Taro [Colocasia esculenta (L.) Schott] production by small-scale farmers in KwaZulu-Natal : farmer practices and performance of propagule types under wetland and dryland conditions.

Abstract

Ethno-archaeological evidence shows that taro [Colocasia esculenta (L.) Schott] originated in Asia. It may have been brought into South Africa a few hundred years after 300 BC from Madagascar, where Malaysian settlers introduced it about 300 BC. The crop is grown in the tropical and subtropical regions of the world, largely for subsistence on farms. In South Africa, taro is mainly produced in the subtropical coastal belt, stretching from Bizana in the Eastern Cape to the KwaZulu-Natal north coast. Although it is a staple crop for the subsistence farmers who grow it, there are no data on taro agronomy in South Africa. The hypothesis of this study was that traditional knowledge about taro production practices is not adequate to form a basis for agronomic and extension interventions to promote the status of the crop to that of a commercial commodity. A survey was conducted at two districts in KwaZulu-Natal, Umbumbulu and Ndwedwe, where taro is a staple crop. The objective of the survey was to determine the cultural practices associated with taro production, including knowledge about varieties, agronomy, plant protection, storage and marketing. Qualitative data obtained from the survey was used to plan an investigation into the agronomy of taro. The survey showed that subsistence farmers at Ndwedwe and Umbumbulu used traditional methods for taro production that had very small influence from the extension services from the Department of Agriculture. The farmers identified three varieties of taro, which they designated as the "red", "white" and "Zulu" types. The "red" and "white" designations were based on consistent crop morphological characteristics. This finding confirmed the reliability of indigenous knowledge for crop classification.The survey also revealed that wetland and dryland conditions are used to produce taro. At Umbumbulu, production occurred predominantly under dryland conditions, whereas at Ndwedwe there was an almost even utilisation of both wetlands and drylands. At both locations, the farmers estimated plant spacing using their feet, which showed that the plant populations would be about 18400 plants ha(-1). Full corms were a predominant type of propagation material. In the light of the survey findings about site types (wetland or dryland), propagation material and plant spacing for taro production, field experiments were designed to 1) determine the effect of site type on taro production, 2) compare three propagule types (full corm, full corm with a shoot and half corm) in taro production and 3) examine the effect of planting density (18400, 24600 and 37000 plants ha(-1) on the performance of propagules with respect to production under wetland and dryland conditions. Field experiments showed that wetland cultivation improved taro yield by 40% compared with dryland production. However, in each of the two site categories, there were significant differences between sites. Using full corms with shoots also enhanced taro yield (42% > full corms without shoots and 66% > half corms), when means were determined across all sites and planting densities. Increasing planting density also caused an increase in taro production (4.9 t ha (-1), 6.8 t ha (-1) and 11.5 t ha (-1), for 18400,24600 and 37000 plants ha,(-1), respectively; LSD (0.05) = 1.4 t ha,1). The enhanced performance of taro under wetland conditions, where corms with a shoot were used and at high planting densities may have been associated with photosynthetic efficiency. Wetland conditions and corms with shoots improved plant emergence and plant growth, which are essential agronomic conditions for efficient capture of the sun's energy for photosynthesis. It is proposed that using propagules with shoots and high plant populations under dryland conditions could enhance taro production. Although wetland cultivation enhanced yield, the survey showed that the total area of land that could be used for wetland cultivation at Ndwedwe and Umbumbulu was too small to warrant sustainable wetland production.