Ha et al. • MARIANA CROW NEST SITE SELECTION 
239 
TABLE 2. Trees species used as nest sites by Mariana 
Crows. 1997-1999. 
Scientific name 
Common name 
Number 
% 
Neiosperma oppositifolia 
Fagot 
28 
22.6 
Eugenia reinwardtiana 
A’abang 
19 
15.3 
Intsia bijuga 
Ifit 
17 
13.7 
Guettarda speciosa 
Zebrawood 
13 
10.5 
Premna obrusifolia 
False elder 
9 
7.3 
Pouteria obovata 
Lalaha 
8 
6.4 
Guamia mariannae 
Paipai 
6 
4.8 
Ficus spp.* 
Fig 
4 
3.2 
Pisonia grandis 
Umumu 
4 
3.2 
Unknown 
3 
2.4 
Psychotria mariana 
Aploghating 
3 
2.4 
Cymmetra ramiflora 
Gulos 
2 
1.6 
Drypetes dolichocarpa 
Mwelel 
2 
1.6 
Maytenus ihumpsonii 
Lulujut 
2 
1.6 
Barrintonia asiatica 
Fish-kill tree 
1 
0.8 
Elaeocarpus joga 
Yoga 
1 
0.8 
Macaranga thompsonii 
Pengua 
1 
0.8 
Tristiropsis obtusangula 
Faniok 
1 
124 
0.8 
100.00 
’’ Two species (Ficus nnctoria and F. prolixa ) occur on Rota, 
nest tree was 7.8 m with a DBH of 0.17 m and 
slope of 10.8°. The mean nest height was 5.9 m 
and the nest was on a branch with a mean branch 
length of 1.7 m. 
Comparisons Between Active Nests and Ran¬ 
dom Plots .—Actual nests were >300 m from any 
buildings, and no distance was recorded for actual 
nests; random sites had a mean distance of 226.7 m 
from a building. Nests were a mean distance of 
223.9 m from a road, while random plots were 
133.5 m from roads on average (/ 74 =1.48, P = 
0.14). There was no significant difference in 
univariate /-test comparisons between nest and 
random sites for the distance from an edge (/ 73 = 
0.32, P = 0.75). There were also no significant 
differences between nest and random sites for 
characteristics of elevation Ugo = 0.46. P = 0.65), 
or slope (7 109 = 0.65. P - 0.52). 
Discriminant Function Analyses. —Forty-nine 
species of trees and two categories (snags and 
woody vines) were recorded in vegetation sur¬ 
veys. Discriminant function analyses indicated 
which variables (Table 3) resulted in a correct 
classification of potential habitat as an active nest 
site versus a random site in 92% (95/103) of cases 
(^ 21,81 = 16.97, P < 0.001), Active nests that 
were successful versus those not successful 
differed (F SM = 10.12, P < 0.001; Table 4). 
This resulted in correct classification of potential 
habitat as successful versus unsuccessful in 85% 
(34/40) of the cases. Slope and some vegetation 
measures varied between study sites (F 70 ,404 = 
9.56, P < 0.001; Table 5), and resulted in correct 
classification of each study site in 71% of these 
cases. 
TABLE 3. Mean ± SE for DBH, and number of stems of tree species discriminating active nests of Mariana Crows 
from random sites. Average nest cover is in percent, DBH is in cm, and the number of stems represents counts by species. 
Measure 
Variable 
Active nests 
Random sites 
Average nest cover 
93.59 ± 0.95 
83.69 ± 3.12 
DBH 
Carica papaya 
9.6 ± 2.10 
0.00 
DBH 
Cycas circinalis 
5.09 ± 3.56 
17.30 ± 6.27 
# Stems 
Eugenia spp. 
37.45 ± 7.00 
24.50 ±4.19 
DBH 
Eugenia spp. 
25.79 ± 2.99 
30.34 ± 2.30 
DBH 
Eugenia thompsonii 
1.03 ± 1.03 
2.16 ± 1.23 
Mean DBH 
Exocecaria agallocha 
1.56 ± 1.56 
0.00 
# Stems 
Guamia mariannae 
33.36 ± 8.06 
19.92 ± 4.07 
Mean DBH 
Guettarda speciosa 
66.74 ± 11.01 
72.25 ± 8.85 
# Stems 
Pandanus spp. 
7.30 ± 1.54 
19.02 ± 4.23 
Mean DBH 
Pemphis acidula 
0.00 
2.22 ± 1.65 
# Stems 
Pipturus argenteus 
2.25 ± 0.81 
3.12 ± 0.79 
Mean DBH 
Pipturus argenteus 
12.31 ± 3.27 
30.29 ± 4.95 
# Stems 
Psychotria spp. 
3.82 ± 1.21 
6.75 ± 1.52 
Mean DBH 
Psychotria spp. 
38.94 ± 6.28 
48.33 ± 5.33 
# Stems 
Randia cochinchinensis 
1.20 ± 0.27 
1.33 ± 0.32 
Mean DBH 
Randia cochinchinensis 
29.73 ± 4.48 
20.77 ± 3.81 
Mean DBH 
Snags 
35.49 ± 6.29 
0.00 
Mean DBH 
Tarenna sambucina 
4.34 ±2.11 
2.86 ± 1.41 
Sum 
Sum # Stems 
233.97 ± 20.59 
252.62 ± 16.40 
