Pizo . LEK BEHAVIOR OF THE PLOVERCREST 
107 
51 H'W; 130 m asl), two small towns in southern 
Brazil. The lek was in the transition area between 
ihe forest and an old field, where small trees and 
dvnhs (<I0 era diameter at breast height, <8 m 
ulll gradually give way to patches of herbaceous 
vegetation averaging ■—1 m in height. C ommon 
•ires included Zanthaxylum sp. (Rutaceae) and 
Cumria sylvestris (Flacourtiaceae) interspersed 
with a few exotic Acacia mearnsii (Mimosaceae). 
The area was surrounded by extensive patches ot 
secondary forest and small farm (usually <10 ha) 
rural properties. Meteorological data collected at 
Sao Leopoldo. -20 km from the study area, show 
that rain is well distributed throughout the year, 
totaling -1,500 mm. The average monthly tem¬ 
perature is 19.8 C (range = 13.8-25 C l. 
Lek Structure and Male Behavior .— I conducted 
in initial survey to locate all territories occupied 
by singing males, and then observed each and all 
males (a = 7) on the largest lek during 15-min 
penods over 20 days from August to December 
2006 totaling 22.75 hrs of observations (3.25 hrs/ 
male). Observations were randomly rotated 
among males so each was observed at least once 
during each hour of the day from 0600 to 18(H) 
hrs 1 recorded during each 15-min observation 
period: (I) the amount of time males spent in their 
territories, which for Plovercrests corresponds 
ilJ singing activity as males were constantly 
mgmg when in their territories: (2) the number 
of songs uttered during a I-min period recorded 
" lll ' e at the middle of each observation period: and 
I'Mhe number of perches used for singing. I also 
"oied ail visits by eonspccifics to lek territories 
mu. due to the short duration of visits and the 
mpid movements of interacting birds, 1 was 
J nable to ascertain the gender of every visiting 
mrd. The heights of the perches used for singing 
iverc measured after each observation. Songs 
produced by lekking males were recorded with a 
Sony TCM 5000 recorder and a Sennheiser ME66 
microphone. Sonograms were obtained using 
R aven Version 1.3 (Charif et al. 2007). 
I measured the area of each territory at the end 
(he study by considering them rectangular and 
takir »g the outermost singing perches as boundar- 
;es This method may have overestimated the area 
°f the territories, but I suspect this was not the 
** because boundaries of each territory were not 
delimited by perches used for singing and were 
Wended to where visiting birds were chased. 
Thus. I am confident the method used provided 
estimates of territory sizes that did not greatly 
depart from real values. The distances between 
neighboring territories were taken from the ap¬ 
proximate center of each territory. I continued to 
visit the lek area twice a month after the hi ids left 
the lek in December 2006 to look for singing males 
that could indicate resumption of lekking activity. 
Males were not marked and 1 could not be 
certain that territory owners did not change during 
the study. However, based on consistent use ol 
particular perches used for singing, the individu¬ 
ally recognizable songs of some males, and the 
short duration of the study, 1 do not believe that 
such changes occurred. Thus, for analysis, I 
assumed the same individuals held their territories 
and were observed throughout the study. 
Statistical Analyses.— One-way ANOVA was 
used to examine possible monthly and among- 
male differences in singing rates. Sample sizes 
were too small for some hours of the day when 
birds were generally absent from the lek, and 
daily variation in singing rate was not examined. 
Lek attendance data (defined as the proportion ol 
the 15-min observation periods spent by males 
inside their lekking territories) were not normally 
distributed, and I used Kruskal-Wallis tests to 
evaluate daily and among-male variation in lek 
attendance. I checked for the homogeneity ot 
variances of ranked data using the Brown- 
Forsythe test (Brown and Forsythe 1974) before 
performing Kruskal-Wallis tests, following the 
recommendations of Ruxlon and Beauchamp 
(2008), which revealed homoscedasticity o 
variances. Pearson correlation tests were used to 
investigate the relationship between singing rate 
and lek attendance All analyses were conducted 
with Statistics 6.1 (StatSoft 1999) with the level 
of significance of 0.05. Values are presented as 
means — SD. 
RESULTS 
The lamest lek consisted of seven male 
Plovercrests in an area ~70 X 30 m. Two other 
wroups were also observed along the road, one 
80 m (3 males) and the other 2 km (2 males) from 
die main lek. respectively. In addition, over a 3- 
week period (23 Sep to 14 Oct 2006), I observed a 
solitary male singing 200 m from Ihe main lek. 
Males were present at the leks (including those 
with 2 and 3 males) from August through mid- 
December 2006 with activity at the leks decreas¬ 
ing during December (Fig. 1A). The leks were 
inactive from December until the next July. 
However, at least six of the seven studied males 
