As part of a comprehensive interdisciplinary research pro,sram the relationships between site conditions and growth of a balsam poplar (P. trichocarpa cv. Muhle Larsen), an aspen (P. tremula x tremuloides cv. Astria) and a willow (Salix viminalis) are being tested since 1983 in three experimental short-rotation plantations on former arable soils distributed over the FRG. The site conditions, the experimental design and the growth behaviour after two (experimental sites Abbachhof in the Jurassic region of Upper Palatia and Canstein in the eastern Rhenanian Mountains) and after one rotation (site Wildeshausen in the lowlands of NW Lower Saxony) are described. Each rotation lasted for five years. The methodological approach is a partly factorial fertilization trial with the nutrients N, P, K, Mg and Ca. Shoot biomass of the tested clones after the first five-year rotation period showed cumulative dry-matter amounts between 10 and 37 t x ha(-1). Highest production was measured from balsam poplars and aspen at Abbachhof, while shoot growth at Canstein and Wildeshausen as a rule was remarkably lower with no clear differences between the tested clones. Based on the first five-year rotation cycle, the average annual shoot axes production of the unfertilized aspen and balsam poplars were 5.1 and 5.9 t ha(-1) at Abbachhof. Production was very low for all clones at Canstein and Wildeshausen and for Salix viminalis at Abbachhof with values around 2-4 t ha(-1). After the second rotation period, a clear differentiation between the sites showed up, resulting in an 2.5-3-fold increase of cumulative shoot biomass of all three species at Canstein to values of ca. 50-70 t ha(-1), resulting in annual shoot axes production of ca. 10 to 14 t ha(-1). Tn contrast, at Abbachhof Muhle Larsen remained at a similar level of production as in the first rotation period (30-35 t ha(-1)), while Salix significantly increased, and the aspen reduced shoot biomass. Salix shoot yields responded significantly to nitrogen fertilization at Abbachhof (50 and 100 kg N ha(-1) year(-1)) by ca. +33 and +73% (first rotation) and +29 and +47% (second rotation), and at Canstein by ca. 25% in the 100 kg N ha(-1) year(-1)-treatment. At Abbachhof, the aspen reduced growth as compared to the first rotation, probably due to heavy attack by defoliating insects. The balsam poplar Muhle Larsen proved to be the most interesting species tested. In all respects, it seems to be well supplied with nutrients for optimal growth and does not need any fertilization or pesticide treatment for at least 10-15 years. So far, no effects of P, K, and Mg fertilization with/without combination with liming were detected. (C) 1999 Published by Elsevier Science B.V. All rights reserved.