An experiment combining the soil pH factors and the form of nitrogen fertilisation was conducted in 1994 at the Bergerac Tobacco Institute, France. The experiment was carried out on a sandy soil on which adjunctions of dolomia (30% CaO + 20% MgO) were applied every other year since 1972, in four liming variants : 0 t/ha (C0), 1 t/ha (C1), 2 t/ha (C2) and 4 t/ha (C3), resulting in 1994 in H_{2O pH's of respectively 6.0, 6.4, 7.0 and 7.2. Nitrogen fertilisation (300 kg/ha) was supplied in nitric (N1) or ammonium (N2) form. The variety studied was Burley ITB 2601 stalk-harvested. The experiment was conducted following a split-plot RCB design with 4 replications. In comparison with N1, the N2 form resulted in tobaccos with less Ca, more K and similar Mg contents. Liming decreased the leaf content in K, Ca and ashes (competition with Mg). The leaf content in trace-metals was modified to a great extent by both factors : in comparison with N1, the N2 form multiplied the content in Mn, Zn and Cd by 3.8, 1.5 and 1.4 respectively, with an increasing effect for decreasing soil pH. For both nitrogen forms, and for all liming variants, the content in trace-metals decreased by up to -73%, -64% and -70% for Mn, Zn and Cd respectively.}These effects can be explained by the increase in the bioavailability of trace-metals due to a general acidification of the medium during nitrification and to rhizosphere acidification during ammonium uptake (C0, C1). Liming compensates the acidifying effect of N2. As the lifetime of ammonium in soil is brief, and the leaf content in trace-metals does not vary much according to stalk position, it is supposed that plants take up importants amounts of these elements at early stages of growing, and spread them thereafter amongst all stalk positions during growth.