The aim of this study was to compare the effects of salmeterol and beclomethasone on lung function and symptoms in children with mild to moderate asthma. Sixty-seven children not treated with inhaled corticosteroids were randomized in a double-blind parallel study either to salmeterol 50 micrograms b.i.d. or beclomethasone 200 micrograms b.i.d. After one year, FEV1 significantly increased in the beclomethasone group, whereas in the salmeterol group there was a small reduction. Differences between groups were 14.2% predicted (p < 0.0001) and 7.0% predicted (p = 0.007) for pre- and postbronchodilator FEV1 values, respectively. PD20 methacholine decreased by 0.73 DD (p = 0.05) in the salmeterol group and increased by 2.02 DD (p < 0.0001) in the beclomethasone group. Morning and evening PEF and symptom scores improved in both groups, although more in the beclomethasone group. Asthma exacerbations, for which prednisolone was needed, were more frequent in the salmeterol group (17 versus two), as were the number of withdrawals due to exacerbations (six versus one). However, growth was significantly slower in the beclomethasone group (-0.28 SDS) compared with that in the salmeterol group (-0.03 SDS) (p = 0.001). We conclude that treatment with a moderate dose of beclomethasone is superior to salmeterol in children with mild to moderate asthma and recommend that salmeterol should not be used as monotherapy.