This innate contribution to vocal learning raises some interesting ques-
tions:by what criteria does the naive bird select a song to memorize?
does a specific note or phrase act as a flag,allowing the bird to recog-
nize the song? or do tonal or temporal qualities of the whole song
provide the necessary cues? what is the neural basis for this type of
guided learning? One can imagine circuits in the brain that act as feature
detectors and,when stimulated in an appropriate combination,cause the
song to be selected as a model for vocal learning.
Working with Jill Soha in Peter Marler’s laboratory at the University
of California,Davis,and in collaboration with Allison Doupe at the
University of California,San Francisco,we approached innate song
recognition using two complementary techniques to study both brain
and behavior of fledgling white-crowned sparrows (Zonotrichia leuco-
phyrus).The behavioral test is simple to perform (Nelson and Marler
1993).Fledgling sparrows are housed alone in soundproof boxes that are
outfitted with a speaker and a microphone.Once an hour they hear ten
repetitions of a taperecording of normal white-crowned sparrow song,
the song of another species (song sparrow,Melospiza melodiaor savan-
nah sparrow,Passerculus sandwichensis),or a white-crowned sparrow
song that has been experimentally altered.When fledglings hear normal
white-crowned sparrow song,they give a series of begging calls that,
under natural conditions,help their parents locate them for feeding.The
songs of other species usually elicit no vocal response.Thus,by counting
the number of begging calls given in response to our altered white-
crowned sparrow songs,we can learn whether the fledglings perceive
them to be acceptable renditions of white-crown song or to be of a
foreign species.
We used this behavior test to compare responses to normal white-
crowned sparrow song with responses to isolate white-crowned sparrow
song,the simple song produced by white-crowns that have not heard
other white-crowns sing.The isolate song consisted of a series of whis-
tles and lacked the trills and buzzes typical of normal song.The isolate
song was as effective as the normal song in eliciting begging calls from
fledglings.One hypothesis to explain the efficacy of isolate song is that
the whistle,universal to all white-crown dialects,acts as a marker for
recognition.We tested this hypothesis using songs artificially constructed
by repeating a single white-crowned sparrow phrase such as a whistle,
buzz,or trill,while maintaining normal song duration and tempo.If the
whistle acts as the critical flag for identifying the white-crowned song,
we predicted that it would elicit as strong a response as the normal song,
and the trill or buzz would be comparatively weaker.
All of the repeated phrase songs,including buzz and trill songs,proved
to be as effective as the normal song in eliciting calls from the fledglings.70 Carol WhalingMUS5 9/14/99 11:59 AM Page 70