226 
THE WILSON JOURNAL OF ORNITHOLOGY • Vol. 123. No. 2. June 2011 
flapping. The males spread the tail four times in 
two videos, and two times in the other two videos. 
The videos looked similar to those shown for the 
Volcano Hummingbird in figures 3B-C. except 
the wings were Happed at the nadir of the dive. 
The tail was spread for an average of 37 ± 2 msec 
(n = 1 spreads), and the intervals between spreads 
were 31 ± 3 msec (n = 6), resulting in a tail 
spread cycle rate of 14.7 Hz—corresponding to 
that at which the pulses of sound were produced. 
The gliding phase (stage 2) apparently explains 
the gap in production of the trill during the dive in 
some of the dive recordings. The dive wing beat 
frequency was 93.7 ± 0.66 Hz, matching the wing 
trill rate during the dive. Two male S. scintilla had 
a mean wing-beat frequency of 68.9 ± 7.6 Hz 
while hovering in a cage. 
The R.2 of both the .Scintillant and the Volcano 
hummingbirds, when placed in a wind tunnel at 
speeds of 10-20 m/sec, produced sounds at the 
same frequencies as the pulsed sounds that the 
birds make while diving (Fig. 2D). The feathers 
generate tones with a fundamental frequency 
ranging from 0.3 to 0.5 kHz (depending on air 
speed), but in which the second or third harmonic 
is dominant. The feathers in the wind tunnel 
generated a stack of 20 or more harmonics, 
similar to the broad swath of sound present in the 
sound pulses (Figs. 2, 4). In particular, the 
emargmated tip of the feather flutters to generate 
the sound. 
DISCUSSION 
Our observations provide new informatio 
about the natural history of Scintillant an. 
Volcano hummingbirds. The shuttle display o 
the male Scintillant, and dive displays produce* 
by both species appear to be courtship displays 
We saw males of both species naturally display t« 
females, when a female was on a male’s territory 
Males rarely displayed to other males, tending U 
chase or ignore them instead. Our experiment 
use of homospecific females was sufficient u 
immediately elicit displays from males, whcrea- 
6. flammula females released on a 5. scintilla-. 
territory were ignored. 
Our goal in conducting this research was tc 
investigate whether Selasphoms species product 
ton n | dS ravT th llK "' Win8S Und lail leathers The 
tonal FMport.on of the Volcano Hummingbird’s 
ive sound , s „o| produced by the tail (Fig. 2E) 
nroduo , e y V0Ca l; conlrasl ' *e pulses of sound 
produced during the dive of both species are timed 
to the rapid tail-spreads in both species. We found 
similar one-to-one correspondences between ki¬ 
nematics and tail-generated sounds in Annas 
(Calypte anna), Black-chinned (Archilocus alex 
andri), and Calliope hummingbirds (Clark and 
Feo 2008, Clark 200*). Feo and Clark 2010) 
Moreover, the emarginated R2 (Fig. 1) of each 
species can generate sounds matching the dive 
sounds when placed in a wind tunnel (Fig. 2D). 
Rufous and Broad-tailed hummingbirds produce 
similar sounds during their dives, and both of 
these species also have emarginated inner rectri- 
ces (Stiles 1972, 1983), similar to Volcano and 
Scintillant hummingbirds (Fig. I). We hypothe¬ 
size that all of these species produce similar 
sounds during their dives via fluttering of the 
emarginated tip of R2. 
There is also a one-to-one match between video 
and sound recordings of the Scintillant Humming¬ 
bird’s shuttle display. Sound elements w and s 
were produced cumulatively at a rate of 93.8 - 
5.1 Hz, closely similar to the measured wing-beat 
frequency of 98.1 ± 2.64 Hz. We conclude that w 
and s arc produced by the wings, but how this is 
done is unclear for two reasons. First, although the 
individual shuttle motions of the shuttle display 
were periodic, they were produced at a frequency 
of 7.8 Hz. This is nearly four times slower than 
the rate ol 27.8 Hz (the rate at which a cycle of a 
w duplet and .v duplet were produced); thus, four 
cycles of such sounds were produced during each 
shuttle motion. We did not discern specific wing 
motions associated with the s elements. Second, 
the s elements are so short in duration that the 
sound spectrogram is intrinsically of limited use 
in detecting if they are truly atonal (broadband), 
or whether they represent a stack of closely- 
spaced harmonics, like the sound pulses of the 
dives. If they are harmonic slacks, perhaps they 
are produced via some form of resonant flutter 
(Clark and Feo 2008). If they are truly broadband, 
this would suggest they are produced by another 
mechanism such as percussion (Bostwick and 
Prum 2003) or rubbing (Bostwick 2006). 
The function of male Volcano Hummingbird 
territories is not entirely clear. Wolf and col¬ 
leagues (Wolf 1976. Wolf et al. 1976) called them 
feeding territories, implying that males were 
guarding a space specifically according to the 
value of the food resources it contained. But, food 
resources seemed nearly ubiquitous at our study 
sites, such that it would be difficult to find an 
open space that did not have some food present. 
