Three-week-old Picea abies seedlings were grown for 7 days in 100 microM aluminium (Al), combined with 1000 or 2000 microM silicon (Si). Solution pH was adjusted to 4.00, 4.25, 4.50, 4.75, or 5.00. In the absence of Si, solution pH had no effect on the decrease in root growth caused by 100 microM Al. Silicon did not ameliorate toxic effects of Al on root growth at pH 4.00, 4.25 and 4.50, whereas significant, and apparently complete, amelioration was found at pH 4.75 and 5.00. An equilibrium speciation model (EQ3NR), with a current thermodynamic database, was used to predict the behaviour of Al and Si in growth solutions. When Si was not present in the 100 microM Al solutions, Al(3+) declined from 92.4% of total Al at pH 4.00 to 54.6% at pH 5.00, and there was a concomitant increase in hydroxyaluminium species as pH increased. The addition of 1000 microM Si to the 100 microM Al solutions caused a reduction in Al(3+) content over the whole pH range: at pH 4.00 Al(3+) fell from 92.4 to 83.3% in the presence of Si; and at pH 5.00 the fall was from 54.6 to 17.7%. These falls were attributed to the formation of hydroxyaluminosilicate (HAS) species. Similar, but somewhat greater, changes were observed in solutions containing 2000 microM Si. The match between root growth observations and the modelling data was not very good. Modelling predicted that change in Al(3+) content with pH in the presence of Si was gradual, but root growth was markedly increased between pH 4.50 and 4.75. Differences between root growth and modelling data may be due to the model not correctly predicting solution chemistry or to in planta effects which override the influence of solution chemistry.