Baker • SILVER EYES SONG EVOLUTION 
459 
Number of syllables recorded 
FIG. 2. Cumulative plot of new syllable types ex¬ 
pressed with number of syllables sung, indicating this 
Silvereyes had no tendency toward a finite vocabulary, and 
instead syllable innovations were being produced as it 
continued singing. 
FIG. 4. Cumulative plot of new syllable types ex¬ 
pressed with number of syllables sung by three Silvereyes. 
These curves do not appear to reach asymptotic plateaus, 
although bird #10 has such a suggestion in its last few 
songs uttered. 
of this bird revealed at leas! 128 different kinds 
of syllables among the 278 lotal syllables 
constituting the 13 songs. The syllable repertoire 
increased in an essentially linear manner as more 
and more songs and syllables were examined 
(Fig. 2) indicating new syllables were continu¬ 
ously produced as the bird continued to utter 
songs. It was clear from visual examination of 
the syllables that some were recurring with 
reasonable fidelity throughout the 13 songs 
(Fig. I), but the great majority did not occur 
more than once or they recurred only a few times, 
although a very few were found more often 
(Fig. 3). Three additional Woody Island birds 
with lesser number of recorded songs were also 
No. times a syllable recurred 
FIG. 3. Recurrence of syllable types in a sequence of 
278 syllables from Silvereyes comprising 128 syllable 
types. A few syllable types recurred frequently, but over 
60% were not repeated and 90% were found three times or 
less in the sample. 
examined for comparison, and their saturation 
plots (Fig. 4) were similar to that of the first bird 
examined, apparently not reaching any finite 
repertoire size and suggesting larger samples of 
syllables from these birds would parallel the 
previous bird (Fig. 2). 1 also examined the songs 
of three birds from the Esperance sample and 
obtained results parallel with Woody Island. An 
additional result was in comparing the Esperance 
syllables to those from Woody Island. The three 
Esperance birds in this limited comparison 
shared, on average, about 25% of their syllable 
types with the four Woody Island birds (6 
examples of syllables shared between popula¬ 
tions in Fig. 5). Thus, whereas Esperance birds, 
like those on Woody Island, performed what 
seemed to be an open-ended production of new 
syllables as they continued singing (no asymptote 
in a cumulative plot) they still shared some of the 
kinds of syllables with the island samples, as 
judged by visual comparison. 
I concluded that individual males continue to 
express new kinds of song syllables as they 
continue singing, and there is no song stereotypy 
across the songs of a given bird. Further, in spite 
of some syllable 'types' recurring, there is too 
little syllable-level stereotypy to suggest a fixed 
repertoire. Therefore, perhaps the best way to 
characterize a sample of songs was to use, in a 
descriptive metric, all the syllables uttered by 
each individual. However, validity of between- 
popuJation comparisons, in any search for cultural 
differentiation, would rest on the assumption that 
even with possibly unbounded generation of new 
