Bird-song functions, such as mate attraction, species recognition, and territory defense, are closely linked to individual fitness. Thus, habitat characteristics that affect song transmission and degradation should exert a strong influence on the evolution of song attributes. Whereas different acoustic environments may influence the evolution of song characteristics, factors such as body size and evolutionary history of taxa are expected to constrain the amount of environmental variation in song properties. In the present study, we controlled for phylogeny and examined the effects of body mass and habitat structure on variation in song structure of 30 taxa of Phylloscopus and Hippolais warblers, which are a closely related group of birds that occupy a wide variety of habitats and show high variation in vocalizations. Habitat structure was strongly correlated with temporal characteristics of songs but not with most of the frequency-related attributes that we measured. Only the highest frequencies of songs varied with habitat structure. As predicted, species occupying closed habitats avoided the use of rapidly modulated signals and had song structures that minimized reverberation. Body mass covaried significantly with most of the song attributes. Smaller species used higher frequencies and had more notes in their songs compared with larger species. Received 9 May 1996, accepted 26 July 1996.
Division of Biological Sciences and Montana Cooperative Wildlife Research Unit,
University of Montana, Missoula, Montana 59812, USA
THE MAJOR FUNCTIONS OF BIRD SONG, such as
mate attraction, species recognition, and the es-
tablishment and defense of territories, are close-
ly linked to individual fitness (Catchpole and
Slater 1995). Thus, songs that maximize effec-
tive communication of signals to intended re-
ceivers should increase the fitness of the sing-
ing individual. Concurrently, songs and acous-
tic communication in general are subject to
spherical-spreading and frequency-dependent
attenuations, reverberations, and irregular am-
plitude fluctuations caused by habitat structure
(Wiley and Richards 1982, Wiley 1991). Thus,
the acoustic environment may exert a strong
influence on the evolution of song attributes.
Previous studies of habitat associations of song
properties led to predictions that lower fre-
quencies should be favored more in closed hab-
itats than in open habitats (Chappuis 1971, Mor-
ton 1975, Cosens and Falls 1984, Waas 1988),
although, theoretically, lower frequencies
should always be favored because they travel
the farthest regardless of habitat (Wiley and
Richards 1982, Wiley 1991). High frequencies
have a greater tendency to be scattered by fo-
liage than do low frequencies, which are af-
fected only by objects larger than foliage (Wiley
E-mail: abadyaev@selway.umt.edu
40
1991). Thus, high frequencies may be more con-
strained by habitat (Wiley 1991). Reverberation
is thought to influence the length of notes (also
referred to as "elements") and the amount of
time between notes (Tubaro and Segura 1995).
Consequently, birds that live in habitats with
numerous scattering surfaces (e.g. leaves and
branches) may: (1) avoid using rapidly modu-
lated signals, (2) use shorter notes, and (3) put
more space between notes to avoid extensive
reverberation (Nottebohm 1975, Wasserman
1979, Richards and Wiley 1980, Gish and Mor-
ton 1981, Handford 1981, Anderson and Conner
1985, Sorjonen 1986, Handford and Lougheed
1991, Wiley 1991).
Song frequencies also may be constrained by
body size because the mass of a vibrating struc-
ture affects the frequency that is produced most
efficiently (Morton 1975, Bowman 1979, Wall-
schager 1980, Shy 1983, Ryan and Brenowitz
1985). Hence, inter-habitat variation in songs
could be confounded by unequal distribution
of different-sized birds among habitat types
(Ryan and Brenowitz 1985). In addition, evo-
lutionary history and relatedness of species can
confound variation in song structures exerted
by habitat parameters (Ryan and Brenowitz
1985), especially if different taxa are unevenly
distributed among habitat types (Wiley 1991).
In this paper we use a phylogenetic approach
Phylloscopus collybita abietinus
-r- SYLLABLE
'" RAN
v
z
o
u. I I I_F
INT
NTLEN
Fig. 1. Sonogram of Phylloscopus collybita abietinus showing variables measured. HF = highest frequency;
INT = internote interval; RAN = frequency range; LF = lowest frequency; NTLEN = note length. A typical
syllable consisting of two notes also is identified.
to examine habitat associations of 30 taxa of
Phylloscopus and Hippolais warblers. These gen-
era are especially well-suited for this study. Most
of the Phylloscopus and Hippolais warblers are
poorly differentiated in plumage characteristics
and morphology but show extensive variation
in vocalizations, even on a subspecies level (e.g.
Dementiev and Gladkov 1954, Cramp 1992; see
Appendix). These warblers are widely distrib-
uted and occupy a broad variety of habitats (De-
mentiev and Gladkov 1954, Cramp 1992). In
addition, the phylogenetic relationships of these
birds, especially Phylloscopus warblers, have been
studied extensively (Ticehurst 1938, Martens
1980, Haffer 1991, Cramp 1992, Richman and
Price 1992, Helbig et al. 1995 ), providing an
opportunity to control for species and subspe-
cies relatedness.
We statistically controlled for phylogeny and
examined the effects of body mass and habitat
structure on variation in song structure of Phyl-
loscopus and Hippolais warblers. Specifically, we
tested the following predictions: (1) frequency
characteristics of song should correlate with
body mass, and high frequencies should vary
with habitat type; (2) temporal characteristics
of songs should be affected by habitat structure.
METHODS
Songs used in the study were obtained from the
song identification guide by Veprintsev (1987). Re-
cordings on these disks contain complete songs and
calls of several individual birds of each species. We
used Canary 1.0 (Cornell Laboratory of Ornithology)
software to measure the following variables (see Fig.
1): (1) highest frequency (HF; kHz), (2) lowest fre-
quency (LF), (3) range of frequency distribution (RAN),
(4) dominant frequency (PEAK; i.e. the most preva-
lent frequency within a song), (5) number of notes
(NT), (6) number of syllables (SYLL), (7) length of
notes (NTLEN), (8) song length (SLEN), and (9) in-
ternote interval (INT). Means of songs of at least three
to four individuals per species or subspecies were
averaged to obtain values given in the Appendix.
We gathered published data on body size and hab-
itat type for all 30 taxa (Dementiev and Gladkov 1954,
Veprintsev 1987, Cramp 1992). We ranked nesting
habitat from open to closed in the following se-
quence: (1) open fields, steppes, and deserts; (2) bush-
es and subalpine bushes; (3) intermediate between
bushes and forest habitats, gardens; (4) coniferous
forests; and (5) deciduous forests. Published data on
habitat distribution were used in conjunction with
habitat types listed on the source of the recordings
(Veprintsev 1987).
The phylogeny of Phylloscopus and Hippolais species
used in this study was constructed by summarizing
available systematics data (Ticehurst 1938, Martens
1980, Haffer 1991, Cramp 1992, Richman and Price
1992, Helbig et al. 1995). Our phylogenetic hypothesis
(see Fig. 2) was largely constructed based on molec-
ular data for Phylloscopus warblers (Richman and Price
1992, Helbig et al. 1995). However, the resulting phy-
logenetic tree is consistent with findings in most oth-
er classification studies as well. Branch lengths were
set as equal because they were available only for a
few species in this study (Richman and Price 1992,
Helbig et al. 1995). Plots of standardized contrasts
against the variances of the untransformed contrasts
showed no significant correlation, justifying the use
of equal branch lengths (Purvis and Rambaut 1995).
To control for species relatedness, we calculated
independent contrasts between nodes for the traits
of interest (Felsenstein 1985) using the CAIC software
package (Purvis and Rambaut 1995). We then exam-
ined the relationships between the variables by cal-
ß e. bre-os
elle
eates
P .
P .beta
H,oa ta
Fig. 2. Phylogenetic hypothesis for the Phylloscopus and Hiolais warblers.
culating linear regressions on these contrasts (Gar-
land et al. 1992, Grafen 1992). All regressions were
forced through the origin (Garland et al. 1992). Anal-
yses of independent contrasts of habitat types were
conducted following Martin and Badyaev (1996). The
numbers of syllables and notes were square-root
transformed; all other song parameters and body mass
were log-transformed before the analyses.
RESULTS
Song characteristics were highly intercorre-
lated (Table !). The taxa we studied used widely
ranging types of syllables in songs, thereby
complicating the direct comparison of original
song variables among species. Therefore, we
used principal components (PC) analysis to de-
scribe variation in song characteristics (Table
!). Principal components were constructed such
that PC I (32% of variation in the model) ac-
counted for most of the variation in frequency-
related variables, and PC II (22.5% of variation)
accounted for most of the variation in temporal
characteristics of songs (Table !). Phylloscopus
and Hippolais warblers of different size were
equally represented across habitat types; body
mass did not vary with habitat openness (Pear-
son r = -0.!, P = 0.6).
Among frequency-related song parameters,
only the highest frequency varied significantly
with habitat openness (Table 2). Frequency pa-
rameters were strongly affected by body mass;
TABLE 1. Pearson correlation coefficients of song properties and eigenvectors for principal components
analyses of song variables (corrected for phylogeny using linear contrasts; Purvis and Rambaut 1995). See
Methods for acronyms of song variables.
SLEN HF LF SYLL NT RAN PEAK INT NTLEN
HF 0.34*
LF -0.10 0.01
SYLL 0.43* 0.50** -0.31
NT 0.59** 0.01 -0.10 0.43**
RAN 0.34* 0.76*** -0.58*** 0.50** -0.03
PEAK -0.00 0.20 0.50** -0.14 -0.06 -0.12
INT 0.00 0.17 0.02 -0.24 -0.49** 0.27
NTLEN 0.25 0.28 -0.02 0.26 -0.35** 0.34*
HT a -0.33* 0.16 0.07 -0.20 -0.51'* 0.15
MASS b -0.07 -0.33* 0.05 -0.17 0.20 -0.36**
PC I c 0.42 0.47 0.50
PC II c -0.60
0.10
-0.06 0.24
0.17 0.25 0.14
-0.28 -0.37** -0.46**
0.53 0.40
*,P < 0.1; **, P < 0.05; ***, P < 0.001.
ß Habitat type.
Body mass (g).
' Only loadings - [0.40[ are shown.
smaller species used the highest frequencies and
had the widest frequency range compared with
larger species (Table 2, Fig. 3). Body mass also
was strongly correlated with temporal charac-
teristics of song; smaller species had shorter in-
tervals between notes and used shorter notes
compared with larger species (Table 2). In con-
trast to the frequency characteristics of songs,
temporal parameters were strongly affected by
habitat structure. Species in closed habitats used
fewer notes and longer notes, and had longer
intervals between notes compared with their
open-habitat relatives (Table 2, Fig. 4).
TABLE 2. Standardized coefficients from multiple re-
gression of song properties on habitat and body
mass (corrected for phylogeny using linear con-
trasts; Purvis and Rambaut 1995).
Sources of variance
Habitat Body
Variable F openness mass
Frequency properties
Highest frequency 3.78** 0.41'* -0.39*
Lowest frequency 0.04 0.03 0.03
Dominant frequency 2.31' 0.37 -0.27*
Frequency range 2.44** 0.14 -0.39**
PC I 3.67** 0.32 -0.39**
Temporal properties
No. of syllables 1.09 -0.19 -0.16
No. of notes 1.69 -0.35 0.29
Internote interval 2.69* 0.22 -0.42**
Song length 0.82 -0.24 -0.04
Note length 3.56** 0.14 -0.48**
PC II 6.22** 0.45** -0.52**
*, P < 0.1; **, P < 0.05.
DISCUSSION
The results support our predictions that tem-
poral characteristics of song vary strongly with
-0.08
Fig. 3.
-0.06 -0.04 -0.02 0.00 0.02 0.04 0.06
BODY MASS 2 GREATER
0.00 0.25 0.50 0.75 1.00 1.25 1.50
HABITAT :: MORE CLOSED
Partial-regression residual plots of inde-
pendent linear contrasts illustrating relationship be-
tween frequency-related characteristics of songs (i.e.
PC I), body mass (upper), and habitat openness (low-
3
2
0
[ -3 -0.06 -004 -0.02 0.i 0.02 0.04 (c) 0.06
BODY MASS GREATER
o.5 o.;5
HABITAT MORE CLOSED
Fig. 4. Partial-regression residual plots of inde-
pendent linear contrasts illustrating relationship be-
tween temporal characteristics of songs (i.e. PC II),
body mass (upper), and habitat openness (lower).
habitat openness, whereas frequency attributes
largely are unaffected by habitat structure. Only
the highest frequencies varied significantly with
habitat structure (Table 2). Closed habitats are
thought to cause greater attenuation in songs,
which in turn would affect the upper band of
acceptable frequencies (Wiley 1991). Thus, only
the highest frequencies are expected to vary
with habitat. The highest frequencies might also
be affected by different levels of background
noise among habitat types (Ryan and Brenowitz
1985). Temporal attributes of song should vary
with habitat because reverberation rises with
increasing habitat closeness (e.g. Wiley and
Richards 1982). As we have shown, birds in
closed habitats avoid the use of rapidly mod-
ulated signals, use shorter notes, and have more
space between notes (potentially to minimize
reverberation). These findings are consistent
with other studies (Wiley 1991, Tubaro and Se-
gura 1995).
Body mass was a significant constraint on most
song attributes (Table 2, Figs. 3 and 4). In par-
ticular, smaller species use higher frequencies
and shorter notes than do larger species (Table
2). This is consistent with the finding that body
size is positively correlated with the mass of the
song-producing structures in birds (Morton
1975, Ryan and Brenowitz 1985). It is thought
that in vocalizations intended for long-range
communication, body mass should strongly limit
the lower range frequency because more energy
is required to produce and transmit low-fre-
quency sounds (Ryan and Brenowitz 1985). Al-
though song structure predictably varied with
habitat type, singing behaviors such as use of
high perches and song-flights also could max-
imize song transmission (Jilka and Leisler 1974,
Wilczynski et al. 1989). However, courtship
song-flights, such as the one of P. sibilatrix, are
very rare among these warbler species. We did
not examine the relationship between singing
height and song structure. These species used
a wide variety of habitats, and taxa with greater
number of species are needed to simultaneously
control for effects of habitat, singing height,
and body mass.
Although habitat structure and body mass
significantly affected song properties in Phyl-
loscopus and Hippolais warblers, a large amount
of variance in song characteristics (46.9%) re-
mained unexplained. These warblers, especial-
ly Phylloscopus, vary widely in such biological
parameters as mating system, territory size, and
population density (reviewed in Cramp 1992).
These factors may determine functions of songs,
i.e. whether songs are intended for long- or
short-range communication, and thus can in-
fluence song structure and composition (e.g.
Salomon and Hemim 1992).
Species recognition frequently is invoked to
explain high variation in characteristics of vo-
calizations, especially in Phylloscopus warblers
(e.g. Salomon and Hemim 1992). The degree of
sympatry varies within Phylloscopus (Cramp
1992); thus, the variation we observed, especial-
ly among subspecies, may be influenced by the
presence of closely related species (Salomon
1989, Salomon and Hemim 1992). Further re-
views and experimental tests of potential mech-
anisms are needed to better address the ecolog-
ical and evolutionary factors causing unusually
high variation in song attributes among closely
related species of Phylloscopus and Hippolais war-
biers.
ACKNOWLEDGMENTS
We thank W. Davison, S. Garner, J. Marks, W. Par-
son, M. Perez, and three anonymous reviewers for
many helpful suggestions. We dedicate this paper to
the memory of B. N. Veprintsev.
LITERATURE CITED
ANDERSON, M. E.,AND R. N. CONNER. 1985. Northern
Cardinal song in three forest habitats in eastern
Texas. Wilson Bulletin 97:436-449.
BOWMAN, R. I. 1979. Adaptive morphology of song
dialects in Darwin's finches. Journal ffir Orni-
thologie 12:353-389.
CATCHPOLE, C. K., AND P. J. B. SLATER (Eds.). 1995.
Bird song: Biological themes and variations. Cam-
bridge University Press, Cambridge, United
Kingdom.
CHAPPUIS, C. 1971. Un exemple de l'infiuence du
milieu sur 1es emissions vocales des oiseaux: L'ev-
olution des chants en foret equitoriale. Terre et
Vie 25:183-202.
COSENS, S. E., AND J. B. FALLS. 1984. A comparison
of sound propagation and song frequency in tem-
perate marsh and grassland habitats. Behavioral
Ecology and Sociobiology 15:161-170.
CRAMP, S. (Ed.). 1992. Handbook of the birds of Eu-
rope, the Middle East and North Africa, vol. VI.
Oxford University Press, Oxford.
DEMENTIEV, G. P., AND N. A. GLADKOv (Eds.). 1954.
Birds of the Soviet Union, vol 5. "Sovetskaya
Nauka," Moscow.
FELSENSTEIN, J. 1985. Phylogenies and the compar-
ative method. American Naturalist 125:1-15.
GARLAND, T., JR., P. H. HARVEY, AND A. R. IVES. 1992.
Procedures for the analysis of comparative data
using phylogenetically independent contrasts.
Systematic Biology 41:18-32.
GISH, S. L., AND E. S. MORTON. 1981. Structural ad-
aptations to local habitat acoustics in Carolina
Wren songs. Zeitschrift fiir Tierpsychologie 56:
74-84.
GRAFEN, A. 1992. The uniqueness of the phyloge-
netic regression. Journal of Theoretical Biology
156:405-424.
HAFFER, J. 1991. Phylloscopus. Pages 1021-1357 in
Handbuch der VOgel Mitteleuropas, vol. 12 (U.
N. Glutz von Blotzheim and K. M. Bauer, Eds.).
Aula Verlag, Wiesbaden, Germany.
HANDFORD, P. 1981. Vegetational correlates of vari-
ation in the song of Zonotrichia capensis. Behav-
ioral Ecology and Sociobiology 8:203-206.
HANDFORD, P., AND S.C. LOUGHEED. 1991. Variation
in duration and frequency characters in the song
of the Rufous-collared Sparrow, Zonotrichia ca-
pensis, with respect to habitat, trill dialects and
body size. Condor 93:644-658.
HELSIG, A. J., I. SEIBOLD, J. MARTENS, AND M. WINK.
1995. Genetic differentiation and phylogenetic
relationships of Bonelli's Warbler Phylloscopus bo-
nelli and Green Warbler P. nitidus. Journal of Avi-
an Biology 26:139-153.
JILKA, A., AND B. LEISLER. 1974. Die Einpassung dreier
Rohrsangerarten (Acrocephalus schoenobaenus, A.
scirpaceus, A. arundinaceus) in ihre Lebensraume
in bezug auf das Frequenzspektrom ihrer Re-
viergesange. Journal fiir Ornithologie 115:192-
212.
MARTENS, J. 1980. LautuOerungen, verwandtschaf-
liche Beziehungen und Verbreitungsgaschichte
asiatischer Laubsnger (Phylloscopus). Advances
in Ethology 22:1-71.
MARTIN, T. E., AND A. V. BADYAEV. 1996. Sexual di-
chromatism relative to nest height and nest pre-
dation: Contributions of females versus males.
Evolution 50:2454-2460.
MORTON, E.S. 1975. Ecological sources of selection
on avian sounds. American Naturalist 109:17-34.
NOTTEBOHM, F. 1975. Continental patterns of song
variability in Zonotrichia capensis: Some possible
ecological correlates. American Naturalist 109:
605-624.
PURVIS, A., AND A. RAMBAUT. 1995. Comparative
analysis by independent contrasts (CAIC): An
Apple Macintosh application for analysing com-
parative data. Computer Application Biosciences
11:247-251.
RICHARDS, D. G., AND R. H. WILEY. 1980. Reverber-
ations and amplitude fluctuations in the propa-
gation of sound in a forest: Implications for an-
imal communication. American Naturalist 115:
381-399.
RICHMAN, A.D., AND T. PRICE. 1992. Evolution of
ecological differences in the Old World leaf war-
biers. Nature 355:817-821.
RYAN, M. J., AND E. A. BRENOWITZ. 1985. The role of
body size, phylogeny, and ambient noise in the
evolution of bird song. American Naturalist 126:
87-100.
SALOMON, M. 1989. Song as a possible reproductive
isolating mechanism between two parapatric
forms. The case of the Chiffchaffs Phylloscopus c.
collubita and P. c. brehmii in the Western Pyrenees.
Behaviour 111:270-290.
SALOMON, M., AND Y. HEMIM. 1992. Song variation
in the Chiffchaffs (Phylloscopus collubita) of the
Western Pyrenees-the contact zone between col-
lybita and brehmii forms. Ethology 92:265-282.
SHY, E. 1983. The relation of geographical variation
in song to habitat characteristics and body size
in North American tanagers (Thraupinae: Piran-
ga). Behavioral Ecology and Sociobiology 12:71-
76.
SORJONEN, J. 1986. Song structure and singing strat-
egies in the genus Luscinia in different habitats
and geographical areas. Behaviour 98:274-285.
TICEHURST, C. B. 1938. A systematic review of the
genus Phylloscopus (willow-warblers or leaf-war-
biers). British Museum, London.
TUBAgO, P. L., AND E. T. SEGUI. 1995. Geographic,
ecological and subspecific variation in the song
of the Rufous-browed Peppershrike (Cyclarhis gu-
janensis). Condor 97:792-803.
VEPRINTSEV, B.N. 1987. Birds of the Soviet Union.
Passeriformes: Leaf warblers and Hippolais war-
biers. "Melodia," Moscow.
WAAS, J. R. 1988. Song pitch-habitat relationships
in White-throated Sparrows: Cracks in acoustic
windows? Canadian Journal of Zoology 66:2578-
2581.
WALLSCHAGER, D. 1980. Correlation of song fre-
quency and body weight in passefine birds. Ex-
perientia 36:412.
WASSERMAN, F. E. 1979. The relationship between
habitat and song in the White-throated Sparrow.
Condor 81:424-426.
WILEY, R. H. 1991. Associations of song properties
with habitats for territorial oscine birds of eastern
North America. American Naturalist 138:973-993.
WILEY, R. H., AND D. G. RICI4ARDS. 1982. Adaptations
for acoustic communication in birds: Sound
transmission and signal detection. Pages 132-181
in Acoustic communication in birds, vol. 1. (D. E.
Kroodsma and E. H. Miller, Eds.). Academic Press,
New York.
WILCZYNSKI, W., M. J. RYAN, AND E. A. BRENOWITZ.
1989. The display of the Blue-back Grassquit:
The acoustic advantage of getting high. Ethology
80:218-222.
Associate Editor: J. S. Marks
APPENDIX. Characteristics of songs (mean values) of Phylloscopus and Hippolais warblers. Song recordings
were obtained from Veprintsev (1987). See Methods for acronym definitions.
Species SLEN HF LF SYLL NT RAN PEAK INT NTLEN
Phylloscopus proregulus 2.192 7.738 2.735 3.7 19.3 5.005 4.35 57.01 94.39
P. i. inornatus 1.433 8.460 3.925 4.5 4.0 4.545 5.39 58.50 129.50
P. i. humei 2.192 8.200 3.400 3.0 7.0 4.800 5.30 53.00 89.00
P. fuscatus 1.160 6.003 2.773 1.3 6.7 3.230 5.16 91.10 118.57
P. griseolus 1.163 5.437 2.607 3.0 9.0 2.830 3.83 32.83 73.27
P. collybita abietinus 2.865 6.810 3.190 3.0 11.0 3.630 5.13 162.00 106.50
P. c. brevirostris 2.898 7.140 2.800 3.0 11.0 4.330 4.52 168.50 105.00
P. c. tristis 5.648 6.775 2.710 3.5 31.0 4.070 4.52 86.52 101.90
P. c. menzbieri 2.556 7.310 3.110 3.0 8.0 4.200 4.26 238.80 105.80
P. c. fulvescens 2.136 7.410 2.760 2.0 7.0 4.650 4.17 267.70 73.00
P. c. lorenzii 2.587 6.900 3.090 2.0 8.0 3.810 5.30 289.00 80.50
P. c. sindianus 3.873 6.230 2.920 3.0 15.0 3.310 4.35 109.00 205.00
P. neglectus 1.113 6.450 2.400 10.0 11.0 4.050 4.35 34.80 69.30
P. t. trochilus 2.513 7.265 2.200 7.0 17.3 5.063 3.65 63.30 134.95
P. t. yakutensis 2.742 6.950 1.880 8.0 18.0 5.070 1.88 68.75 102.50
P. sibilatrix 2.823 7.620 3.070 1.0 29.0 4.550 4.87 98.30 36.30
P. coronatus 1.722 5.987 2.823 2.0 8.7 3.163 4.78 94.57 132.07
P. occipitalis 4.536 7.510 2.980 6.0 22.0 4.530 3.91 74.30 127.70
P. tenellipes 1.727 6.755 5.400 1.0 24.5 1.350 5.70 12.50 44.65
P. borealoides 2.584 7.887 4.290 1.3 9.3 3.590 4.67 47.87 256.27
P. b. borealis 2.316 6.360 2.040 2.0 18.0 4.320 4.61 41.00 32.80
P. b. xanthodryas 2.025 5.750 2.470 3.0 20.0 3.290 4.43 39.30 54.70
P. trochiloides nitidus 3.091 7.600 2.980 4.0 13.0 4.620 4.96 72.30 112.30
P. t. viridanus 1.507 6.570 2.880 3.0 14.0 3.690 3.82 112.50 83.00
P. t. plumbeitarsus 1.846 8.050 2.500 4.0 15.0 5.550 5.30 47.30 110.50
P. schwarzi 1.459 6.440 1.850 3.0 24.5 4.585 3.48 17.00 50.80
Hippolais icterina 5.951 7.520 1.540 5.0 23.0 5.990 3.48 63.50 145.00
H. pallida 2.881 7.280 1.700 6.0 18.0 5.580 -- -- --
H. languida 6.132 4.640 2.190 2.0 39.0 2.450 3.82 63.50 61.00
H. c. caligata 2.765 4.840 1.780 1.5 19.0 3.060 3.04 80.00 102.30
H. rama 5.816 6.520 1.670 15.0 47.0 4.850 3.82 54.50 109.00