THE WILSON JOURNAL OF ORNITHOLOGY • Voi 124. No. 1. March 2012 
1 14 
used, (2) Jump Circle-a 110-ha circular grassland 
used as an air-drop zone, and (3) Westfield-a 
240-ha area encompassing two active airstrips. 
Fields are dominated mainly by warm-season 
grasses (e.g., Schizachyrium spp., Paniciun spp.) 
with significant amounts of bare ground (e.g., 
sand, lichens) and early-stage shrub encroachment 
(mainly Pinus rigida and species in the family 
Ericaceae). 
Nest Survival and Predation Rates .—Nest 
searching occurred within 16 irregularly shaped 
plots totaling 257 ha (mean plot size = 16 ha, 
range = 9-32 ha). Plot boundaries were delineat¬ 
ed to provide an ample amount of searchable 
habitat (i.e., the maximum deemed feasible to 
search) within each of the three grassland areas of 
the base. Plot distribution was: six in Test Site 
(76 ha), four in Jump Circle (109 ha), and six in 
Westfield (71 ha). We searched two to three plots 
each weekday for ~2 hrs per plot beginning on 15 
April (2009) and 28 April (2010). Searching 
involved one to three observers walking parallel 
transects and agitating vegetation with 2-m 
bamboo poles to flush nesting birds. Plot visits 
rotated so that each plot was searched at least 
once every 1-2 weeks. 
Geographic coordinates of located nests were 
obtained using a global positioning system, and 
two small pieces of pink flagging were placed on 
vegetation 2-3 m from the nest (i.e., creating a 
line with the nest at the center). Flagging was 
inconspicuous among surrounding vegetation, and 
intended to aid in relocation at close range; we do 
not believe that it drew the attention of predators. 
Nests were generally checked every 2-3 days 
until fledging, failure, or until young could no 
longer be located. Five check intervals for three 
nests exceeded 3 days due to logistical constraints, 
four intervals were 4 days and one was 5 days. 
Nest searching concluded on 15 July both years, 
although all active nests at that time were 
monitored until completion. 
Young nighthawks arc semi-precocial and te 
to move from the original nest site before fledgi 
with movements that appear to increase 
distance and frequency with age (Fowle 19< 
DeXt ? r J 952 ; MCA, pers. obs.). We thorougl 
searched the area within -30 m of the nest (or i 
last Io cat, 0n , which chjcks wcrc ohscrved 
in C,, ' P,y " CStS - *» follow 
in most cases by at least one subsequent sear 
during the next nest-check. Typically" youngT 
days post-hatch were easy to locate as they we 
invariably < 1 m from the original nest site, while 
older young could not always be found. Thus, wc 
calculated nest survival and predation raio. 
through hatch-date only (i.e.. success = hatching 
This is also the approach taken by Perkins am! 
Vickery (2007), who also worked in grassland 
habitat. Daily nest survival rates and confidence 
intervals were calculated using the logistic nest 
survival model within Program MARK (While 
and Burnham 1999. Dinsmore et al. 2002). Ths 
program was also used to evaluate whether or not 
daily nest survival rate varied over the course 
of the season. We used the likelihood ratio lest 
(Shaffer and Thompson 2007; alpha = 0.05) to 
evaluate model performance versus the null : i.e. 
constant survival or 'no effect’) model. Nesi 
failures were classified as either abandonment or 
predation based on timing and evidence at the 
nest. Daily predation rates were measured bv 
calculating the daily survival rate based on 
predation failures only, and subtracting this value 
from one. 
Clutch Size and Movements of Young .— 1 Clutch 
size calculations were based only on active nests 
that were visited at least twice prior to hatching to 
avoid uncertainties associated with mobile young 
Observations made during checks after hatching 
were used to generalize pre-fledging movements 
of young from the original nest site. 
Nest-site Characteristics and Habitat Selection 
We measured maximum vegetation height (cm), 
after finding each nest, at which vegetation touched 
a meter-stick at five locations: at the nest and 0.5 n 
from it in each of the cardinal directions. Weal? 1 ’ 
measured litter depth (i.e., unrooted, dead vegeta¬ 
tion) in 2010 at the same five locations. The mean 
value of the five measurements was used " 
subsequent calculations and analyses. We visual!) 
estimated the percent cover (±5%). alter noai - 
attempts were completed, of four vegetation caie 
gories within a 1 X 1-nr quadrat centered on tfc 
nest: (I) grasses (including rooted live and dc. 
grasses), (2) forbs, (3) shrubs (woody perennial- 
and (4) open (including sand. lichens, mosses. 
lilter). Only nests found prior to hatching were us* 
in vegetation calculations, as young found afio 
hatching may have wandered from the original " 
We also measured vegetation characteristics h 
2010 at 80 randomly-generated points within '• 
16 plots searched for nests to better assess habiM 
preferences. Points were generated in a geograP hK 
information system, and constrained to be at k 1 ' 1 
