Lishman and Nol • MAGELLANIC PLOVER HABITAT SELECTION 
489 
different sets of behavioral patterns. Transient 
birds would walk or fly long distances (e.g.. 500 m 
to 2 kin) within the lake area, forage for extended 
periods of time (>2 hrs), and neither associate 
with other birds, nor exhibit territorial behaviors. 
Territorial birds vocalized in the presence of 
an observer or conspecific. were aggressive with 
conspecifics and. at times with other species 
(Charadrius falklandivus: but see Jehl 1975), 
associated strongly with a male (copulating, 
following, or giving joint territorial displays), 
and/or incubated a nest or brooded and fed a 
chick. 
We recorded the location of each plover 
encountered on the shoreline of lakes, using a 
Global Positioning System (GPS) (Garmin Inter¬ 
national Inc.. Olathe, KS. USA). The spatial 
distribution of observations of territorial birds was 
examined using Program Garmin Mapsource®. 
Locations within 200 m of one another were 
classified as a single territory, based on previous 
observations that pairs were spaced along lake- 
shores every 200-300 m (Jehl 1975. Ferrari el al. 
2008). The validity of classifying clumps of 
observations as a territorial pair was confirmed 
by observations of territorial disputes at territory 
boundaries and by stability of territory locations 
on successive visits within and between the two 
field seasons. The number of territories per lake 
was calculated for 18 occupied lakes that were 
searched thoroughly in both seasons. 
Nests were considered successful if one or more 
eggs hatched, young were seen, or small eggshell 
fragments were found in the nest cup (Mabee 
1997) and the parents were still on territory. They 
were considered unsuccessful if we found no eggs 
or adults on territory (Mabee 1997). We presumed 
predation was the cause of failure if eggs and 
young were absent. Eggs crushed inside the scrape 
were considered trampled. We used the Mayfield 
(1961 1 method and Johnson’s (1979) alternative to 
incorporate exposure days (days between nest 
visits) and number of total nest losses to calculate 
daily nest survival. 
Magellanic Plover habitat was described at 
three spatial scales: lake, putative territory, and 
nest (microhabitat). We did not measure actual 
territory sizes and assume our measurements at 
the scale of 200-m radius around nest sites and/or 
feeding areas of pairs of plovers, represent broadly, 
territories of this species, given the average dis¬ 
tance between adjacent pairs. Wc considered all 
endorheic lakes in the known geographic distribu¬ 
tion in Santa Cruz Province, Argentina (Ferrari 
el al. 2003) as available at the lake scale and 
searched 53. based on road accessibility. Lakes 
were considered occupied if individual Magellanic 
Plovers were observed at least once during the 
2006 and/or 2007 breeding season. We used data 
from 22 lakes searched on five occasions (Aug, 
Oct. Nov. and Dec 2006. and Dec 2007) to estimate 
monthly detection probability, using the occupancy 
model of Program MARK (White and Burnham 
1999). 
We examined satellite images from Google 
Earths and measured lake perimeter, total area 
and size of aeolian lunettes using the "distance 
ruler’. The high water line of the lakes, clearly 
visible on satellite images, delineated the lake 
perimeter. The aeolian lunette's shape w'as approx¬ 
imated to that of a rectangle. Lunette area was 
calculated using the equation: area (in ha) = length 
(m) X width (m) X 0.0001 ha/m 2 . Elevation of the 
shoreline weis taken from the digital elevation 
model (DEM) provided by Google Earth®. 
Electrical conductivity was measured by CL to 
approximate salinity on a subset of 22 lakes that 
were visited from 14 to 20 December 2007 using a 
Eutech ECTestr 10 high' M tester (Oakton Instru¬ 
ments, www.instrumart.eom/Oakton) to the near¬ 
est 0.1 mS/cm. Measurements >19.9 rnS/cm were 
nol detected by this device. The measurement was 
taken by directly immersing the device’s probe 
in the hike water at the shoreline on 12 lakes. 
Lakes that were dry Of = 10) were measured by 
dissolving 10()-ml hike bottom sediment in 100 ml 
distilled water and immersing the probe in this 
solution. Salinity metrics were binomial in 
distribution with all readings between 0 and 
0.8 mS/cm (freshwater) or >19.9 mS/cin (saline). 
Salinity measurements W'cre confirmed by taste. 
A 300-m buffer at the high water line of the 
lake delineated the available area at the territory 
scale. Wc measured 33 sites within that area used 
by Magellanic Plover and 33 available sites along 
the lake shores (used and available, respectively). 
A territory was considered used if a nest or chick 
was found and the location of the nest or chick 
defined the center of the used territory. Unused 
sites at this scale were chosen using randomly- 
selected angles and distances (0-200 m) from 
the center of evenly spaced point s around the 
perimeter of the lake. The next appropriate 
randomly generated distance was used if the 
unused site w'as outside of the designated 
