80 
THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 123, No. 1, March 2011 
N 
I 
o 
c 
0 
3 
O' 
0 
i- 
II 
N 
I 
> 
O 
c 
0 
3 
O’ 
0 
Li. 
H 1 ; h ' 1 h 11 1, 
f | i! 1 i U 1 ' 
H M ; : i 11 ! 1111 
iiiiijmtfMt 
! f 111 i i I < i 11 j i 
nmtrmfGf 
mttnmmt 
[ f i ' 1 * 1 1 | 1 I 1 f » 
* 
I I I 
0 1 2 3 4 
Time (sec) 
B 
Time (msec) Amplitude (dB) 
Toril/ G R 3 'n A : ° sclll °g ram (above) and spectrogram (below) of a variant form of the characteristic song of the Cuban 
y- . e ailed spectrogram (left) and power spectra (right) of the central notes of the song marked in A with an asterisk. 
ranged between 0.0 to 38.5%. Individuals from 
Pinai del Rio were statistically reassigned to Isla 
de la Juventud, while individuals from Matanzas 
and more eastern provinces were statistically 
teassigned to Guantanamo. Two individuals from 
Matanzas were reassigned to Pinar del Rio and 
one from Ciego de Avila was reassigned to Isla de 
la Juventud (Table 2; Fig. 5). A MANOVA 
showed the model was significant (Wilks’ X = 
0.099, X 2 = 243.3, df = 42, P = 0.000) and that 
94.3% of the variation was explained by the two 
discriminant fiinctions (Fig. 5). The proportion of 
correctly classified vocalizations by cross-corre¬ 
lation was 61.2%. The first discriminant function 
which explained 86% of the variation, was more 
influenced by interval between notes and number 
of notes, while the second function was influenced 
by interval between songs and peak frequency 
(Table 3). 
DISCUSSION 
Vocal Repertoire of the Cuban Tody .—The 116 
birds studied across the Island of Cuba emitted 
only three different types of calls. Two of these 
sounds, the characteristic song of the species and 
the sound presumably produced with the wings, 
have been previously described onomatopoeically 
(Raffaele et al. 1998, Garrido and Kirkconnell 
2000). The variant form of the characteristic song 
had not been previously reported. Its similarities 
to the characteristic song could have been 
sufficient to consider both as the same by 
