Baker • SILVEREYES SONG EVOLUTION 
455 
taxon (Diamond et al. 1976, Warren et al. 2006). 
The unusually high rate of Zosterops diversifica¬ 
tion in species and subspecies, together with rapid 
expansion and colonizing ability (potential gene 
flow), presents something of a paradox: i.e., how 
can one explain a high rate of differentiation from 
island to island (e.g.. as little as 2 km: Mavr and 
Diamond 2001) in a genus with an apparently 
exceptional dispersal capacity? Could learned 
mating signals have an isolating effect and serve 
to repel new attempts to colonize an island with 
an established population? 
Zosierops appears to be an ideal taxon with 
which to examine patterns of evolutionary differ¬ 
entiation in association with possible diversifica¬ 
tion of reproductive communication signals. Local 
populations are commonly dense with singing 
males during an early dawn chorus of —30 min. A 
given male's song is —5-7 sec but can range up to 
10 sec with noticeable spaces between successive 
songs, which are produced at a rate of five or so 
songs per min and consist of a series of discrete 
sound units usually referred to as notes or 
syllables (1 use the latter term). A song sounds 
like a warbling sequence of similar syllables not 
possessing any obvious syntactical variation to the 
human listener. Ecological studies and research on 
vocal communication behavior of Zosterops are 
scant with only a handful of studies reporting on 
ecology or behavior, and most have been 
conducted on a single island without comparison 
to other island or mainland populations. Thus, 
knowledge of song structures across the genus is 
lacking. 
These factors led me to conduct research on an 
island population (Woody Island) and an adjacent 
mainland population 15 km distant (near Esper- 
ance) of Zosterops in Western Australia. The 
objectives were to: (I) compare the populations to 
see if they differed in song traits, and (2) explore 
methodology that could be used in an extensive 
survey of Zosierops island endemics. Choice of 
the study areas was based upon two factors. First, 
they were relatively near where 1 lived. Second, 
given the species reputation for colonizing ability 
(and banding-recovery data from mainland loca¬ 
tions) the 15-km distance between the island and 
mainland populations seemed trivial. Therefore, if 
the two study populations differed vocally it 
would be consistent with the hypothesis that a first 
step in behavioral isolation is differentiation in 
reproductive communication signals. The two 
target populations of the present paper are 
considered conspecific; there has been no evidence 
collected to address their morphological or molec¬ 
ular differences, nor are there data on possible 
exchanges of birds between them. Exchanges 
between any Zosierops populations on different 
islands remain undocumented. Work on Zosterops 
populations occupying islands elsewhere indicated 
inter-island distances of 8-12 km or less separate 
morphologically differentiated forms (e.g., Mees 
1961). 
METHODS 
General Biology, Study Locations, and Sam¬ 
pling of Vocalizations .—Most Zosterops are 
known as White-eyes or Silvereyes for the 
characteristic white-feathered eye ring and tend 
not to exhibit sharply defined plumage patterns 
having more or less gray green to yellow green 
upper parts, green or brownish sides, and gray or 
yellow below. They are socially and genetically 
monogamous (Robertson et al. 2001). breed in 
dense colonies (11-30 pairs/ha: Catterall et al. 
1982). and males sing intensely for somewhat 
<30 min in the pre-dawn. 1 and a field assistant 
(M. S. A. Baker) tape recorded a mainland 
population sample of Zosterops lateralis gouldi 
(Johnstone and Stori 2004; Z. /. cliloronotus in 
some treatments) occupying coastal scrub habitat 
(low sclerophyllous shrubs/heath) near the town 
of Esperance, Western Australia on two succes¬ 
sive mornings during the pre-dawn chorus of the 
breeding season (Nov 2004). 1 recorded a second 
sample in similar habitat on two successive 
mornings, within a few days of the Esperance 
sample, on Woody Island, a 240-ha island 15 km 
offshore from Esperance in the Recherche Archi¬ 
pelago in the Southern Ocean. The second 
morning of recording of both populations began 
where recording the previous morning had ended. 
No searching for nests was conducted to follow 
breeding status. The song of Z lateralis has often 
been described as a ’'sequence of warbling notes" 
but. a wide variety of songs has been described in 
onomatopoeia (Pratt et al. 1987) among different 
island populations of Zosterops in the tropical 
Pacific; many of these verbal descriptions hold 
little resemblance to "a sequence of warbling 
notes". 
Birds were not marked, which would have been 
of limited help as most singing was under dim 
pre-dawn light when the recording occurred, and 
reading color bands would have been uncertain or 
impossible. I moved continuously through the 
