Mageretal- LOONS SIGNAL HIGH AGGRESSION WITH LONGER YODELS 
Flyover Distant intruder Close intruder Social gathering 
FIG. 2. Mean (± SE) probability of attack under different social contexts based upon >3,530 hrs of observation of 
Common Loons between 2002 and 2004 (/•\ 4g = 75.448. P < 0.0001, n = 84 males). 
fights in which individuals actually grab one 
another by cither the bill or neck and beal one 
another with their wings, following descriptions in 
McIntyre (1988) and Piper et al. (2008), and 
yodels. We recorded the time of day. the yodeler's 
physical posture (i.e.. ‘crouch’ or ‘vulture’ after 
Rummel and Goetzinger 1978). the estimated 
distance between yodeler and perceived receiver, 
and the number of repeat phrases given when each 
male yodeled. 
Playback Experiment.—We assessed how loons 
responded to yodels having fewer/more repeal 
syllables by recording vocal responses of local 
pairs to broadcast yodels that simulated an 
unfamiliar male’s intrusion into the territory. We 
““d a Portadat MDP 500 acoustic recorder 
HHB. London, UK) equipped with a Sennheiscr 
MHK-70 shotgun microphone (Old Lyme, CT, 
^A) to broadcast yodels from males recorded 
e «fflier that season at a distance between 15 and 
^ m from males in the study area whose 
territories were >8 km from a focal pair's 
territory (to reduce social recognition) (Waas 
lw k Mager et al. 2010). and did not play the 
^ me yodel to more than one pair (to prevent 
pscudoreplication) (Kroodsma 1989). Wc created 
undistorted yodels containing one. lour, and seven 
repeat phrases by adding identical repeat phrases 
t'Jthe same yodel using Canary acoustic soilware 
lV enion 1.5, Cornell University Bioacoustics 
^‘search Program, Ithaca. NY, USA), Thus, each 
yodel was identical in all acoustic parameters 
ex «pt for the number of repeat phrases. We 
playbacks between 2130 and 0230 hrs on 
three successive nights. We randomized the cider 
yodels were played on the three nights (e.g., 
1-repeat phrase the first night, 7-repeats the 
second, and 4-repeats the third night to one pair; 
4 , 7, I repeat phrases on successive nights to 
another pair, etc.). Each night, we broadcast 
yodels through a RadioShack (Fort Worth, TX. 
USA) 20-Watt amplifier and Super PowerHorn 
model 40-1.445 speaker at 90 dB (measured 10 m 
from the speaker) and rebroadcast the same yodel 
(having the same number oi repeat phrases) at 5 
and 10 min after the first. We recorded the time 
and type of all vocal responses during the 15-nun 
period following the first playback using the same 
recorder and microphone from which we recorded 
the playback yodels, 
We quantified the number of tremolo, wail, and 
yodel responses from focal pairs in addition to the 
latency before first vocalization. We interpreted the 
number of tremolos to reflect the extent pairs were 
threatened by the yodel, and the number of wa.l 
responses to reflect the pair’s level of alarm and 
interest to contact following Barklow (1979) and 
McIntyre (1988). We used repeated-measures 
ANOVA and associated post-hoc tests to investigate 
differential responses hy pairs to the three classes ol 
playback. We used the PROC MIXED procedure 
(SAS after Singer 1998, Johnson 2002) to construct 
linear growth models to examine whether responses 
given could be attributed to more repeat phrases 
while controlling for variability among pairs. We 
accepted significance at a Bonferroni-corrected a of 
0.05 for all statistical tests. 
RESULTS 
We recorded conspecific intrusions on 57 
territories and the yodeling responses of 58 banded 
males defending those territories (x ± SE number 
