838 
THE WILSON JOURNAL OF ORNITHOLOGY • Vol 123, No. 4, December 2011 
TABLE 2. Mean ± SD and potential for information coding (PIC) values for 25 variables measured for five to 20 wee- 
ooo songs for 17 male Eastern Wood-Pewees. T = temporal measures (sec); TP = temporal proportion measures: F = 
frequency measures (kHz); and FM = frequency modulation measures (kHz). CV b = between individual coefficient of 
variation; CV W = within individual coefficient of variation. PIC = CVVCV W . 
Variable 
T1 (A to F) 
T2 (A to C) 
T3 (C to E) 
T4 (E to F) 
T5 (A to B) 
T6 (B to C) 
T7 (C to D) 
T8 (D to E) 
T9 (B to D) 
TP1 (T2-T6/T1) 
TP2 (T2/T1) 
TP3 (T2 + T7/T1) 
FI (A) 
F2(B) 
F3 (C) 
F4 (D) 
F5 (E) 
F6 (F) 
FM 1 (A to B) 
FM2 (B to C) 
FM3 (C to D) 
FM4 (D to E) 
FM5 (E to F) 
FM6 (A to F) 
FM7 (D to F) 
Mean ± SD 
aw 
CV„ 
CV„ 
PIC 
0.96 ±0.15 
97.8 
15.78 
5.24 
3.0 
0.38 ± 0.05 
83.3 
12.95 
4.60 
18 
0.10 ± 0.01 
24.6 
13.31 
6.69 
20 
0.48 ±0.13 
76.6 
27.25 
9.75 
2.8 
0.18 ± 0.03 
28.9 
16.21 
8.12 
2.0 
0.20 ± 0.04 
98.7 
17.79 
6.48 
2.7 
0.04 ± 0.01 
4.5 
29.08 
23.37 
1.2 
0.06 ± 0.01 
13.3 
16.50 
11.31 
15 
0.24 ± 0.04 
59.3 
16.36 
6.64 
2.5 
NA 
NA 
17.99 
8.93 
2.0 
NA 
NA 
15.14 
6.29 
2.4 
NA 
NA 
14.73 
6.52 
2.3 
3.21 ± 0.20 
21.8 
6.33 
3.74 
1.7 
4.72 ± 0.26 
244.1 
5.46 
1.12 
4.9 
3.80 ±0.16 
97.1 
4.07 
1.36 
3.0 
6.20 ± 0.50 
64.3 
8.10 
2.99 
2.7 
3.24 ±0.12 
77.0 
3.58 
1.29 
2.8 
2.78 ±0.17 
25.0 
6.05 
3.47 
1.7 
1.51 ± 0.27 
34.9 
18.02 
9.01 
2.0 
0.92 ± 0.23 
170.3 
25.15 
6.93 
3.6 
2.41 ± 0.48 
61.2 
20.12 
7.61 
2.6 
2.96 ± 0.54 
77.1 
18.31 
6.08 
3.0 
0.45 ±0.13 
10.3 
28.57 
23.56 
1.2 
0.43 ±0.18 
10.9 
42.80 
39 25 
1.1 
3.42 ± 0.55 
56.7 
16.18 
6.42 
25 
F-values for one-way ANOVAs comparing between and within individual 
variation for each song variable: all P < 0.0001. 
Thus, the song being classified is not used to both 
develop the linear function and to test that 
function. We used one-way ANOVAs to compare 
within and between individual variation for each 
song variable. We conducted all statistical testino 
using SPSS Version 17 statistical software (SPSS 
RESULTS 
All 25 variables measured for pee-ah-wee an< 
wee -000 songs had PIC values >1 (Tables 1.2 
PIC values for pee-ah-wee song variables wer 
similar for temporal and frequency modulatioi 
measures but higher for frequency measure 
(ANOVA: F 2a o = 5.21, P = 0.01; Tukey HSI 
Post-hoc test) (Table 1). PIC values were simila 
tor a" wee _ 0 ng variab|e categories (AN 
samni " °' 74 ' P = °' 49 > (Table 2). Tin 
of temporal proportion measures was tor 
Sr S r TT UVe anal - Vsis - P1C values wen 
* C ' for P ee -»h-'vee rhun wee -000 songs whe, 
all variables were combined (/-test: (47 = 2.62, 
P = 0.01). 
Principal component analysis of the 25 pee-ah- 
wee song variables generated six principal 
components w ith eigenvalues > 1.0 that explained 
88.4% of the variance. We used those principal 
components as the basis for a discriminant 
function analysis and correctly assigned 97 
of pee-ah-wee songs (315/323) to the correct 
individual. Principal component analysis of the 25 
wee -000 song variables generated seven principal 
components with eigenvalues > 1.0 that explained 
88.9% of the variance. We used those principal 
components as the basis for a discriminant 
Junction analysis and correctly assigned 95.0'' 
of wee -000 songs (191/201) to the correct 
individual. 
DISCUSSION 
Each recent study of suboscine song, including 
ours, found sufficient variability to demonstrate 
