240 
THE WILSON JOURNAL OF ORNITHOLOGY • Vo/. 123. No. 2. June 20/1 
TABLE 4. Mean ± SE for DBH (cm) and number of stems (counts) of species discriminating successful from 
unsuccessful nests of Mariana Crows, 1997-1999. 
Measure 
Species 
Successful nests 
Unsuccessful nests 
Mean DBH 
Ficus tinctoria 
18.69 ± 5.56 
32.96 ± 7.35 
Mean DBH 
Ochrosia mariannensis 
16.17 ± 4.83 
9.10 + 4.72 
Mean DBH 
Polyscias grandifolia 
10.47 ± 4.08 
7.79 ± 3.88 
Mean DBH 
Pouteria obovata 
56.03 ± 10.04 
23.64 ± 10.07 
# Stems 
Psychotria spp. 
6.36 ± 2.31 
1.44 ± 0.47 
DISCUSSION 
Eugenia reinwardticina , Neisosperma oppositi- 
folia, fntsia bijuga , and Guettarda speeiosa were 
chosen significantly more often than 16 other 
species of trees used by Mariana Crows for 
nesting. All trees species used for nesting are 
native to Rota. This finding could be due to the 
higher proportion of these commonly used species 
in the forests of Rota, and does not necessarily 
reflect a preference for these species. There were 
no significant differences between nests and 
random plots in elevation, distance or direction 
to edge, distance or direction to road, slope, or 
total number of stems. 
Mariana Crows chose nest sites that: had 
greater canopy cover, larger stems of papaya 
{Carica papaya ), contained species associated 
with limestone forests, and had larger woody 
vines than random sites. Mean canopy\'ovcr was 
about 10 percentage points greater for nest than 
random sites. Nests were high in the canopy 
(mean ± SE nest height = 6.6 + 0.3 m) but were 
placed, on average, 2 m below the crown (nest 
tree height = 8.7 ± 0.3 m). Higher canopy cover 
may reflect efforts to hide the nest from predators 
protect the nest from the adverse effects from sun, 
*' ,nd dama ge from typhoons (Bowman 
and Woolfenden 2002; J. C. Ha. unpubl. data). 
Nest sites were also associated more frequently 
with larger papaya stems. Mariana Crows have 
been observed eating wild papaya (R. R. Ha, pers 
°»ce ,?fT Ver - “ iS U " ClCar " hether lhe 
“es wi h P Paya ,rees SUggeS,S crows choose „e sl 
tes with papaya nearby or whether papaya trees 
““T due lo ,he — P y r 
papay 4 by crows and subsequent seed dispersal in 
the nesting territory. We believe the latter may be 
habSr n, e r eaUSe ClOSed ““W are poor 
' ypically ^ 
(Raiiinr jr- ” exposure to sunlight 
be active 1 an I nChart l99l) - Wood y ^nes may 
be actively selected by Mariana Crows. Lusk and 
Taisacan (1996) found that one particular native 
vine, Marianne jasmine (Jasmmum marianum 
comprised 84% of the materials of nest platforms, 
and other vines and twigs were also used. Nest 
trees were >300 m from buildings suggesting 
Mariana Crows generally prefer to nest away from 
human disturbance. Baker (1951) and (NRC 
1997) indicated Mariana Crows avoided human 
habitation. Roads on Rota may not be as 
disturbing as houses to Mariana Crows. Rota only 
has 3,283 residents (U.S. Department of Com¬ 
merce 2000 ) and vehicle traffic is light. 
Mariana Crows were more likely to have a 
successful nest on sites within the closed-canopy 
limestone forest with smaller Ficus tinctoria, 
greater stem densities of Psychoma spp. and 
larger Ochrosia mariannensis, Polyscias grand- 
ifolia, and Pauteria obovata. Ficus tinctoria is 
associated with the beach or back strand vegeta¬ 
tion, and was linked with unsuccessful nests. The 
other vegetation associated with successful nests 
is reflective of limestone forests (Raulerson and 
Rinehart 1991). Five of the six study sites were 
logged by the Japanese between 1917 and the end 
of WWll and the forest we surveyed is known to 
be 50-90 years of age. Fosberg (I960) reported 
that much ol Rota was cleared between 1932 and 
1935; by 1946. only one-fourth of the island was 
covered by “well-developed"* forest. Older lime¬ 
stone forest growing on the steep slopes of the 
Sabana tends to have a much more open 
understory and much lower densities of Manana 
Crows (Fancy et al. 1999, Plentovich et aL2005) 
We do not believe that vegetation species linked 
with successful nests necessarily convey unique 
species-specific advantages to Mariana Crow 
productivity. Collectively, they describe forest 
composition and structure that is consistent with a 
maturing, undisturbed, native limestone forest m 
the Mariana Islands. If the forest is too young, tire 
canopy is too open and may not provide 
Protection from typhoons, sun, (Bowman and 
