Reproductive Behavior and Communication in Blue-Throated Hummingbirds

13 Pages·2002·9.2 MB·English

by  Millicent Sigler Ficken

Wilson Bull., 1 14(2), 2002, pp. 197-209
REPRODUCTIVE BEHAVIOR AND COMMUNICATION
IN
BLUE-THROATED HUMMINGBIRDS
MILLICENT SIGLER FICKEN,' ^ KATHRYN M. RUSCH,24 SANDRA J. TAYLOR,^
AND DONALD POWERS^
R.
—
ABSTRACT. Previous studies of Blue-throated Hummingbirds {Lampornis clemenciae) showed that males
have elaborate songs and females also sing. Here we report complex territorial systems, sexual behavior, and
communicatory repertoires of both sexes. Males typically defended territories along streams throughout the
reproductive season. Time budgets revealed that males allocated more time to vocalizing in territorial advertise-
ment than in more energetically costly activities such as chases. Both observations and playback experiments
indicated that two vocalizations have different roles in territorial advertisement. Strings ofchips served in long
distance territorial advertisement, while the lower amplitude song, more prevalent early in the season, attracted
females and also repelled males from the immediate area. Agonistic interactions included long chases, some
with vocalizations. During close range encounters an array ofpostures and sounds occurred. Females defended
territories against other females during a briefperiod preceding egg laying. Sexual interactions were unusually
intricate, consisting ofa series ofactivities in which the female played an active role. She delivered loud calls
advertising sexual readiness and performed aerial displays. Males competed with other males for access to
females. We observed four presumed copulations, all preceded by a song fragment from the male. The vocal
repertoire included at least 16 different sounds, all with different functions, an unusually large number for a
nonoscine. Male songs were the most complex sounds. Behaviors that differed from many otherhummingbirds
that have been studied included strong reliance on vocal signals, lack of male aerial displays, and female
advertisement of sexual readiness. These attributes may be linked to reduced sexual dichromatism and related
to the species’ preference for wooded riparian habitats. Received 15 May 2001, accepted2 April 2002.
Hummingbirds provide a fascinating and species, but are absent in others (Wagner
difficult challenge for behavioral ecologists. 1954, Johnsgard 1997). Males ofsome species
The family Trochilidae includes >300 species congregate in leks during the breeding season
occupying a variety of habitats (Schuchmann (Johnsgard 1997).
1999). Although subjects ofnumerous ecolog- Detailed studies of visual and vocal signal-
ical and physiological studies, the reproduc- ing in reproductive interactions are lacking for
tive behavior of relatively few species has most hummingbird species with the exception
been assessed in detail, which is not surprising ofcomprehensive studies ofcourtship and ag-
considering their general tropical distribution gression in Streamertails, Trochilus polytmus
and the difficulty of studying such rapidly (Schuchmann 1980) and Anna’s Humming-
moving small birds. The general picture of bird, Calypte anna (Stiles 1982). The mono-
hummingbirds that has emerged is the con- graph on ecological and evolutionary aspects
servative nature of some behavior patterns of the lek mating system of Long-tailed Her-
and the diversity of others (Johnsgard 1997, mits (Phaethornis siiperciliosiis) includes a
Schuchmann 1999). Mating systems are po- description of the communicatory repertoire
lygynous, and males have no role in parental involved in reproduction (Stiles and Wolf
care, yet other behaviors vary according to the 1979). Our present study is an initial contri-
species. For example, elaborate aerial court- bution in assessing the nature and variety of
ship displays are performed by males ofmany hummingbird behavior.
The main focus of this study is visual and
' Dept, of Biological Sciences, Univ. of Wisconsin- vocal signaling systems and their role in me-
Milwaukee, Milwaukee, WI 53201, USA. diating territorial interactions and sexual be-
^Univ. ofWisconsin-Milwaukee Field Station, 3095 havior in Blue-throated Hummingbirds (Lam-
Blue Goose Rd., Saukville, WI 53080, USA. pornis clemenciae). We document and discuss
^ Biology Dept., George Fox Univ., Newberg, OR some previously undescribed behavior of this
97132, USA.
Current address: 1510 W. El Rancho Dr., Mequon, species that contributes new perspectives for
WI 53092, USA. integrating the role of intersexual communi-
^Corresponding author; E-mail: [email protected] cation in reproduction. Most studies of hum-
197
198 THE WILSON BULLETIN • Vol. 114, No. 2, June 2002
mingbird reproductive behavior have focused and chips utilizing playback experiments. Us-
almost exclusively on males, as they are much ing a time budget analysis, we documented
easier to observe because ofbrighter plumage the number of these two types of male vocal-
and more conspicuous behavior. The marked izations as compared with chases and other
sexual dichromatism of many species indi- aerial territorial behavior. We also tested the
cates that female choice may select for plum- prediction that longer songs should predomi-
age and behavioral attributes. Vocal signals nate earlier in the season when sexual activity
have received less attention. Here we suggest is highest, if a principal function is the attrac-
females have an active role in initiating sexual tion of females. Females also utter eomplex
behavior, as may be the case in other hum- muted songs, typically when close to a male
mingbird species as well. (Ficken et al. 2000), and we played back these
The Blue-throated Hummingbird has an ex- songs to males. We describe usage ofa female
tensive range in Mexico but breeds in only a call (designated Tia following its description
few localities of the southwestern United by Wagner 1952) and supplement observa-
States where males typically defend territories tions with playbacks. In addition, we describe
along streams in montane regions from their a variety ofvisual signals by both sexes, some
arrival following migration (typically April) with complex components that are important
until departure in late summer (Williamson constituents of reproductive behavior.
2000). Eight male territories were spaced at In addition to studies of general reproduc-
uneven distances along 1,500 m of stream in tive behavior and some visual displays, we
Arizona (Williamson 2000). Territory size and provide the hrst sonagraphic analysis of the
structure are variable and related to the distri- acoustic repertoire of a hummingbird and
bution of nectar supplies (Lyon 1976; DRP compare the behavior of Blue-throated Hum-
unpubl. data), but since birds also are breeding mingbirds with some other avian species re-
during their time in the southwestern United garding acoustic repertoire categories and
States, the functional importance of male ter- characteristics.
ritories probably extends beyond simple en-
METHODS
ergy acquisition to include a staging ground
for displays and vocalizations (Powers 1987). We conducted the study from 1995-2001, with the
The sexes are spatially segregated during most intensive data collection during May, and some
most of the reproductive period, with only a additional observations during late April and August
at two sites in southeastern Arizona (Cochise Co.):
b1r9i5e2f;oMveSrFlaupnpduubrli.ngdatsae)x.uaBlluaec-ttihvriotyate(dWaHgunme-r NR,am1s0e9y° 1C0a'nyWo),n ainndthteheHuSaocuhtuhwceastMeorunntRaeisnesar(c3h1°S5t0a'-
mingbirds are unusual in having a more pro- tion in the Chiricahua Mountains (31°26' N, 110° 19'
tracted period of association of the sexes than W). Both sites had numerous feeders visited by a va-
other hummingbirds studied in Mexico by riety of hummingbird species. Although many obser-
Wagner (1954). Females defend two kinds of vations were made <10 m from feeders, others were
made at a greater distance. Some behavior was much
territories, with no temporal overlap: the im-
easier to observe near feeders, because the interaction
mediate nesting area for a few days before egg rate was increased by competition for access to feed-
laying (Wagner 1952), and temporary feeding ers. Eeeders undoubtedly affected many aspects ofre-
territories later in the season (Lyon 1976). productive behavior, such as nest site selection and
Our previous studies described the structure population density, but we were unable to assess all
and usage of complex male songs (Ficken et these effects. We described different typesofagonistic
interactions, but because of the rapidity of certain
al. 2000). These songs comprised clusters of types of encounters we were not always able to de-
notes (termed “units") that we designated A, scribe all components for both participants.
B, C, D, and E. Our initial observations in- We determined time budgets for nine males at the
dicated they may serve as short range teni- Southwestern Research Station and two males at Ram-
torial advertisement as well as female attrac- sey Canyon that had established territories around
tion. Other vocalizations, particularly strings feeders (June to July 1995). We quantified two 30-min
periods for each individual, one during the morning
of chips (the loud "seeps" of Marshall 1957), (09:00-12:00 MST) and one during the afternoon (12:
also are associated with reproductive behavior 00-17:00). The observer was positioned about 20 m
(Williamson 2000). Here we address questions from the feeder being defended by the focal male. We
concerning the different functions of songs recorded behaviors related to territorial defense. We
Ficken et al. • BLUE-THROATED HUMMINGBIRD BEHAVIOR 199
edgiovriideesdt(chheasseeasg,oncirsotuiccheesn,coaunndtehrosvienrtso)tfhorreeamnaaljyosriscaot-f TABLE 1. Time budgets indicate that male Blue-
throated Hummingbirds (Lcimpornisclemenciae) spent
tshoeuinrdasssoofctieantiaocncowimtphanvioecadlitzhaetsieonbse.haTaviilorsppraettaedrinnsg.aWned more time vocalizing in territorial advertisement and
defense than in active aggression (chases and other
analyzed a sequence of activities of one encounter in
nonvocal behaviors). Values are mean number of acts
slow motion from a published videotape (Nature Sci-
per 30-min observation period for individually identi-
ence Network, Inc. 1988).
We recorded vocalizations and notes using Sony fiable males. Dataare from the SouthwesternResearch
Professional Walkman WM-D6C cassette tape record- Station (birds 1-9) and RamseyCanyon (birds 10-1 1),
southeastern Arizona, June and July 1995.
ers and either a Nakamichi directional microphone or
an Audiotechnica AT 877 microphone. Some natural Aggressive
behavior and many playbacks were videotaped using Bird acts Chipping Singing
a Sony camcorder. We made sonograms with a Kay 1 0.36 18.03 0.93
7800 Sona-graph at 8 kHz with a 150-Hz bandwidth 2 0.00 0.53 2.04
filter. For measurements of chips and Tias, we used 3 0.23 7.92 0.87
Avisoft SONAGRAPH Pro (ver. 2.5). 4 0.65 0.28 0.45
We conducted playback experiments with the Sony 5 0.28 17.83 1.78
Professional Walkman and a Sony SRS-27 speakeraf- 6 0.58 0.57 1.04
fixed about 2 m high to abranch in areas nearfeeders. 7 0.05 9.28 0.23
The feeders were >50 m apart, increasing the proba- 8 0.94 13.12 0.39
bility that different individuals were tested. 9 0.11 22.13 0.13
To test whether responses differed for certain vo- 10 0.17 15.50 1.07
calizations associated with reproductive behavior, we 11 0.43 0.40 0.07
conducted playbacks during May 1998 and May and Grand mean 0.35 9.60 0.82
early June 1999. We completed 14 experiments at two (SD) (0.29) (8.25) (0.65)
sites (five males at Ramsey Canyon and nine males at
Southwestern Research Station). Eachpresumablywas
a different male, based on distance between playback
sites. Each experiment consisted of three randomized whether both males and females responded by ap-
series of vocalizations, each lasting 85 s with a 90 s proaching the speaker.
interval between playbacks. Playback durations were We measured sonagrams for comparisons of calls.
similar to normal singing in bouts. In natural singing, Because of the extremely varied usage, we analyzed
a bout began with the first song and terminated when chips separately according to sex and context. Values
the bird ceased singing or flew. The amplitude for all are presented as means ± SD. Unless indicated oth-
playback sounds was adjusted to 70-75 dB (measured erwise, r-tests are two-tailed. For comparisons of re-
with a Yu Fung YF-20 sound level meter). For male sponses to sound playback we used the sign test, com-
song (Type 1 of Ficken et al. 2000), we used a com- paring approaches during playbackto the nonplayback
mon variant consisting of a string of six units situation. P values <0.05 were considered significant.
(ABCDEB) repeated 11 times. Playback of female The data in Tables 2 and 3 are based on >8 males and
song (Type 2 of Ficken et al. 2000) lasted 1.9 s and >3 females. We compared the incidence of male and
was presented nine times, and chips were acontinuous female agonistic behavior patterns, but because the
series (Serial Chips) using natural intervals. All males data were not obtained from known individuals, they
in the population used the same song variants, and only suggest trends and statistical tests were not per-
chips were very stereotyped. Response measures for formed.
hummingbirdswerenecessarilydifferentthanthosefor
RESULTS
oscines because ofthe behavioral differences between
the two groups. Hummingbirds rarely perched near a Territorial behavior and time budgets re-
I'jII sqia(m2upnn)evedonoaftolk(lbvet3yvihr).inegoaWdsupuprespreaairpmksnoietgodaarcvr.ahtepgeRlmidetateosyapntpblte<oaadcn1yaksb0fwealisacagmkynhistdnsowcwflhavhuoneetdcdnhenaedlhpiao(slzv1paem)eeydarabnlikaoeenocrgnak,lwpynapsuesrtsaoiaurnra<aftclt1rehlhed5;ye-; eotlsvradpeetleernyndtdtwwtameogaorgsarregeecgslasretfieafiosremnsce.uitnoeTvnddowe.c—robaylTtimhizaefmilesneeeegdsecbrotussnhdp,dageinnetbttiuesotannlgsuma.naodgsMmoetaaudljbtetohis-ren
speaker. We noted the actions ofthe birds before play- same amount of time chipping (Serial Chips)
back,duringpausesbetweenplaybacks, andafterplay- as engaged in aggressive acts, but all the oth-
back. For the analysis we used only behavior during ers spent much more time chipping (Table 1).
playback.
‘ Similarly, singing occupied more time for 8
j SouItnhawessetpearrnatReesseetarocfhexSptaetriiomnenwtes p(lJauyneed1T9i9a9)caaltlsthien of the 1 1 males than overt aggression. Most
I areas where females had been observed producing birds (8 of 11) spent more time chipping than
i these calls. These experiments were designed to test singing.
;
200 THE WILSON BULLETIN • Vol. 114, No. 2, June 2002
There may be a relationship between terri- = 28, P = 0.36) or the mean high frequency
tory size (and possibly quality) and degree of (t = 0.99, df = 28, P = 0.33), but we detected
aggression exhibited by the territory holder. tonal differences between the sexes. Male
The most dominant male in our study (bird 1 chips sounded harsher, perhaps due to more
Table 1) defended a large territory that ap- high frequency bands than those of females,
peared to be about 10,000 m^ in size. When with the inclusion of a second band from
bird 1 was removed, his territory was broken 10.9-12.2 kHz (present in all 16 samples),
into several smaller pieces defended by males and sometimes a third even higher band 15.9-
exhibiting less dominant behavior. Another 18 kHz (present in 6 of 16 samples). We ob-
male (bird 2) who defended a territory adja- served less banding in females, with a second
cent to a dry creek encountered substantially band of 11.6-13.4 kHz (present in 7 of 14
fewer intruders and exhibited no aggressive samples) and no third band (n = 14). The
behavior duri—ng observations (Table 1). most striking difference, however, was that fe-
Male song. Song length was related to the male Single Chips were about half the dura-
stage ofthe breeding cycle, with shorter songs tion of male Sin—gle Chips (Table 2).
more prevalent later in the season (one-tailed Serial Chips. Males delivered long strings
t = 1.8, df = 286, P = 0.037). During May ofchips, particularly in the early morning and
songs had a mean of 4.99 ± 3.46 units {n = late evening, and less frequently at othertimes
242 songs), and during July a mean of 4.09 of day. Calling was most prevalent early in
± 2.75 (/? = 46). Although some long songs the breeding season. A single male called for
persisted throughout the summer, the short 20 min on one occasion, and many bouts were
song ofjust two units (AB) was the most com- as long as 10 min. The mean rate was 7.21 ±
mon song variant, and as the season pro- 1.37 (/? = 23), and the chips were delivered
gressed the proportion of AB songs increased at uniform intervals. Single Chips and Serial
(18% of all songs early in the season, 33% Chips did not differ significantly in frequency
later). {t = 1.96, df = 32, P = 0.058), but individual
Although single songs were frequent, songs chips in a series were shorter than male Single
also were performed in bouts. Pauses between Chips {t = 6.2, df = 30, P < 0.0001).
bouts always were much longer than intervals Males typically perched near the creek on
between songs within a bout. The longest bout high perches. Serial Chipping often was ac-
was 18 songs, but most bouts were composed companied by head movements (Wagner
of only a few songs. Early in the breeding 1954; MSF pers. obs.) through an arc ofabout
season bouts were somewhat longer (mean = 150°. While chipping, the male occasionally
4.3 ± 3.8 songs, n = 59) than later in the altered the pitch of the call (Willard 1911), as
season (mean = 2.8 ± 1.4 songs, n 17; one- indicated in Fig. 2. To our ears, the pitch shift
tailed t = 1.6, df—= 1A, P = 0.045). and the changing amplitude during head
Single Chips. Males often were silent movements produced a “stereo” effect. In
when perched on their territories, unless an- fact, we believed two birds were calling when
other male Blue-throated Hummingbird or we first noted this behavior.
Magnificent Hummingbird {Eugenes fulgens) Female Serial Chips were rare, and all five
approached. The other hummingbird usually cases occurred when a female was in the last
left the vicinity following one or two chips by stages of nest construction. During this time
the resident Blue-throated Hummingbird. females engaged in territorial interactions with
These Single Chips also preceded flight, or neighboring females. Female Serial Chips
were uttered as the bird approached a feeder. were associated with a head tossing move-
Females usually were silent both in the vicin- ment, but unlike the males, we detected no
ity of feeders and when perched away from pitch shifts in females. In females, individual
feeders (<1% ofchips we recorded were from chips in a series were longer than Single
females). Chips (r = 3.15, df = 30, P — 0.004), in con-
Single Chips were very short, high pitched trast to males.
notes (Fig. la). No significant differences oc- Playback experiments of Serial Chips and
—
curred between the sexes with respect to the male andfemale song. Males responded sig-
mean low frequency ofthe chips (r = 0.99, df nificantly with approaches to male song (sign
Ficken et al. • BLUE-THROATED HUMMINGBIRD BEHAVIOR 201
4-
3-
2-
1 1 1 1 1
0.5 1.0 1.5 2.0 2.5
Time
(s)
FIG. I. Sonograms of Blue-throated Hummingbird (Lampornis clemenciae) vocalizations recorded at the
Southwestern Research Station, southeastern Arizona, by M. S. Ficken, AprilthroughAugust, 1995-2001. Black
circles on horizontal axis separate different kinds of calls. Sonograms were prepared using a Kay 7800 Sona-
Graph with 0-8 kHz settings and a 150-Hz bandwidth. In a few cases, some elements were higher than 8 kHz.
(a) Single Chip ofa male, (b) Staccato Trill, (c) Bill Crack, (d)Zeet. (e) ShortCall, (f) LongCall (mostcommon
form), (g) Complex Long Call, (h) Tia. (i) Mobbing Chips (more than one individual is calling).
202 THE WILSON BULLETIN • Vol. 114, No. 2, June 2002
TABLE 2. Acoustic characteristics of Blue-throated Hummingbird {Lampornis clemenciae) chips illustrate
the variation associated with sex, age, and context. Values are means ± SD {n). Data are from the Southwestern
Research Station, southeastern Arizona, May through June, 1995-1999.
Lowest Highest
Duration(s) frequency(kHz) frequency^(kHz)
Single chips (female) 0.11 ± 0.05 (14) 5.64 ± 0.35 (14) 6.79 ± 0.33 (14)
Single chips (male) 0.21 ± 0.05 (16) 5.50 ± 0.50 (16) 6.90 ± 0.28 (16)
Mobbing chips 0.08 ± 0.02 (23) 5.68 ± 0.21 (9) 7.63 ± 0.80 (9)
Fledgling chips 0.11 ± 0.02 (10) 7.86 ±0.15 (9) 9.39 ± 0.17 (9)
Series chips (female)^ 0.16 ± 0.04 (18) 5.14 ± 0.17 (11) 6.35 ± 0.14 (11)
Series chips (male) 0.13 ± 0.02 (18)*^ 5.31-6.20 (18)^ 6.44-7.10 (18)‘i
Energydistributioniscenteredinthisband.
^Meanofmeansfortwofemalechipseries.
Meanofmeansforfivemalechipseries.
Meanofrangesforfivemalechipseries,depictingthepitchshiftforthesevocalizations.
test, L — I, n — \4, P = 0.002), but only two were elevated and vibrated, only eight times,
approaches by males occurred during play- but as it occurred very rapidly, many probably
back of female song (L = 2, n = 14, P = were not detected. Each case occurred when
0.012). Responses to playbacks of Serial a resident was approached by another male, or
Chips were ambiguous, with two males leav- another male flew directly over the perched
ing during playback, five males approaching, resident. Usually the resident’s tail was
and six males not responding at all (L = 5, P spread, and sometimes the bird emitted a Sin-
= 0.58). No females responded to playbacks. gle Chip or Zeet. The opponent then left. Tail
Behavior during agonistic encounters be- spreading occurred alone or in combination
—
tween males. Hummingbirds exhibited an ar- with many other acts and exposed the prom-
ray ofpostures and sounds associated with ag- inent white tail spots. In some cases the tail
onistic interactions. Some activities occurred also vibrated rapidly.
in conjunction with others, e.g., sounds ut- Chases varied greatly in duration; some
tered when the bird was crouched with a were very long, apparently covering more
spread tail. Such composite displays were than the length of one territory. They often
quite common. were initiated by one bird flying directly to-
We observed wing waving, in which wings ward another, but as both birds disappeared
10-
^
'• ^ K
I-
T"
I I I I
0 2 4 6 8 10
Time
(s)
FIG. 2. Sonogram ofmale Blue-throated Hummingbird (Lampornis clemenciae) Serial Chips showing pitch
shifts. Recorded by S. J. Taylor at Southwestern Research Station, southeastern Arizona, June, 1995. Prepared
using Canary 1.2.
Ficken et al. • BLUE-THROATED HUMMINGBIRD BEHAVIOR 203
from view quickly, we could not distinguish TABLE 3. Male and female Blue-throated Hum-
between shorter supplantings and longer chas- mingbirds {Lampornis clemenciae) differed in the
es. Long chases involving three (and more numberofintrasexual agonistic actsandassociatedvo-
rarely four birds) were frequent, particularly calizations. Data are from the Southwestern Research
early in the breeding season, but were impos- Station, southeastern Arizona, April through June,
sible to follow in detail. One bird chased an- 1995-2000.
other, with the third bird joining later. Males Agonisticactivity
sometimes supplanted females, but long inter- Chase Crouch Hover
sexual chases were uncommon except in a
Male- Female- Male- Female- Male- Female-
sexual context. Male Blue-throatedHumming- Calltype male female male female male female
birds, the dominant species, sometimes chased Long call 13 0 0 0 0 0
male Magnificent Hummingbirds, but heter- Short call 4 0 9 0 3 0
ospecific encounters were rarer and briefer Zeet 6 0 4 0 4 1
than conspecific ones. Encounters with Mag- Bill crack 6 0 0 0 10 1
nificent Hummingbird males were confined to No calls 8 9 18 5 5 6
Total 37 9 31 5 22 8
chases and did not include hovers or crouches
described below.
Another type of interaction involved
crouching in which a perched bird (resident) as I hovered over it. I left and R remained.
was approached by another (intruder) that Hanging upside down was relatively rare in
hovered within a few centimeters of the resi- our sample of crouches (19.4%, n = 31), and
dent. The outcome was variable. A chase we never observed one with a bill-to-bill com-
sometimes ensued immediately with the resi- ponent as in this video.
dent pursuing the intruder. Alternatively, the Hovering was another type ofconfrontation
perched resident crouched with tail spread. (Table 3) in which both birds flew up face to
The position of the hovering bird sometimes face only a few centimeters apart with widely
changed from a frontal approach to hovering spread tails, sometimes flying several meters
directly over the resident, and in rare cases the into the air while maintaining the same ori-
intruder even landed on the back of the entation. Hovers sometimes terminated in
crouched bird. An even more unusual posture chases.
—
occurred in which the crouched bird fell over Agonistic sounds. Various sounds were as-
backwards and hung upside down while the sociated with agonistic interactions. One was
opponent continued hovering over it (see a trill (unit B), a common component ofmale
drawings in Rising 1965). The bird that songs (Ficken et al. 2000), and the only song
crouched always won (« = 31 encounters); the unit delivered without any accompanying
bird that approached the resident left or was units. The B unit occurred in several different
chased, in contrast to Rising’s (1965) conclu- contexts. The loud Staccato Trills (Fig. lb) oc-
sion that the crouched bird was the subordi- curred during some agonistic encounters. Soft
nate. Trills of much lower amplitude with shorter
A published videotape of male Blue-throat- internote intervals than the staccato ones
ed Hummingbirds (Nature Science Network, sometimes occurred during close range ago-
Inc. 1988) allowed slow motion analysis. Ini- nistic interactions between males, and all ap-
tially the resident (R) was perched, the intrud- parent copulations were preceded by these
er (I) landed next to R, and then I flew and whispered sounds. We excluded these trills
hovered over R. R shifted position and hung from the quantitative analysis of sounds dur-
upside down with his tail spread and bill ing encounters (Table 3) because the soft ones
pointed upward at the hovering I, who had its were inaudible unless we were within a few
bill pointed at R (at one point their bills al- meters of the bird.
most touched). I changed position, and the Bill Cracks were nonvocal sounds (Fig. Ic)
two birds briefly were tail-to-tail (R was still produced during hovers and chases (Table 3).
upside down). I called and moved, still hov- In the few cases where we had good views of
ering a few centimeters above R. R righted the combatants, the sounds were produced as
itself on the perch but maintained tail spread the birds hovered and struck bills at about 2/
204 THE WILSON BULLETIN • Vol. 114, No. 2, June 2002
3 the length ofthe bill. The sounds were loud, observed lining a nest adjacent to an area
indicating a forceful impact. Bill Cracks oc- where another female was engaged in the
curred during chases when the bird being same activity. Both females were involved in
chased turned and confronted its pursuer. Bill many agonistic interactions with each other.
Cracks were not heard during crouches. Both returned to their original sites following
Four major types of vocalizations also oc- encounters. We never observed agonistic en-
curred only during agonistic interactions. The counters between incubating females or those
Zeet was a short, harsh, snarly call ofvariable with nestlings, despite about 20 h of obser-
duration, covering a wide frequency range vation.
(Fig. Id). Short Calls were a series of several The types of agonistic encounters between
(typically 3) transients covering a wide fre- females included the same general categories
quency range (Fig. le). Long Calls (Fig. If) as those of males (Table 3), but female en-
were more protracted strings of notes with counters were of lower intensity. During 22
more complex structure than Short Calls. We encounters, sounds occurred only twice. Chas-
recorded an unusually complex call only twice es usually were shorter as well as less frequent
(both times delivered by the same male) dur- than those of males.
—
ing our extensive recording (Fig. Ig). Its rarity Male-female interactions. During a period
in our sample may be due to its occurrence in of about two days, the male often remained
an unusual situation when two males hovered near the female and in the area of the nest.
close to each other during an agonistic en- We observed behavior that seemed sexual, but
counter near a sexually active female. Many because most birds were unmarked and the
of the notes in these calls are transients cov- stage of the nesting cycle for many females
ering a wide frequency range. The Complex was unknown, we were able only to piece to-
Long Call was more complex than other Blue- gether a number of separate observations to
throated Hummingbird vocalizations except produce a general pattern. Females sometimes
for song, as there was a mixture ofnote types, approached singing males and produced a
and some notes covered a narrower frequency very distinctive but muted song that often
range than is typical forother agonistic sounds overlapped the male’s song (Ficken et al.
such as Short Calls and the typical Long Calls. 2000). In all cases where visibility was good,
Some sounds were associated with certain the female left after <10 s. On one occasion
types of interactions more than others (Table when there were no prior interactions, a male
3). Long Calls were performed only during hovered over a crouched female. She uttered
chases, while Short Calls and Zeets occurred a Zeet and he left, but he returned about 5 min
during all three types of encounters (hovers, later to the same place and delivered long
chases, and crouches). Zeets also were given songs. The female then disappeared from
when another individual flew near the caller view.
—
but no obvious interaction occurred. They Male-male interactions nearfemales. The
may inhibit approaches, in contrast to the oth- period of late nest building through nest com-
er sounds that occurred once an interaction pletion coincided with sexual behavior and
was already underway. Crouches often were was marked by especially intense agonistic
silent, while hovers, and particularly chases, encounters between males, often involving
often had accompanying sounds. long chases. This was the only time we ob-
—
Female territorial behavior. We noted ag- served physical contact between males. On
gression between neighboring females at the three different occasions males tumbled to the
time that sexual activity was occurring (late ground, hitting with a strong impact (“thuds”
nest building and perhaps egg laying). This were heard) with one grasping the other. Fe-
was the time that females uttered Serial Chips males always were nearby.
near the nests {n = 3 females). While feeders Female sexual invitation behavior.—An un-
may have modified spatial relationships, most usual aerial display that we named “River
interactions were not near feeders. Two fe- Dance” was performed only by females
males were observed for several hours each known to be sexually active, just following
on six days, and their territories did not over- nest completion and during the period of Tia
lap. One female had fledglings and later was calling (see below). Just prior to dancing, a
Ficken et al. • BLUE-THROATED HUMMINGBIRD BEHAVIOR 205
female would be perched very low near a seemed to call spontaneously, but at other
creek or would fly to the site ofthe dance. All times they called in response to a male flying
dances took place over a creek, a dry creek overhead. Some calls from perches elicited no
bed, or more rarely (<10% of cases) <10 m apparent responses from other birds. On other
from a creek. Dances did not take place near occasions the female called in flight as she
nests. Dances may have originated from ritu- approached and then engaged in a chase with
alized insect-catching movements, but no in- a male, although it was not clear which sex
sects were present as we inspected the sites was the ultimate pursuer. In one case, afemale
carefully immediately after dances. No dances uttered a very muted Tia as the male hovered
were observed during a study session in which over her, but no further interactions took
no females were at the right stage ofthe nest- place. In two instances encounters with Tias
ing cycle. ended in copulations.
River Dance was a series of short flights, We also conducted seven playbacks ofTias.
usually only a few cm high and 1-3 m in di- Responses occurred at all sites where we had
ameter, often centered over the creek. The fe- observed females calling that day. In all cases
male did not vocalize during the dance, and the females approached, and in three cases she
she was quiet before starting a display. Each uttered Tias, chipping only once. In four of
short flight was followed by a hover (typically the seven cases a male also approached during
<2 s long) and a change in position. During playback. One time the male chased the fe-
some but not all hovers, females exposed their male, and another time two males approached
white outer tail feathers, sometimes while and chases ensued. In two other cases, chases
flicking the tail back and forth. Dances incor- occurred but the sexes of the participants
porated all directions (up, down, left, right, could not be de—termined.
forward, backward, and diagonals) in 3-di- Copulations. We observed three probable
mensional space. Mean duration (timed from copulations, and E. Sandlin (pers. comm.)
video recordings) was 12.4 s ± 6.0 (n = 19), provided data on another. All happened during
with about 10 movements per dance. In some late May or early June at the Southwestern
cases we observed males within a few meters Research Station. The events occurred very
of the dancer, but in many cases we were not rapidly and cloacal contact was not observed.
certain if males were present, as they tended All were on the ground in dense vegetation at
to perch low in very dense vegetation and low light levels during early morning or even-
chipped only occasionally. We observed no ing. In three cases a chase preceded copula-
interactions immediately following the danc- tion, in two cases Tias occurred, and in all
es. four cases the male performed a soft B trill
During periods of sexual activity, females fragment.
—
uttered special calls (Tias) that were not heard Calls ofthe young. Nestlings were silent,
at other times. We measured the calls from a as was the female in approaching the nest,
pooled sample of females (>3 different birds, giving only a soft call when a human was near
8 sonograms); mean duration was 0.58 s ± the nest (Wagner 1952; MSF unpubl. data).
0.25, highest frequency was 7.89 kHz ± 0.76, However, fledglings called prior to a feeding
and lowest frequency was 5.75 ± 0.09. Unlike when the female was several meters away, and
Serial Chips, Tia calls consisted of a series of sometimes in her absence. Sometimes the fe-
rapid chip notes with the first note always lon- male also chipped before approaching the
ger and slightly higher pitched than the others fledglings for feedings (Wagner 1952; MSF
(Fig. Ih). unpubl. data). When the female was close, the
The calls were given either when the bird young quivered their wings. The chips uttered
was perched or when she flew. In contrast to by fledged young just before independence
many calls of this species, Tias were unusu- differed from the other chips in the species’
ally loud, and functioned as long distance sig- repertoire (Table 2) in being higher pitched,
nals, as individuals >50 m distant approached with the main band between 7.9-9.4 kHz and
in response to the calls. Because of observa- a second band around 20 kHz. —
tional difficulties, the stimuli eliciting the calls Calls in response to a predator. We ob-
were not always apparent. Sometimes females served only one case of a reaction to a pred-
206 THE WILSON BULLETIN • Vol. 114, No. 2, June 2002
ator, a Northern Pygmy-Owl (Glaucidiiim forbeing very aggressive. However, time bud-
m
gnoma), which was perched about 1 from gets showed that the use ofmore energetically
a feeder. Five Blue-throated Hummingbirds expensive behavior involving flights (short
(unknown sex) flew at the owl giving loud chases, hovering, and the long chases char-
chips. The owl then flew and was followed acteristics of this species) were rarer than the
closely by the chipping hummingbirds. Calls presumably lower cost vocalizations. The dis-
of different individuals overlapped, a pattern covery that vocalizations are an important
often associated with oscine mobbing (MSF component of hummingbird territory mainte-
unpubl. data). Because ofthese conditions, the nance is not new, having been suggested pre-
exact number of individuals calling was un- viously for the highly vocal Anna’s Hum-
known. Mobbing Chips (Table 2) were of mingbird (Stiles 1982, Powers 1987).
shorter duration than either male or female These hummingbirds were unusual in hav-
chips (t = 11.3, df = 31, P < 0.0001) and ing two different vocal systems involved in
were higher in frequency (t = 3.38, df = 23, territorial behavior. Oscine song typically
P = 0.030) compared to female Single Chips. serves dual roles in territorial advertisement
Mobbing Chips also differed from other types and mate attraction (Catchpole and Slater
ofchips in having a second band ranging from 1995). Blue-throated Hummingbirds have par-
7.8-9.3 kHz (/? = 10; Fig. li), instead of the titioned these functions in a different way,
usual harmonic band for Single Chips or Se- with song mediating close range interactions
rial Chips. between males and attracting females. Serial
Another type of call, described as a Chips, on the other hand, were involved only
“squeal,” is uttered by both sexes and young in long distance territorial advertisement.
in stressful situations, such as being handled Males advertised territorial boundaries
(Williamson 2000). through strings of loud Serial Chips often
from an exposed perch near a stream, whereas
DISCUSSION
songs were low amplitude and sometimes
—
Reproductive behavior. As hummingbirds were delivered when a male was hidden. Blue-
are polygynous with no pair bonds and no throated Hummingbirds preferred riparian
male parental care, females would be expected habitat, a noisy environment in which chips
to exert strong selection for male traits. In probably were effective for long distance
many hummingbirds extreme sexual dimor- communication. Head turning movements
phism, brilliant male colors, and often acro- during chipping broadcast the signal in many
batic aerial displays are associated with court- directions and increased its effectiveness.
ship (Wagner 1954, Johnsgard 1997, Schuch- Even when silent, they were alert and sur-
mann 1999). In Blue-throated Hummingbirds, veyed their environment, immediately pursu-
sexual dichromatism is minimal, the male be- ing conspecific interlopers. Chipping birds
ing distinguished from the female by a blue usually were not approached by conspecifics.
gorget, and in the dim light of its typical hab- Playback results also strengthened our in-
itat even that difference may not be readily terpretations that Serial Chips and song had
apparent. No aerial displays orexaggerated vi- different roles in territorial behavior, as close
sual signals by males are associated with approach to playback was associated only
courtship (Wagner 1954; MSF unpubl. data). with song. Serial Chips seemed to transmit the
Our observations of the reproductive be- message of territory occupancy, and may in-
havior of Blue-throated Hummingbirds re- hibit intrusions by other males. Song also in-
vealed some new aspects of hummingbird be- dicated ownership, but was not the primary
havior. Unusual features included female ter- means of excluding other males, serving in-
ritoriality near the nest, active and complex stead as a short distance threat (characterized
role of females in initiating sexual encounters by very low amplitude). Song often was as-
(including aerial displays), courtship occur- sociated with agonistic encounters, and other
ring in several stages, and the diversity and males sometimes approached a singer. Other
complexity of communicatory signals. indications for song playing a strong role in
Males of many hummingbird species are sexual behavior were that longer songs (hence
teiTitorial, and Blue-throated males are noted more complex ones), and longer singing bouts