* 
Townsend et a !. • BICKNELL’S THRUSH IN WINTERING AREAS 
369 
predictors, weighted by sample size, specifying 
binomial errors and logit-link function. 
RESULTS 
The island-wide sex ratio for all sites combined 
was significantly male-biased (0.64. x 2 i = 9.7, 
P = 0.002), but did not differ from the proportion 
male expected (0.67) based on the sex ratio 
observed in breeding areas (x 2 i = 0.6, P = 0.45). 
The sex ratio varied among geographic areas with 
two sites male biased (Pueblo Viejo x*i - 6.8, 
P - 0.009; Cordillera Central r, = 5.4. P = 
0.02) and five sites not varying significantly from 
50:50 (Table 1). The number of shrubs/ha in a 
GLM was a significant predictor of the proportion 
male (x 2 i = 4.24, P = 0.04) whereas all other 
vegetation characteristics showed no significant 
relationship (small trees: x z i — 0.47, P = 0.49; 
medium trees: x'i = 1.5, P — 0.22; large trees: 
T\ - 0.26, P = 0.61). The proportion male at 
three sites with >10,000 stcms/ha was signifi¬ 
cantly male-biased (x'i = 13.8, P < 0.001), 
whereas the proportion male did not differ from 
Parity (X z i = 0.9, P - 0.35) at four sites with 
<8,000 stems/ha (Table 2). There was little 
variation in non-shrub-layer vegetation character¬ 
istics between sites with one notable exception. 
Total basal area of trees at Los Haitises was 39- 
56% lower than at other sites, reflecting the 
historical and on-going deforestation impacts of 
shifting agriculture on this area (Table 1). Pro¬ 
portion male did not vary significantly with 
elevation (F = 0.03, P = 0.9). 
The age ratio was significantly adult-biased at 
two sites (Cordillera Central: 1.0, x 3 i = 15, P < 
0.001; Los Haitises: 0.79, x 2 i = 8.2, P = 0.004). 
The proportion adult for all sites combined was 
0-73. None of the vegetation classes in a GLM 
with proportion adult as the response variable had 
significant predictive value (shrubs: x : i = 2.52, P 
~ 0.11; small trees: x : i = 114, P = 0.29; 
medium trees: x : i = 2.99. P = 0.08; large trees: 
Vi = 3.05, P = 0.08). Proportion adult did not 
var y with elevation (F = 0.3, P = 0.6). 
DISCUSSION 
Our results indicate the island-wide sex ratio of 
wintering Bicknell's Thrush is equivalent to the 
male-biased sex ratio documented in breeding 
areas (Townsend et al. 2009). This suggests a 
year-round male-skewed sex ratio of —2:1 for 
Bicknell's Thrush and raises the question of how 
•his might affect the overall population dynamics 
> S 
-o 
• e 
C oo m -© O -o 
-3- rf OO — O' — <N 
+ 1 +1 +1 +1 +1 +1 +1 
3 n h O' - O' ri 
OO r*~, rf O' rj- — 
2 CO T) -- r^i r- 
+1 +1 +1 +1 +1 +1 
O oc c*~, O O u-. 
- W r, IT, DC IT. 
— r- — Vi — O' 
+1 +1 +1 +1 +1 +1 
oo O r. oc <r. i/-. 
r~- rr o oc v. 
T O ‘/'i 'C -a- 
awei 
-J 
"'2 — 0 
, . . . -rt r~ 
— cn -r- r<-, 
+| +1 +1 +1 +1 +1 
rl u-, 
g N JO 9 C f 
o’ —’ r 
-c -c r-~ 
8 3 3 : 
cc oo oo r 
1 « ^ M t 
— r-l — - 
§ E C r 
c s EE 
3 & * 'i 
U S ai ifl U 
Los Haitises 24 0.63 0.79 50 2,020 ± 274 365 ± 57 55 ± 9 
