The Reed Warbler shows the most specialized morphological traits for vertical climbing among the six central European Acrocephalus species [A. scirpaceus, A. melanopogon, A. palustris, A. schoenobaenus, A. arundinaeceus and A. paludicola]. We consider this as an adaptation to the vertical structures in its habitat. In experiments with young Reed Warblers I tested whether early experience with different perches has an influence on the choice of these structures later on in life. Locomotory activity of the following groups was tested in double choice experiments (horizontal versus vertical perches): "Control group vertical" (Ks): raised on vertical perches, "Control group mixed habitat" (Km): raised on vertical and horizontal perches, "Test group horizontal" (Vw+): raised on horizontal perches, "Test group horizontal, plus rewards" (Vw+): raised on hirozontal perches and additional feeders attached to the bars. With the exception of "Control group vertical" (Ks) which shows no choice preference (Fig. 2) all other birds preferred the vertical test substrate (Fig. 2 and 3). Even the reward (food) as a positive reinforcement of horizontal elements had little effect on their attractiveness (Tab. 1). An increase in the choice of the vertical substrate could be observed throughout the 3-day test period in all groups (Tab. 1). This increase was only due to the experience in the test situation for the birds in both control groups (Ks and Km) and the "Test group horizontal" (Vw) (Tab. 2). Therefore the use of vertical or horizontal structures in Reed Warblers is determined by: 1. An "innate" preference for species-specific substrate, 2. Novelty due to early experience (choice behaviour of "Control group vertical", Fig. 2), 3. Experience while using different substrates (experience in the test situation) which optimizes substrate choice according to proprioceptive learning. Additional choice experiments with the same test birds later on the year (October to December) revealed the same results. Therefore it seems unlikely that ther exists a preprogrammed change in substrate choice in the course of a year (Fig. 4). Although choice experiments of this kind can only include a limited part of the "habitat-scheme", they are useful experimental designs for investigating specialized species whose habitats can be simulated closely in the laboratory.
