The Wilson Journal of Ornithology 123(2):236-242, 201 1 
NEST SITE SELECTION AND CONSEQUENCES FOR REPRODUCTIVE 
SUCCESS OF THE ENDANGERED MARIANA CROW 
(CORVUS KUBARYT) 
RENEE ROBINETTE HA, 1 - 5 JOHN M. MORTON, 2 * JAMES C. HA, 1 
LAINIE BERRY,’ AND SHELDON PLENTOVICH 4 
ABSTRACT.—Reasons for the decline of the Mariana Crow (Con us kubaryi) on the Western Pacific island of Rolaare 
currently unknown, but a need to protect nesting habitat has been suggested. We examined 55 actual nest sites and 60 
random sites from 1997 to 1999 to investigate habitat characteristics specific to crow nest .sites. Both nests and random plots 
were predominantly in limestone forest habitat. Discriminant function analyses indicate actual nest sites were differentiated 
from random sites based on a higher percentage of canopy cover and mean DBH of papaya (Carica papaya) and wood;, 
vines, as well as a higher stem count ot species associated w ith limestone forests This resulted in correct classification of i 
potential site as nesting versus random in 92% of the cases. Actual nests were >300 m from buildings, while random site- 
averaged (± SE) 226,7 ± 71.6 tn from a building. Actual nest sites were about twice as far from a road as random nest sites 
Twenty-eight ol the 55 active nests fledged young. Nests in native forests were associated with higher reproductive sucres 
than nests in more disturbed areas. These findings suggest that damage to habitat from anthropogenic or natural causes nw; 
be limiting nesting success. Received 19 February 2010. Accepted 22 November 2010. 
Some species of Corvidae appear to select nest 
sites to reduce risk of predation (e.g„ Hooded Crow 
[Corvus comix], Loman 1979; Florida Scrub-Jay 
[Aphelocoma coerulescens J, Bowman and Wool- 
fenden 2002), while others may select nest sites to 
minimize exposure of the nest to harsh weather 
conditions (Pinyon Jay [Gymrtorhinus cyanocepha- 
lus], Baida and Bateman 1972; Brown Jay |P.v/- 
lorhinus mono ], Lawtoti and Lawton 1980). These 
different preferences represent adaptive responses 
to environmental variables (Bowman and Wool- 
fenden 2002) to maximize fecundity. 
The Mariana Crow (Corvus kubaryi) is listed as 
an endangered species (USDI 1984) and originally 
occurred only in the Mariana archipelago on two 
adjacent islands. Rota and Guam. It has been 
extirpated on Guam by the brown tree snake 
(Boiga irregularis) (Wiles et al. 2003). Mariana 
Crows can re-nest and have an extended breeding 
season, but the peak season appears to be October 
through March (Lusk and Taisacan 1996). They 
occasionally build multiple dummy nests prior to 
incubation, lay 1-4 eggs, and typically fledge 1-2 
young (Morton et al. 1999). Both adults care for 
1 Animal Behavior Program, Department of Psychology. 
University of Washington. Seattle, WA 98195 USA 
Refm» S ai ' d Wi ' dlifC Scrvicc ' Kenai National Wildlife 
Refuge, Soldotna, AK 99669, USA. 
Box D 37 P 7oT nt D MarinC Und Wikl,ife ^sources, P. O. 
x . 0 , p a g 0 Pago, American Samoa, 96799 
lulu. mS™ usr'° sy ’ University of Hawaii - H °"°- 
s Corresponding author; e-mail: robinet@uw.edu 
the young and defend the nesting territory: there is 
no evidence of cooperative breeding in this 
species (Morton et al. 1999). Little is known 
about characteristics of their preferred nest sites 
except they appear to favor primary and second¬ 
ary limestone forest (Plentovich et al. 2005). 
Most studies comparing nesting and random 
sites focus on proximate cues associated with 
nesting habitat, while those that use measures of 
reproductive success associated with nesting 
characteristics explore the evolutionary implica¬ 
tions of nest site selection (Krebs 1994, Dunkd 
al. 1997). Analysis of nest site selection of the 
Mariana Crow has implications for conservation 
ot its habitat; an analysis of the fitness implica¬ 
tions ol habitat selection can reveal whether sonic 
selections confer higher reproductive success 
(Clark and Shutler 1999). Nest site characteriza¬ 
tion involves comparing sites where birds build 
nests to the habitat available for nesting. This 
level of analysis does not necessarily reveal 
optimal nesting habitat, as some birds may 
suboptima] sites due to lack of available habitat or 
inexperience (Clark and Shutler 1999). Ir is also 
necessary to compare successful and unsuccessful 
nest sites (successful nests produce fledglingsUu 
discern whether nest sites are characterized by 
what is available or what is optimal for nesl 
success. Our objectives are to: (1) report nest site 
selection by Mariana Crows using active nest sites 
and random sites within the same study areas, and 
(2) examine how successful and unsuccessful nesl 
sites differ. 
236 
