Chemical and physical factors limiting regeneration of natural vegetation on lands reshaped six years earlier in the Scotland District of Barbados were studied. Soils from both the Scotland sandstones at Walkers and Scotland clays at Colon were saline-alkalis, conductivities of saturation extracts and exchangeable Na percentages commonly being as high as 7000 = ?mho/cm and 30 per cent respectively, but salt distribution was patchy. Soluble salts varied between 0·2 and 13 mequiv./100 g along a line transect at Walkers which had 206 sampling points at 1 m intervals, and pH varied from 6 to 9. Weight of vegetation was low at high salt concentrations but was sometimes high and sometimes low at low salt levels. Sparseness of vegetation over most of the area was attributed more to erosion than to salt, islands of vegetation often standing 12 to 15 cm higher than surrounding bare land. Chloris barbata was the most frequent grass and together with Euphorbia prostrata, Rhynchosia minima, Neptunia plena, leucaena glauca, and Crotalaria verrucosa was found at salt concentrations greater than that (1·6 mequiv./ 100 g) which prevented the establishment of the other 30 species. Reclamation could be improved by collecting and distributing seed of those salt-tolerant plants which have value as fodder. Erosion control was obtained by mulching with cut grasses or by products of sugar factories. A ridge and furrow system graded at 1 in 15 (clay) and 1 in 20 (sandy loam) controlled surface drainage. Mulching reduced both the rate and amount of run-off, losses of up to 70 per cent of the rainfall being common on bare dry soil. The average weekly moisture content of mulched topsoil was 7·6 per cent v/v (Walkers) and four per cent v/v (Colon) more than bare topsoil over a six month period. Evaporation losses did not differ markedly on bare and mulched land but were greater from the clay with a deeply cracked soil surface than from the sandy loam with a smooth uncracked surface. Soil temperature and physical conditions near the soil surface were more suitable for seedling establishment in the mulched than in bare soil. Nutrients available from different mulch materials were negligible and responses to fertilizer were obtained on mulched but not on bare land.