SONG FEATURES ESSENTIAL FOR SPECIES DISCRIMINATION AND BEHAVIOR ASSESSMENT BY MALE BLACKBIRDS (TURDUS-MERULA)

The present study tested the signal value of different parameters of the blackbird song by comparing territory owner's responses to unmodified motif parts. Parameters were modified with reference to their natural limits of variation which were known a priori for some of the parameters. Those for the frequency of occurrence of the three types of motif sound figures (unmodulated and/or complex toned FM-figures, and compound MIX-figures composed of pieces of both the CF- and the FM-type) and their modulations, and the relative sound pressure (SP) of their harmonics, were determined in this study. The following results were obtained: The frequency of occurrence of the sound figure types is essential for species discrimination in the sense that motif parts must contain CF-pieces with a relatively constant frequency (either as CF-figures or as parts of MIX-figures), and it communicates additional information: a high proportion of FM-figures giving a gurgling sound character communicates a high arousal and vice versa. The ratio of SP of the fundamental to that of the 1 st and the 2nd overtone is also essential for species discrimination. The ratio must be greater than about 1:0.2:0.2 making the motif parts relatively pure toned. Above this value the SP ratio may communicate additional information: a relatively low ratio giving a hoarse sound character seems to represent a high arousal and vice versa. Also, a high sound pressure level of the motif parts communicates a high arousal and vice versa. Neither the order of sound figures, nor the specific variation across the motif part of the sound figures' frequency level, duration and amplitude is essential for species discrimination. So is, however, a certain variation per se of each of the three features across the motif part. The results fit in with the 'room for variation hypothesis' on the principles of coding information into bird song. Thus the species-essential parameters seem to be 'tuned' to their natural limits of variation in the sense that species discrimination ceases when they are changed beyond these limits; within these limits some of them convey additional behavioural information; and they include all sorts of parameters, e.g. basic as well as higher order syntactical and sound type parameters, and invariant as well as variable ones.