Summary Slash‐and‐burn farmers in northern Laos consider weeds, insufficient rainfall and rodent damage as the most important constraints to upland rice (Oryza sativa L.) production. Labour inputs of 140‐190 days ha‐1 for weed control result in low labour productivity. Average weed cover observed in rice fields was 5.6, 4.1, 2.1, 1.7, 0.7 and 0.7 cm m‐1 (transect) for Chromolaena odorata (L.) King and Robinson, Ageratum conyzoides L., Commelina spp., Lygodium flexuosum (L.) Sw., Panicum trichoides Sw. and Corchorus spp, respectively. Chromoiaena odorata, which was introduced in the 1930s, has become the main fallow species and is considered a desirable fallow plant by most farmers. Average fallow periods reported for the 1950s, 1970s and 1992 were 38, 20 and 5 years respectively. Reduced fallow periods in the last decades have re‐suited in a marked increase in weeding requirements. Above‐ground biomass for rice stem, herbaceous plants and trees after rice harvest was 168, 67 and 60gm‐2 in 1991 and 115, 43 and 24 gm‐2 in 1992. Weeding at 14‐day intervals did not increase rice grain yield.

Decline in soil fertility accelerated by shorter fallow periods was expected to be a major constraint in slash-andburn rice production systems in northern Laos. In this paper we describe relationships between fallow period, soil fertility parameters, weeds and rice yield. Soil infertility is not perceived a major yield constraint by the farmers. Of the various soil parameters observed only soil organic matter showed consistent association with rice yield (r =0.42, p<0.01). Fallow period and rice yield showed no association and the relationship between fallow and organic matter was very weak (r=-0.16, p<0.01). Rice yield was negatively related to densities of Ageratum conyzoides and Lygodium flexuosum. Soil loss during the cropping period ranged from 300-29.300 kg ha -l . For the same period organic matter, total N, available P and available K content in the top 0-3 cm decreased by 11,12,17, and 17%, respectively, and loss of total N for the soil depth of 0-25 cm was estimated at 400 kg ha-I. Soil physical properties, moisture stress and available N are the most likely detriments to rice yields. Further attempts to relate soil properties to rice yield should include repeated measurements during the cropping season and observations on soil physical properties.

No abstract

No abstract

Reduced fallow length in slash‐and‐burn rice (Oryza sativa L.) production systems of northern Laos increased weed pressure, labour requirement and the need for soil and moisture conservation. On‐farm and on‐station studies and on‐farm surveys were used to evaluate the effect of residue management and cropping intensity on weed population, rice yield and nematode density. Residue loads were 2.3–4.4 t ha−1 after a rice crop and 9.5 t ha−1 after 1 year of fallow. Compared with farmers' traditional burning of crop and weed residues, mulching reduced rice yield by 43% in one out of four comparisons and increased weed biomass by 19–100%. Compared with continuous rice treatments (averaged over burning and mulching treatments), treatments with fallow or cowpea [Vigna unguiculata (L.) Walp.] in the previous year had 32% less herbaceous weed biomass, 90% fewer Ageratum conyzoides L. and over 99% fewer Meloidogyne graminicola Golden & Birchfield. Rice yield was negatively associated with A. conyzoides density (−0.62, P < 0.01) and M. graminicola number (−0.42, P < 0.05). Less striking effects of fallow period on A. conyzoides and M. graminicola, observed from on‐farm surveys, demonstrate the limitations of on‐farm studies because of undocumented effects of farmers' management decisions.