Pieplow and Francis • SONG VARIATION OF YELLOW-EYED JUNCOS 
465 
Arizona Oaxaca Baja 
“1A 
D 
G 
i j 
°1B ,0 ' 
s WVJJ- 
K ! 2 
E ,0 - 
K 1 2 
'“Xi w. > 
•ec ! 2 
°lc 10 - 
*K 1 2 M 
F ? 
K 1 2 
FIG. 1. Spectrograms of Yellow-eyed Junco songs from Arizona (A-C), Oaxaca. Mexico (D-F), and Baja California 
Sur. Mexico (G-I). All recordings by NDP except D (courtesy of Florida Museum of Natural History), E (Charlie Vogt), 
and F (Peter Boesman 2006). Complexity of song syntax increases from left to right. 
USA and Oaxaca, Mexico). This study represents 
an initial quantitative assessment of differences 
between these taxa to identify potential harriers to 
gene flow between the isolated populations and to 
examine if re-evaluation of the taxonomic status 
of J. p. bairdi may be warranted. 
METHODS 
Recordings. —Songs of J. p. bairdi were 
recorded on a Western Field Ornithologists and 
Sonoran Joint Venture expedition to the Sierra 
La Laguna, Baja California Sur, Mexico, during 
14-19 July 2008. Songs of J. p. palliatus were 
recorded in the Huachuca and Chiricahua moun¬ 
tains, Arizona in May 2009. We considered the 
difference in the month of recording unlikely to 
influence study results, as both study sites were 
visited during the lengthy active breeding season; 
Arizona Yellow-eyed Juncos raise up to three 
broods per year from late April to the beginning of 
August (Sullivan 1999), while juncos in Baja 
California during the 2008 expedition were at all 
stages of breeding from nest-building to indepen¬ 
dent young (Carol Beardmorc ct al.. unpubl. data). 
Recordings were made in 24-bit WAV format 
with a sample rate of 44.1 kHz on a Fostex FR2- 
LE recorder with a 55-cm Telinga parabola and 
stereo DAT microphone. Additional recordings of 
J. p. bairdi from Baja California Sur. J. p. 
palliatus from Arizona, and nominate J. p. 
phaeonotus from Oaxaca were assembled from 
the collections of the Macaulay Library at the 
Cornell Laboratory of Ornithology, the Florida 
Museum of Natural History, the Borror Labora¬ 
tory of Bioacoustics, the Xeno-Canto online 
collection, and the private collections of S. N. 
G. Howell, Andrew Spencer, and Richard Web¬ 
ster, as well as the commercially published audio 
collection of Boesman (2006). Recordings from 
the Macaulay Library, the Borror Laboratory of 
Bioacoustics, Xeno-Canto, and Boesman (2006) 
were originally obtained in MP3 format and 
converted to WAV format at 44.1 kHz for 
analysis. Recordings by S. N. G. Howell were 
digitized in WAV format at 44.1 kHz from the 
original cassette tape by NDP, Typical songs from 
each region are illustrated in Fig. 1. 
We randomly selected five songs of the same 
song type from each recording for each individual 
sampled. The recording contained a single song 
type in almost all cases; if a recording included 
more than one song type, we randomly selected 
five songs of the song type that was most common 
on the recording. We used Marler and Isaac’s 
(1961) definitions of junco song features: a note is 
a continuous vocal utterance, a syllable is two or 
more notes grouped to form a single coherent unit, 
a trill is a syllable repeated at least twice 
consecutively, and a phrase is one or more 
dissimilar syllables that are separated from other 
phrases by a trill. We measured 13 song features 
(Fig. 2): song length (sec), total number of 
