• Shackleton, Scott
  • Ratcliffe, Laurene M.
  • Weary, Daniel M.


Black-capped Chickadee (Parus atricapillus) song consists of two pure tonal notes: "fee bee" (see NORM in Fig. 1). Frequency descends during the fee note (termed glissando), and there is a drop in frequency between the fee and the bee (termed pitch interval; Weisman et al. 1990). The frequency ratio of the glis- sando (fee-start/fee-end), and the ratio of the pitch in- terval (fee-end/bee) are highly stereotypic, varying by less than two percent among and within birds in pop- ulations from Alberta, Wisconsin and Ontario (Weis- man et al. 1990). The frequency at any given point within a song is highly predictable from any other point, that is, highly correlated to all other points. Although chickadees shift their songs over a wide range of ab- solute frequencies (Shackleton 1991), the glissando and pitch interval remain constant. Playback studies con- ducted in the laboratory showed that adult male Black- capped Chickadees respond less aggressively to songs where the pitch interval (fee-end/bee) had been altered, compared to normal songs (Weisman and Ratcliffe 1989). Therefore, to examine the importance of the glissando (fee-start/fee-end) to song recognition in this species, we played back songs to wild males in which we altered both of these relative frequency parameters.


We synthesized three playback songs using the Mitsyn sound synthesis package (Bregman 1982). The amplitude envelopes of the three songs were identical and the volume settings were held constant during playback sessions. All three songs had identical bee notes. The only difference between the normal (NORM) and alered songs (ALT1 and ALT2) was the presence or absence of a glissando within the fee note. All frequency ratios in NORM were set at the species' mean (fee- start/fee-end = 1.06, fee-end/bee = 1.13; Weisman et al. 1990) and the absolute frequency of NORM was set at the mean frequency of songs recorded during the year previous to this study (fee-start = 3.8 kHz). Both ALT1 and ALT2 had no frequency sweep within fee (i.e., fee-start/fee-end = 1.0). The frequency of fee in ALT1 was equal to the frequency at fee-start in NORM and the frequency of fee in ALT2 equalled the frequency at fee-end in NORM. Thus in ALT1, the pitch interval was larger than normal, while in ALT2 the pitch interval was normal (Fig. 1). Since all other features of these songs were identical (amplitude, bee note and frequency), any difference in response between the normal and the altered songs must therefore be due to the presence or absence of the glissando in fee. A repeated measures experimental design was used to account for inter-subject variation. All three stim- ulus songs were presented to each subject in a single session consisting of a priming phase followed im- mediately by three playback trials. Sessions were al- ways performed near the center of the subject's territory. During the priming phase we played a tape consisting of 30 sec of chickadee calls, which simulated an intruding male and served to attract the subject to the vicinity (<40 m) of the playback speaker. The effect of these calls on subsequent response was controlled by presenting the same tape to all birds. The calls on the priming tape were from another population, and were unfamiliar to any of the subject birds. During each trial, we played the test song at a rate of 10 songs/ min for 90 sec followed by 120 sec of silence. This playback rate simulates natural counter-singing. We fully counterbalanced the assignment of songs to trials in a block randomized list prepared prior to the experiment. One observer (SAS) recorded the subjects' movements (number of flights and distances from the speaker) and vocalizations (fee bees, "chickadees" and "gargles"; Ficken et al. 1978) at a distance of 10 m from the speaker. As an aid to distance estimation, a 20 m measuring tape was laid on the ground with the speaker at the centre. Distances were estimated to the nearest meter. In equivocal cases, perch locations were noted during the trial, and distances measured directly thereafter. Playback songs and calls were played on a Sony Professional WM D6C cassette recorder and broadcast through an Aiwa SC-A8 speaker/amplifier. Playbacks were performed between 26 April-20 May 1991 to territorial males in two populations: at the Queen's University Biology Station (QUBS), Chaffey's Locks, Ontario (n= 18 birds) and near Chelsea, Quebec (n = 6 birds). These two sites are separated by ap- proximately 200 km. These populations were part of a larger study of the singing behavior of this species (Shackleton 1991).