Brand and Noon • SHRUB-NESTING BIRDS IN A RIPARIAN CORRIDOR 
51 
TABLE 1. Reproductive parameters (mean ± SE d ) for Abert’s Towhee, Yellow-breasted Chat, and Bell’s Vireo in 
saltcedar (SC), cottonwood (CW), mesquite (MQ), and across all vegetation types. 
Species 
Habitat 
n a 
% nest survival w 
Clutch size d 
Fledglings 
per nest d 
Fledglings per 
successful nest d 
Seasonal 
fecundity cd 
Abert’s 
SC 
36 
30 
9 
2.80 
± 
0.11 
1.03 
± 0.22 
2.13 
0.22 
1.23 
± 0.43 
Towhee 
CW 
45 
29 
8 
2.79 
-t- 
0.09 
1.19 
± 0.20 
2.43 
+ 
0.13 
1.32 
± 0.40 
MQ 
66 
30 
6 
2.81 
0.09 
1.07 
± 0.17 
2.50 
0.15 
1.43 
± 0.36 
ALL 
147 
30 
4 
2.80 
0.05 
1.09 
± 0.11 
2.38 
0.09 
1.34 
± 0.27 
Yellow 
SC 
25 
39 
+ 
11 
3.31 
0.18 
1.13 
± 0.25 
2.17 
0.21 
0.60 
± 0.18 
-breasted 
CW 
71 
43 
7 
3.17 
0.08 
1.16 
± 0.20 
2.64 
± 
0.22 
0.81 
± 0.16 
Chat 
MQ 
58 
41 
+ 
7 
3.56 
0.09 
1.31 
± 0.21 
2.62 
0.20 
0.77 
± 0.16 
ALL 
154 
42 
4 
3.36 
+ 
0.06 
1.20 
± 0.13 
2.54 
■± 
0.29 
0.75 
± 0.14 
Bell’s 
SC 
11 
9 
8 
2.86 
+ 
0.31 
0.09 
± 0.09 
1.00 
± 
0.00 c 
0.10 
± 0.09 
Vireo 
CW 
15 
34 
13 
3.10 
0.18 
1.07 
± 0.32 
2.29 
0.18 
0.66 
± 0.27 
MQ 
60 
12 
4 
3.13 
0.13 
0.47 
± 0.14 
2.33 
± 
0.28 
0.28 
± 0.10 
ALL 
86 
14 
± 
3 
3.09 
+ 
0.10 
0.54 
± 0.11 
2.14 
± 
0.19 
0.31 
± 0.09 
d n = total nest sample size. 
b From Brand et al. 2010. 
c Seasonal fecundity = mean number of female offspring (number of fledglings divided by 2) successfully fledged per adult female per year. 
d 90% Cl = mean ± 1.645 x SE. 
e SE = 0 as only one of 11 Bell’s Vireo nests found in saltcedar was successful. 
(Molothrus ater) may nest more frequently. We 
used average nesting attempts weighted by the 
proportion of nests parasitized by vegetation type 
from Brand et al. (2010) to account for this 
variation, and assumed that Bell’s Vireos nested 
1.7 ± 0.2, 2.0 ± 0.1, 2.2 ± 0.2, and 2.0 ± 0.2 
times (mean ± SE) in cottonwood, mesquite, 
saltcedar, and across all vegetation types, respec¬ 
tively (Budnik et al. 2001). 
We used estimates of annual adult survival 
from the literature specific to the Southwest. The 
estimates were Sa = 0.574 ± 0.072 for Bell’s 
Vireo (mean ± SE), Sa = 0.518 ± 0.028 for 
Yellow-breasted Chat, and Sa = 0.486 ± 0.126 
for Abert’s Towhee obtained from 6, 18, and 5 
Monitoring of Avian Productivity and Survivor¬ 
ship (MAPS) stations in the southwestern U.S., 
respectively (DeSante and Kaschube 2006, Mich¬ 
el et al. 2006). Juvenile survival estimates are 
lacking for these species, and we assumed that 
annual survival of juveniles was half that of adults 
(Temple and Carey 1988, Donovan et al. 1995, 
Budnik et al. 2000). Use of the same adult and 
juvenile survivorship estimates for the different 
vegetation types will tend to underestimate 
differences in X between them. 
We estimated the standard error of |3 and X 
using the delta method (Armstrong et al. 2002, 
Powell 2007). We assumed covariances = 0 since 
S^, ns, and a were estimated with different data. 
Only one successful Bell’s Vireo nest was 
observed in saltcedar (thus SE = 0), and we 
substituted the estimated standard error of the 
number of fledglings per successful nests ob¬ 
served in mesquite (SE = 0.28) to represent 
maximum observed variability in the system for 
that species for the delta method estimation of 
SE(P) and SE(?Q. We set a = 0.10 to minimize 
the probability of a Type II error and interpreted 
meaningful differences in reproductive parame¬ 
ters among vegetation types in terms of non- 
overlapping 90% confidence intervals. 
The population for X> 1 was considered to be a 
potential source of emigrants and the population 
for X=\ was considered stable (Pulliam 1988). 
The population for X,<1 was considered to 
demonstrate characteristics of a population sink 
(Pulliam 1988, Battin 2004). We set a = 0.10 to 
minimize the probability of a Type II error and 
interpreted results with 90% confidence intervals 
(Cl) around We also separately estimated adult 
survivorship, seasonal fecundity, and the average 
number of nesting attempts required to obtain a 
stable population (X = 1) with other vital rates 
held constant. We used this approach to assess 
whether estimated rates were within 90% confi¬ 
dence intervals of what was observed or assumed, 
and to provide insight into what survivorship 
levels would be needed to conceivably maintain a 
stable population on the San Pedro. We conducted 
a sensitivity analysis to assess what levels of 
parameters drawn from the literature—adult 
