for all the parcels. The upland parcel contains adequate moisture for rice growing for a shorter period, and hence the farmers may prefer to grow a short-duration one, although it might be lower yielding. Shorter duration variety may also be preferred to smooth out the seasonality of labour demand and to obtain higher prices. On the low-lying parcels long duration varieties may be more suitable. For the Boro season farmers look for shorter duration varieties with cold tolerance to avoid the yield loss from cold injury and early floods. Many low-yielding traditional landraces are retained by farmers due to other useful traits such as better grain quality (for which consumers are willing to pay premium prices), adjustability to problem soils, tolerance to droughts (uplands) and submergence (lowlands), suitability for special rice products such as rice flakes and puffs (used as snacks), quality of straw and resistance to pests.

Raison d'etre

To study farmers' perceptions regarding maintenance of cultivar diversity the authors asked them to state the reason for growing more than one variety on the farm. The responses obtained in order of importance are: a) to suit different topographical locations of the parcel (70), b) suitability for different soil types (57), c) suitability for different growing seasons (52), d) marketing advantages (49), e) shorter maturity to help fit other crops in the system (39), f) reduce the risk of crop failure from natural calamities and pest pressures (30), g) different quality of grains (30), and h) different quality of straw (figures within parentheses are percentage of total cases). Since the complexity of these factors increase with the geographical scale, i.e., from the household to the village to the district and to the nation, the number of rice varieties grown and the varietal diversity increases with the broader geographical coverage (Perfect, 1991).

Conclusion

In Bangladesh the coverage of MVs has expanded to over 60 percent of the rice area. The technological progress has contributed to more than doubling of rice production with the same amount of land over the last 35 years. It is estimated that the growing of MVs contributed to an additional production of unhusked rice to the extent of 13.4 million tons, enough to feed 59 million additional people. The diffusion of MVs has reduced the unit cost of rice production, enabling farmers to offer rice to consumers at affordable prices. The real rice price has declined by over 40 percent since the mid 1960s, a major factor that contributed to moderate alleviation of rural poverty, in spite of slow growth in the overall economy.

The concern that the diffusion of MVs has eroded cultivar diversity was tested with limited data. While the number of rice varieties grown by farmers may be declining, the cultivar diversity is still very high and is not found to be negatively related to the diffusion of modern varieties of rice. It is estimated that Bangladeshi farmers still grow over 750 rice varieties in the country as a whole, 50 to 100 at the district level, and 10 to 20 at the village level. A large proportion of households grow more than one rice variety on the farm to suit the diverse agro ecological conditions of the parcels in their land-portfolio, and for different economic considerations such as obtaining better prices, smoothing out seasonality in labour demand and reducing the risk of yield loss from pest pressure and climatic disturbances.

Note on negatives

But that does not mean that MVs would not have any adverse impact on environment. In fact, the lack of biodiversity is one of the environmental outcomes arising from adoption of MVs. There could be, for example, more use of chemical fertilisers, pesticides or water contamination following the use of modern technology. That part should left to further research.

Abdul Bayes is Professor of Economics at Jahangirnagar University.