MILLER_605_Ornis_Scand by panniuniu


									                19:                 1988
ORNISSCANDINAVICA 257-267. Copenhagen

                       of                            and
Breedingvocalizations Baird's SandpiperCalidis bairdii
relatedspecies, with remarkson phylogenyand adaptation
E. H. Miller, W. W. H. Gunn* and B. N. Veprintsev

                                       Miller,E. H., Gunn, W. W. H. and Veprintsev,B. N. 1988. Breedingvocalizations
                                       of Baird'sSandpiper                                             on
                                                                 bairdiiand relatedspecies, with remarks phylogeny
                                       and adaptation.OrnisScand. 19: 257-267.

                                       Vocalizationsof Baird'sSandpiperwere recordedin northeastern    Alaska, U.S.A.
                                       Malesbeat the wingscontinuously a loose deep flutterduringdisplayflights.They
                                       utterRhythmically RepeatedCalls(RRCs), SongandChatterin displayflights,anda
                                       Laughwhen mildlydisturbed.RRCs averaged690 ms long with intervalsbetween
                                       them of 280 ms. They have a "buzzy"quality because they comprise a series of
                                       rapidlyrepeatedpulses (pulse rate = 77 Hz). Songs are long, complex,multipartite
                                       utterances,and usuallyinclude Chatterat the end. Laughs are brief, simple trills
                                       consisting a singleelementtype repeatedseveraltimes. RRCs, ChatterandLaugh
                                                      homologouswith call types describedfor other calidridinespecies.
                                       are structurally
                                       Phylogeneticaffinities and signallingdistance account for many of the acoustic
                                       featuresof Calidridini.

                                        E. H. Miller,CollectionsProgram,RoyalBritishColumbia Museum,Victoria,British
                                        Columbia,CanadaV8V 1X4 and Biology Dept, Univ. of Victoria,P.O. Box 1700,
                                        Victoria,BritishColumbia,CanadaV8W2Y2. B. N. Veprintsev,  Institute Biolog-
                                        ical Physics, U.S.S.R. Academy of Sciences, Pushchino-on-Oka,
                                                                                                    Moscow Region
                                        142292, U.S.S.R.
                                        * Deceased.

                                                            Miller et al. 1987), and has revealed low intraspecific
                                                            variation for one type of sound in a representative spe-
Research on acoustic communication in songbirds has         cies (Miller 1986). Thus vocalizations of calidridine
contributed useful information to systematics, espe-        sandpipers are potentially valuable in systematic re-
cially below the species level (Payne 1986). Less com-      search, and provide a useful contrast to the more widely
parative information on vocalizations in other taxa is      studied and better-known songbirds.
available, probably because such species are harder to         The present study contributes new information on
record and because most do not possess a discrete class     vocalizations used by breeding calidridines, with special
of loud, complex vocalizations like "song" as songbirds     reference to the Baird's Sandpiper Calidris bairdii. Our
do. The 24 or 25 species of calidridine sandpipers          findings support the suggestion that vocal attributes of
(Gochfeld et al. 1984, Hayman et al. 1986) offer oppor-     calidridines are evolutionarily conservative, and suggest
tunities for comparative research because they are          that phylogenetic relationships and selective pressures
closely related, exhibit a range of mating and spacing      related to transmission distance can explain most fea-
systems, and share certain homologous displays (Pitelka     tures of calidridine nuptial displays (contra Loffredo
et al. 1974). Earlier work has identified several homolo-   and Borgia 1986).
gous classes of sounds in calidridines (references in          Acoustic communication in shorebirds has been re-

Received 2 July 1987
Revised 11 January1988
Accepted 3 February  1988

ORNIS SCANDINAVICA 19:4 (1988)                                                                                    257
                                                                                                   Fig. 1. Calidris bairdii,
                                                                                                   Rhythmically Repeated
                                                                                                   Calls (RRCs). A - Portion
                                                                          I ..I..,                 of long sequenceof RRCs
                                                                                                   (analyzingfilter bandwidth
                                                                                   *f_-        ~   (afb), 150 Hz). B -
c     1-                                                                                           PenultimateRRC before
                                                                                                   Song (afb, 300 Hz). C -
                                                                                                   Same RRC analyzedat half
                                                                                                   speed over 0-8000 Hz; three
v     2-       *        -                                                                          pulses that are analyzedin
EL    1,                                                                       -   litt i   ~      part G are marked(afb, 600
           ',:.:-   ?
                                                                                                   Hz). D - Beginningand end
                                                                                                   of same RRC analyzedon
                                                                                       ....-       logarithmic frequency scale
                                                                          1Ijykl^                  at quarterspeed over 0-8000
                                                                                                   Hz, with same pulses
                                                                                                   marked(afb, 180 Hz). E =
                                                                                                   Oscillogram same RRC,
                                                                                                   with same pulses marked.F
              E                                 Fmli           zzkki
                                                                                                   - Oscillogram of 17 terminal
                                                                                  AL? AApulses of same RRC, with
                                                                                          same pulses marked(signal
                            z         _
                                      _   w                                               digitizedat 20 kHz). G -
                                               ~G          f,
                                                                                          Power spectrumof three
                                                                              '-~-~_      markedpulses (signal
                                -'0   4   6.
                                                                                         digitizedat 20 kHz; FFT
                                                        I'       II  I
                                                                                -      t based on 1024points; see
                                                0       1        2     3
                                                              Frequency - kHz
                                                                                          Methods).Time markersin
                                                                                         ms, except part A.

viewed by Miller (1984). Additional references are sum-      itized at 20,000 Hz (it is desirableto digitizeat a rate at
marized in the Appendix to the present paper.                least 2.5 times a signal'smaximumfrequency).Spectra
                                                             discussed below were computed using no pre-emphasis,
                                                            a Hammingwindow, and low smoothing (average in-
                                                            corporating   four bins on either side of the currentbin).
                                                               We made additionalmeasurementson calls used in
We recorded C. bairdii, Pectoral SandpipersCalidris         aerialdisplaysby the White-rumped       SandpiperCalidris
melanotos,and Buff-breasted  SandpipersTryngites sub-       fuscicollis (see Fig. 8A): total call durationfrom begin-
ruficollisnear the mouth of the CanningRiver and on         ning of pulsed section to beginning of isolated pulse
BarterIsland,in northeastern Alaska, in June 1979.We        (a-d); durationof pulsed plus terminal section (a-c);
used a Nagra IS recorderwith a SennheiserMKH816             duration of terminal tonal section (b-c); interval be-
ultradirectional("shotgun") microphone with wind            tween terminaltonal section and beginningof isolated
sock. Informationon recordingsof other species re-          pulse (c-d); and intervalfrombeginningof that pulse to
ferredto below is summarized Miller (1983a, b) and          the start of pulsing in the next call (d-e). Data are
Veprintsev(1982). Sound spectrograms   were prepared        summarizedin Tab. 2. The variables (a-d) and (d-e)
on a Kay ElemetricsDigital Sona-Graph   7800. We used       were taken as estimatesof the durationof Rhythmically
various tape speeds and both linear and logarithmic         Repeated Calls (RRCs) and Inter-callIntervals(ICIs),
frequency scales to investigate fine structure. Low tape    for comparisonwith other species. Pulse rates were
speed and a logarithmicfrequencydisplaywere partic-         based on the numberof pulses per 125 ms for both C.
ularly useful in emphasizing fine temporal features         fuscicollisand the Broad-billedSandpiperLimicolafal-
while increasingtheir contrastwith backgroundnoise.         cinellus; the samples were taken working backwards
Oscillogramsand power spectra were prepared with            fromthe beginningof the terminaltonal portionof each
MicroSpeechLab (MSL,version2.2a),a softwarepack-            RRC (see Fig. 8E). For other species with pulsed
age for digitizing and analyzing sound signals. MSL was     RRCs, the final 250 ms were used to estimate pulse
runon an IBM XT PersonalComputer,with a Hewlett-            rate.
PackardThinkjetprinter.We filteredinputsignalswith             The adjustedcoefficient of variation(V*) was com-
a Krohn-Hitemodel 3750 filter below 500 and above           puted for several variablesusing the formulaof Sokal
8000 Hz, at 24 dB per octave. Signals were dig-              and Rohlf (1981): V* = V(1+(1/4n)) (see Fig. 9).

         from SpeechTechnologyResearchCentre, Universityof Victoria,P.O. Box 1700, Victoria,B.C. CanadaV8W 2Y2.

258                                                                                                ORNIS SCANDINAVICA 19:4 (1988)
Fig. 2. Calidrisbairdii,                  A-1
termination sequenceof
Rhythmically  Repeated           , S
Calls (RRCs) followed by
Song; the Song ends in
Chatter,whichis overlainby                                               '                    :
                                          A-2                          1^[' :Wii'1 t
                                                                               i         It        1 1
part of LaplandLongspur              4,
Calcarius lapponicussong.            3
A - Complete RRC-Song-           u
Chattersequence;parts 1                                                i   .     !   .    .   '.

and 2 overlap(arrowsmark                        3~   ~       ~..gI ~
                                                                 g     I         : 1                 Iti t'   Ci
where they meet) (analyzing
filter bandwidth, Hz).
                  300                                -
Part of Chatteris analyzed                               i
furtherin Fig. 4. B -
Oscillogram same
sequence (signaldigitizedat
20 kHz). Time markers,500

Fig. 3. Calidrisbairdii,
furtheranalysisof song. A -
End of one pulse train,
entire couplet train, and
beginningof second pulse
trainof Song, analyzedat
half speed over 0-8000 Hz
(analyzingfilter bandwith        :
(afb), 300 Hz). Marked
portionsare analyzed                 I
furtherin other partsof this
Fig. B - Last six couplets in    c
trainand beginningof
second pulse train (first
segmentmarkedin part A),
analyzedat quarterspeed
over 0-8000 Hz (afb, 1200
Hz). C- As part B,
analyzedon logarithmic
frequencyscale. D - End of
first pulse train, entire
couplettrain, and beginning
of second pulse trainin part
A, analyzedon logarithmic
frequencyscale at half speed
over 0-8000 Hz (afb, 90
Hz). E - Oscillogram of
beginningof second pulse
trainshown in part A
corresponds the second
segmentmarkedin part A);
F - Powerspectrumof eight
pulses markedin part E (for
E and F, signaldigitizedat
20 kHz; FFT based on 1024
points). G - Oscillogram of
five couplets, from couplet
trainshownin partA.
Markedportionsare ana-
lyzed furtherin part H. H -
Power spectraof first and
second pulseswithincouplet
markedin part G (for G
and H, signaldigitizedat 20
kHz; FFTswere each based
on 256 points). Time
markers ms.                                                                Frequency - kHz

ORNIS SCANDINAVICA 19:4 (1988)                                                                                     259
                      Ali                                      ?

         A                      i            *       *                 B~~~~~~~~~~~~~;i                                     Fig. 4. Calidrisbairdii,
                 <,         ,   -                1        'I                                                                furtheranalysisof Chatter
     4H                                                                                                                     (from Chatterterminating
     2                                                                                           00I.                       Song shown in Fig. 2). A-F
                 1    1.4
                                *0 ,e
                                 0.                      i.
                                                                                '.           1
                                                                                                                            - Portions of Chatter
                                                                                                                            0-8000 Hz using different
                                                                                                                            tape speeds and analyzing
                                                                                                                            filter bandwidths (afb) (A -
                                                                                                                            half speed, afb 600 Hz; B -
I                                                                                                                           quarterspeed, afb 1200 Hz;
                                                                                                                            C - half speed, logarithmic
I                                                                                                                           frequencyscale, afb 600 Hz;
                                                                                                                            D - quarter speed,
                                                                                                                            afb 1200Hz; E - half speed,
                      '                                                                                                               frequencyscale,
IL   4-E
     2 ....           *-F               5I       F                                                                      I
                                                                                                                            afb 90 Hz; F - quarter
                                                                                                                            speed, logarithmic   frequency
                                                                                                                            scale, afb 180 Hz). G-
                                        '                                                                                                 of
                                                                                                                            Oscillogram triplet(signal
                                                                                                                            digitizedat 20 kHz). H, I -
          4  ~                                                                                    X             p   *       Oscillogram(H) of three
                      $, . . *.
                             ?2-,                        .......                     1g                                     tripletsand power spectra

                                                                                                                            (I) of the first and third
                                I0                                                               10                         pulses (as markedin part H)
                                                                                                                            of one triplet (signal
             ?J~                                         I"    (       .             .                                  I
                                                                                                                            digitizedat 20 kHz; FFTs
                                                                                                                            were each based on 256
                                                                                                                            points). Time markersin

                                * pbeo

                                                                   6   i                 i
                                                                                                        3   4           5
                                                                                  kH2 -
                                                                               Frequency - kHz

Results                                                                              tion of each pulse became increasingly dominant as an
                                                                                     RRC progressed, and was strikingly so by an RRC's end
Vocalizations Baird'sSandpiper
                                                                                     (Fig. 1D). Pulses were about 7-10 ms long (slightly
C. bairdii uttered four kinds of calls that were immedi-                             briefer at the start) with intervals between of about 4-5
ately identifiable as classes distinguished for calidridines                         ms.
in previous research: Rhythmically Repeated Calls,                                      The fundamental frequency of RRCs was about 1-1.5
Song, Chatter, and Laugh.                                                            kHz; frequency increased then levelled off throughout
                                                                                     each call (Fig. 1A, B, D). Most energy was in the
Rhythmically Repeated Calls                                                          second harmonic (Fig. 1A, G).
Unpaired males of many calidridine species engage in                                    While in aerial display, male C. bairdii fluttered the
lengthy aerial displays over their future nesting area                               wings continuously in a loose deep flutter, in the man-
(Miller 1979). Male C. bairdii gave such displays, and                               ner of the Stilt Sandpiper Calidris himantopus (Miller
uttered long series of calls (RRCs) during them. These                               1983b) and Surfbird Aphriza virgata (Miller et al. 1987).
RRCs were about 700 ms long with intervals between
them of about 300 ms, thus were emitted at a rate of                                 Song
about 60 min-' during unbroken sequences (Tab. 1).                                   A complex kind of calidridine vocalization ("song")
Drury (1961, p. 196) described them as "long-drawn-                                  punctuates RRC sequences, occurs when a displaying
out, slurred, hoarse tooowee-tooowee calls."                                         bird's attention is directed towards another bird, and is
   RRCs were composed of a series of pulses with a                                   usually given during descent (Miller 1983a, b). It is also
repetition rate of about 77 Hz (Fig. 1, Tab. 1). Early                               used in other contexts.
pulses in an RRC were low in amplitude, and each pulse                                  Song of C. bairdii lasted several seconds long and
had a rapid rise then fall in frequency. Pulse amplitude                             included two alternating types of elements plus Chatter
increased throughout the course of an RRC, then lev-                                 (see below; Fig. 2). The first type of element was a
elled off near the end. The descending-frequency por-                                simple sequence of pulses of about 250-650 ms duration

260                                                                                                                                          19:4
                                                                                                                            ORNISSCANDINAVICA (1988)
Fig. 5. Calidrisbairdii,
Laugh. A - One-element
laugh (analyzingfilter
bandwidth(afb), 300 Hz).
"2"were 50 and 40 ms long,
            their power
spectraare shownin part E.
B - Two-element Laugh            94

(afb, 300 Hz). C- Same
Laughas in part A, shown         I'
on logarithmic  frequency        e
scale (afb, 45 Hz). D-           ma
Oscillogram laughshown
in partsA and C. E - Power
spectraof markedsegments
in part A; fundamental
frequencyand second har-
monic are markedfor each
(signaldigitizedat 20 kHz;                                          0         1             2        3          4            5
FFTswere each based on
1,024 points). Time markers                                                               Frwqu ncy- kHz
in ms.

(Figs 2 and 3). Each pulse was 3-4 ms long with in-            In C. bairdii we only heard Chatter in association
tervening intervals of 3-4 ms; the pulse repetition rate     with Song: Chatter always terminated Songs. Chatter
was about 160 Hz. The dominant frequency of this pulse       exhibited sequential grading in amplitude, repetition
train was 2-2.3 kHz (Fig. 3).                                rate, and frequency (Fig. 2). It consisted of rapidly
   The second type of Song element was a rapid series of     repeated triplets of pulses. Each triplet was about 30-35
pulse pairs. Each pulse pair was about 10 ms long with       ms long, separated from adjacent triplets by 15 ms.
intervals between of 7 ms; the first and second pulses       Pulse durations were about (in sequence) 5, 4 and 10 ms
were each 4 ms long, separated by about 2 ms (Fig. 3).       long, with intervening intervals of 3 and 10 ms. The
The first pulse of each pair had a distinctive harmonic      fundamental frequency of the first and second pulses in
structure with a strong fundamental frequency of             each triplet was about 2.5 Hz, and the pulses showed no
around 1.4 kHz. In contrast, the second pulse had a          higher harmonics. The third pulse was distinctly differ-
higher fundamental frequency (around 2.2-2.3 kHz)            ent, with a fundamental around 1.5 kHz but with most
and no detectable higher harmonics (Fig. 3H).                energy in the second harmonic (Fig. 4).
   Songs typically ended (and sometimes began) with
Chatter                                                      Laughs were uttered by birds early in the breeding cycle
Chatter is a prominent kind of call in several calidridine   when mildly disturbed by our presence. They sounded
species (Miller 1983a, b). It is frequently associated       similar to the Laugh of Aphriza though the elements
with Song but also occurs by itself, so is treated as a      were longer, and closely resembled the Laugh of Red
distinct class of vocalization.                              Knot Calidris canutus (see Cramp 1983).

                                             .~~~~~~~O-                               n    U~w    w   ww   w   ww   w   w

Fig. 6. Calidris
hoot sequenceof displaying
male. A - Logarithmic
frequencyscale, analyzing
filter bandwidth(afb) 11.3       I
Hz. B - Linear frequency
scale, afb 75 Hz. C -                 500'   -     -~   -            ~
                                                                  JCJIJJr         ~               TJJr
Logarithmic  frequencyscale,
afb 22.5 Hz. D - Linear                          O~~~~~~~~~1
frequencyscale, afb 150 Hz.
Time marker,1 s.

ORNIS SCANDINAVICA 19:4 (1988)                                                                                              261
                                                                                            Fig. 7. Tryngites
                                                                                            displaying ("gulp"
                                       II~~~~~~~~~~~I                                       followed three"ticks")
                                                                                            (analyzing bandwidth,
                                                                                            300Hz). Timemarker   in

        1-                              ,I                  I                -

                  t~~~~~~.~g 4l

  We analyzed 18 Laughs of one presumed male C.                 constant-frequency  portion. Later hoots had an initial
bairdii.Thirteenof these consistedof two elements,and           portionat 325-350 Hz, then showed a sharpfrequency
the other five consisted of only one. First (or lone)           rise to just below 450 Hz (Fig. 6).
elementsaveraged531 ms long (S.D. = 44.1), intervals
between elements averaged145 + 13.1 ms, and second
                                                                             of            Sandpiper
elementsaveraged425 ? 24.4 ms. Each element began
at around 1.5 kHz, droppedquicklyby about 500 Hz,               Displays  of male Tryngiteshave been describedby Par-
then rose gently in frequencyand amplitudeuntil near            melee et al. (1967) and Myers (1979). During these
the end when frequencyincreasedrapidlyto 1.5-2 kHz              displaysmalesutteredtwo kindsof wide-bandsounds:a
(Fig. 5). The second harmoniccontained much more                brief "tick"sound and a slightlylonger frequency-mod-
energy than the fundamental.                                    ulated sound (Fig. 7). The latter was uttered when
                                                                males engagedin a verticalpumpingaction of the head
                                                                and neck ("gulping" Parmeleeet al. 1967, p. 32). The
                    Sandpiper                                   soundscould only be heardover short distances.Males
We describe here only the well known "hooting"of                were silentwhenthey jumpedin theirlow flutterdisplay
male C. melanotos;the species uttersmanyother kinds             (a ritualizedform of flutter-fighting;
                                                                                                     Miller 1984).
of vocalizationsthat deserve specificstudy.
   Aerial displays of male C. melanotos are different
from those of most calidridinesin being low to the              Discussion
ground(only a meter or two high), in involvingdirected          Phylogeneticconsiderations
flightwith continuouswingbeats,and in havingdistinct
boutsof calls (hoots). Malesbegansuchdisplayssilently           RRCs are homologous as a rhythmicallyrepeated,
and, througha curious"pumping"       action involvingthe        structurallysimpletype of call. A dichotomyinto pulsed
wings, somehow expandedthe lower throat and upper               and non-pulsedRRCs seems to be distinct,thoughsev-
chest. The expansionappearedas a pendulantlaterally             eral species (e.g., C. himantopus,Long-toed Stint C.
compressed"dewlap".When the expansionwas com-                   subminuta)show strong rhythmic modulation of the
plete, a hoot series was initiatedwith one to severalsoft       carrierfrequency(Fig. 8). Repeated, loud pulsed calls
calls. The hoots then became louder. Hoots were not             are also used in long-distanceadvertisementby other
utteredin synchrony   with the slow deep wingbeatsused          species (e.g., Common Nighthawk Chordeilesminor;
during the display (contra Pitelka 1959 and Myers               Davis 1962), includingother Scolopacidae(e.g., Asiatic
1982).                                                          Dowitcher Limnodromus semipalmatus, Eurasian
   Displaying males uttered up to 28 hoots per series           Woodcock Scolopax rusticola, American Woodcock
(Pitelka 1959, our observations).Correspondingly,    du-        Philohelaminor;see Ferrand(1983), Fiebig and Jander
rationsof hoot serieswere up to about8 s (thatshownin           (1985), Veprintsev (1982), and references in Miller
Fig. 6 was 6.5 s long). Early hoots in each series were         (1984)). However, non-pulsed RRCs are much more
longest in duration.In the hoot series illustrated Fig.                         in
                                                                widelydistributed the Scolopacidae,occurring trin-
6, the secondhoot (the firstwas too faintfor an accurate        gines, Willet Catoptrophorus semipalmatus, godwits,
estimate) was about 320 ms long, and the duration               and curlews (Glutz von Blotzheim et al. 1975, 1977,
declinedto about 190ms for the seventhhoot. Duration            Sordahl 1979, Cramp1983, Miller 1984), so we judge
declined slowly thereafter, reaching a minimum of               them to be the form from which pulsed RRCs have
about 160 ms for the terminalhoot.                              evolved. By this reasoning,pulsedRRCs in calidridines
   Hoots were low in frequency.Early hoots in a series          musthave evolved independently  fromthose of L. semi-
were between 300-325 Hz in frequency, and each                  palmatus, Scolopax, and Philohela. If pulsed RRCs are
showed a slight rise in frequencyfollowed by a nearly           a derived character, then one phylogenetic grouping

262                                                                                         ORNIS SCANDINAVICA 19:4 (1988)
Tab. 1. Descriptivestatisticsfor temporalattributes Rhyth- tion. Several calidridine species that lack pulsed RRCs
micallyRepeatedCalls(RRCs) in displayflightsof Calidridinia. have strongly ritualized bouts of RRCs, notably the
                                                              Western Sandpiper Calidris mauri and C. melanotos
                         Duration      Duration Terminal
                          of RRC         of ICIb   buzz rate  (and presumably also the Sharp-tailed Sandpiper C.
                            (ms)          (ms)        (Hz)    acuminata); their bouts are similar in showing sequen-
                                                              tial grading from long to brief calls, which also succes-
GroupI                                                        sively increase slightly in frequency (Holmes 1973,
  1. Aphriza virgatac 326?31.0          109+7.8               Myers 1982, Veprintsev 1982). Aphriza (and presum-
                            (107)          (95)               ably also the Great Knot Calidris tenuirostris) also ut-
  2. Calidris canutus 1113?85.5        153?21.6
                             (91)          (78)               ters RRC bouts, each followed by Song (Miller et al.
  3. C. ferruginea       903?42.9       121?378               1987). It is not clear to us whether the Little Stint
                              (7)           (6)               Calidris minuta also utters RRCs in bouts (see Fig. 1 of
  4. C. himantopus       726?86.0      446?97.9
                                          (104)               Cramp 1983, p. 308). Only C. mauri and C. melanotos
                            (140)                             seem similar enough in this attribute to be placed to-
  5. C. minutillad           367            103
  6. C. ruficollis       516?12.2      352+40.1               gether phylogenetically.
                             (43)          (42)                   RRCs comprise one to several parts. In species with
  7. C. subminuta        627?19.5          235
                                            (1)               non-pulsed RRCs, differentiation has occurred at the
                                                              beginning of RRCs, with the derivation of several brief
Group   II                                                    introductory elements there (Aphriza, C. mauri, Least
  8. Calidris alba       458+36.4      864?193.6 44?1.3       Sandpiper C. minutilla). Rhythmic frequency modula-
                             (94)          (92)        (99)   tion occurs at the beginning of RRCs in other species
  9. C. alpina           609?90.8      159+44.8      79+3.4
                            (491)         (422)       (500)   (e.g., Rufous-necked Stint Calidris ruficollis), and this
10. C. bairdii           690?99.3      280?138.9 77?6.8       has probably provided the raw material for such deriv-
                             (28)          (22)        (37)   ations. C. canutus has a unique kind of complex non-
11. C. fuscicollis       186?12.5           305e     74+3.2
                             (60)                      (61)   pulsed RRC, with its two portions distinguished by fun-
12. C. ptilocnemis       522+79.4      596?164.6 42?2.0       damental frequency and harmonic structure (Fig. 8K).
                             (26)          (23)        (25)   We are not comfortable with proposing any phyloge-
13. Eurynorhynchus                                            netic groupings using these attributes.
     pygmeus             492?29.6        79?4.1     165?5.2       Several species do not have RRCs: Semipalmated
                              (4)           (4)         (5)
14. Limicola                                                  Sandpiper Calidris pusilla, Temminck's Stint C. tem-
     falcinellus         324+28.3        48+3.6      48?0.0   minckii, Ruff Philomachus pugnax, and Tryngites. The
                             (53)          (38)        (53)   latter two species have highly divergent signalling sys-
                                                              tems because of their lek social organization, so the
a. Each entry is Y + S.D. (N).
b. ICI = Inter-call Interval (interval between successive absence of RRCs for them may just reflect conver-
    RRCs).                                                    gence. The absence of RRCs in Calidris pusilla and C.
c. Data from Milleret al. (1987).                             temminckii may be a shared derived homology, howev-
d. Data from Miller (1986). N was > 2,000. See Tab. 1 of er.
    Miller(1986) for detailedstatisticalsummary.
e. For C. fuscicollis,this estimatewas computedas the sumof       Punctuated Fluttering (PF) is a distinctive flight mode
    the mean intervals(c-d) and (d-e) (see Fig. 8 and Tab.2). used by several species during DFs, including C. alpina,
                                                               C. minutilla, C. ptilocnemis (and presumably C. mar-
                                                               itima), and C. pusilla. Species known to lack PF, based
                                                               on our observations, all share a deep loose flutter in
includes Sanderling Calidris alba, Dunlin C. alpina, C. DFs: Aphriza, C. bairdii, and C. himantopus. We are
bairdii, C. fuscicollis, Rock Sandpiper C. ptilocnemis,        surprised at the incongruence between distributions of
Spoon-billed Sandpiper Eurynorhynchus pygmeus, and one of these flight modes and pulsing of RRCs.
 Limicola (the Purple Sandpiper C. maritima will pre-             In summary, derived states of RRCs include rhythmic
sumably also fall here, when its RRCs are analysed).
The only obvious candidate for a group nested within Tab.2. Descriptivestatisticsfor temporalattributes Rhyth-    of
 this one includes C. fuscicollis and Limicola, which have micallyRepeated Calls in displayflightsof C. fuscicollis.
 differentiated RRCs with "tonal" (non-pulsed) termina- Variable*                        Y             S.D.          N
 tion (see Fig. 8).
    RRCs occur in discrete bouts or trains in various Duration                         252 ms          12.0          60
 species of scolopacids including Wandering Tattler Het- Duration(a-c)                 186 ms          12.5          60
 eroscelus incanus and Lesser Yellowlegs Tringaflavipes        Duration(b-c)            23.5 ms         4.64         61
                                                               Interval(c-d)            54 ms           8.8          62
 (Miller 1984). The evolutionary transition between            Interval(d-e)           251 ms          21.9          52
 trains and ill defined long calling sequences seems, a Pulse rate                      74 Hz           3.2          61
 priori, to be an easy one, so the possession of one or the
 other state may carry little useful phylogenetic informa- *Forexplanationsee Fig. 8 and Methods.

ORNIS SCANDINAVICA 19:4 (1988)                                                                                      263
                                                                                                                                Fig. 8. Calidridinespecies,
                                                                                                                                Rhythmically  Repeated
                                                                                                                                Calls (RRCs). A - Calidris
  i                                                      k
                                                                                                                                fuscicollis(two RRCs
                                                                                                                                separatedby natural
                                                                                                                                points for measurements
  Ii .'
                                    F"'         T1           i
                                                                                                                                (see Methods)). B - C. alba.
                                                                                                                                C - C. ptilocnemis. D -
                                                                                                                                C. alpina. E - Limicola
                                                                                                                                C. subminuta.  G-
                                                                                    --- -- ---
                                                                                                                                C. himantopus.H-
   1::-                   -                                            ...*: i i                             !                  C. ruficollis(two RRCs
                                                                                                                                separatedby natural
                                J                                                                            L
                                                                                                                                interval). I - Aphriza
  4-      ..
           ........                   Ia             .           .          ..                                                  virgata.J - C. minutilla.
                                                                                                                                K - C. canutus. L -
      1      -
          v|IB-               Ss.-^        t-                        *'.*          air'--
                                                                                                                       -        C. ferruginea.Analyzing
                                                                                                                                filter bandwidth,300 Hz.
                                                                                                                                Time marker500 ms.

FM, pulsing, organization into bouts, and differentia-                                           (Fig. 9). Such a trend may signify that there are general
tion into parts. The latter (plus bout organization?) has                                        constraints to evolution, and may reveal that characters
probably evolved several times. PF is probably derived                                           of low variability contain most phylogenetic informa-
relative to continuous deep loose fluttering in DFs                                              tion.
(Aphriza glides when vocalizing in DFs). To refine phy-                                             A comprehensive discussion of relationships based on
logenetic analysis of this group, it will be necessary to                                        acoustic and anatomical characters would be inappro-
analyze more call types in other species, such as Laugh,                                         priate at this stage of our knowledge. Some of our
Chatter, and brooding calls. These are more difficult to                                         findings are not consistent with generally accepted rela-
record, but are far more evolutionarily conservative                                             tionships within the Calidridini, for example. Informa-
than Song (especially) or RRCs, and will permit in-                                              tion on more call types and more species, and on other
clusion of deviant species like Philomachus and Tryn-
gites, and of species that have lost certain call types
(e.g., C. pusilla, C. temminckii). The fine structure of                                                           0
calls should also be revealing. In C. bairdii both Song                                              50-
and Chatter show fine details of frequency and timing                                                                                * calls pulsed
                                                                                                                                     o calls not pulsed
that are at or below the limits of avian perceptual reso-
lution. For example, pulse duration and inter-pulse in-                                              40-
tervals are only a few ms long, and in Chatter adjacent
pulses differ strikingly in fundamental frequency and
harmonic structure. Presumably the characteristics of
                                                                                                 V   30-           0
pulse structure in these two call types are perceived in
an integrated manner; this would tend to conserve such
features as couplet or triplet organization through evo-
lutionary time. Certainly the structure of Chatter in C.                                                          0
bairdii is strikingly similar to that in C. pusilla ("motor-                                                                    a
boat sound"; Miller 1983b) and C. minutilla (Miller                                                                             0
1983a). Further fine analyses of calls for other species is
therefore likely to be revealing.
   Finally, further research should be directed towards                                                                         0            t-
comparative analysis of character variation in sounds.

In an earlier study, the interval between RRCs of C.                                                             ICI          RRC         Pulse
minutilla was found to be more variable than RRC                                                                                          rate
duration, and frequency characteristics of RRCs were                                             Fig. 9. Calidridinespecies, relationshipof coefficientof var-
even less variable; the latter is in keeping with a broad                                        iationto variabletype. Data for Calidris
                                                                                                                                        minutilla for totals
trend in birds. Analysis of the limited material available                                       on variables T1 and T3 in Tab. 1 of Miller (1986). Short
to us extends this analysis and reveals that RRC dura-                                           horizontalbars representmedianvalues for each variable.V*
                                                                                                 = adjusted coefficient of variation (see Methods); RRC =
tion is less variable than the interval between RRCs,                                            durationof Rhythmically  Repeated Call; ICI = inter-callin-
while the pulse rate in pulsed RRCs is even less variable                                        terval (between successiveRRCs).

264                                                                                                                             ORNIS SCANDINAVICA 19:4 (1988)
                                      *8                                                  strong, species such as Limicola, whose RRCs have a
     800-                                                   * calls pulsed                broad bandwidth, would be predicted to have shorter
                                                            ocalls not pulsed
                                                                                          average signalling distances.
                                                                                             RRCs have a median duration of 519 ms for species
     600-                                      12
                                                                                          listed in Tab. 1, with a range of about 200-1100 ms; ICIs
I                                                                                         have a median duration of 197 ms within a range of
                                                                                          50-900 ms. Durations of RRCs are thus slightly posi-
c 400-
                                                                                          tively skewed, and of ICIs strongly so (Fig. 10). Consid-

     300-         .11                                                                     ered together, these estimates reveal that RRCs ac-
                                                          10                              count for a median percentage of 77% of total time
     200                                             96
                                                    9.                                    during calling sequences (this figure includes estimates
     100                  10 04                                        30                 of 80% for bouts of C. melanotos and 92% for C.
                                                          t 9            ?       .
                                                                                          mauri). The distribution is negatively skewed, with only
                         .T                .         .
            100   200   300     400       500       600   700   800   900 1000 11001200
                                                                                          3/18 species below 50%, and 13/18 above 70% (Fig. 11).
                                Duration of RRC- msec
                                                                                          There are no differences between species with pulsed
                                                                                          and non-pulsed RRCs on these measures. It appears
                   species, bivariateplot of meandurationof
Fig. 10. Calidridine
the interval between RhythmicallyRepeated Calls (RRCs)                                    that most species use a strategy of densely "packing"
(ICI = inter-call interval) against mean durationof RRC.                                  their sound transmissions, regardless of RRC duration.
Numbers refer to species listed in Tab. 1. Sticks show dis-                                  Calidridine Song and Chatter cannot be analyzed as
persionof species along each axis. Joint medianis shownas a
cross.                                                                                    simply as RRCs. In some species (e.g., Aphriza, C.
                                                                                          alpina, C. ptilocnemis) the loud Song is transmitted
                                                                                          effectively over long distances and exhibits high sequen-
                                                                                          tial redundancy based on brief, rhythmically repeated
kinds of characters, is needed to resolve and clarify                                     subunits (these show sequential grading in C. alpina and
relationships in the group.                                                               C. ptilocnemis) (Miller 1983b, Miller et al. 1987, E. H.
                                                                                          Miller, unpubl. data). In most other species Song has
                                                                                          much less internal repetition, has longer subunits, and
Adaptiveconsiderations                                                                    generally is more structurally complex (e.g., C. himan-
                                                                                          topus, C. minutilla; Miller 1983a,b). Chatter, like Song,
 In the preceding section we summarized information                                       has been well described for very few species, but it is
which shows that a range of vocal attributes, from entire                                 organized consistently as long series of rapidly repeated
classes of vocalization to the fine structure of acoustic                                 couplets, triplets, quadruplets, etc. (Miller 1983a,b). In
 organization, have been retained in different calidridine                                C. pusilla, RRCs are lacking and Chatter ("motorboat
species over evolutionary time.                                                           sound") has replaced them as the major long-distance
    RRCs are the most important kind of long-distance,
undirected, broadcast sound signal in most calidridines.                                            100-
Their brevity and stereotypy, and their rhythmic re-                                                        -w                      13
                                                                                                     90-               14.                                     0
peated utterance, enable listeners to integrate informa-                                                                      4
 tion over successive calls and thus to accurately receive                                           80-    -P
                                                                                                                                   49 1
information about a caller's location and attributes                                                 70-
 (e.g., individuality, size). The rapid rhythmic repetition                                          60-                                0              07
of RRCs probably permits reception of details, though                                      0                                            5

the importance of details in complex RRCs (e.g., in C.                                               50-                                    *12
fuscicollis and C. minutilla) may be to nearby listeners                                   oc 40-                *11
rather than distant ones. Stereotypy of RRCs is equiv-                                         o-                                  *8             * calls pulsed
                                                                                                     30-                                          o calls not pulsed
alent to high correlational redundancy (e.g., frequency
characteristics of RRCs in C. minutilla are highly inter-                                            20-
correlated; Miller 1986); rhythmic repetition of ster-                                                10-
eotyped calls leads to high temporal predictability or                                                            i I'              r (                     I , I         i
sequential redundancy. Both aspects of redundancy                                                                200         400            600        800         1000       1200
should promote accurate reception over long distances
                                                                                                               Duration of RRC-msec
in a noisy environment.
                                                                                          Fig. 11. Calidridine                    of
                                                                                                             species, relationship the percentageof
    RRCs in some species (e.g., C. canutus, Curlew                                        time occupied by Rhythmically Repeated Calls (RRCs) during
Sandpiper Calidris ferruginea) are narrow and fairly                                      display flights, to call duration (ICI = inter-call interval).
constant in bandwidth, attributes that may be adapta-                                     Numbers refer to species listed in Tab. 1. Sticks show dis-
tions to minimize frequency-dependent attenuation                                         persionof species along each axis. Calidrismauri(W) and C.
                                                                                          melanotos(P) are shownon verticalaxis(at 92.1%and79.6%,
over distance (Miller and Baker 1980, Wiley and Ri-
                                                                                          respectively;see text). Joint median, for 16 values on the
chards 1982, Miller 1983a,b). If this correlation is                                      verticalaxis and 14 on the horizontal,is shown as a cross.

18 ORNIS SCANDINAVICA 19:4 (1988)                                                                                                                                              265
sound signal (Miller 1983b). The comments made ear-           cussed on the behavior of communicating, rather than
lier about high sequential redundancy being an effective      on the physical nature of the sounds used.
way to transmit RRCs can also be applied to Chatter
(and to some species' Song), with a much briefer in-                             -
                                                              Acknowledgements We thank S. F. MacLean,Jr., and Phil
                                                              Martinfor theirhospitality the CanningRiverdelta, Joe T.
tegration time required. However, Song and Chatter                    for
                                                              Marshall his thoughtful   criticism some of our ideas, Craig
appear to have more diverse uses in calidridine commu-        Dickson for encouraging  and advisingus on the use of MSL,
nication than do RRCs, so the relationships of structure      Erica Bates for her help throughoutthe analysisand writing,
to function cannot be adequately discussed until more         and Ed Longpresfor photographing figures. This project
detailed behavioral observations are available.               was supportedby the NaturalSciences and EngineeringRe-
                                                              search Council of Canada (individual operating grants to
   Loffredo and Borgia (1986) recently surveyed acous-        EHM) andthe Friendsof the RoyalBritishColumbia     Museum
tic characteristics and mating systems of birds and con-      (researchgrantsto EHM).
cluded that sound signals of polygynous species tend to
be noisier, broader in bandwidth, and less melodic than
in monogamous species. Our findings disagree with
theirs: we believe that phylogenetic relationships and        Baker,A. J. and Hockey, P. A. R. 1984.Behavioral vocaland
transmission distance are the main determinants of               affinitiesof the AfricanBlack Oystercatcher    (Haematopus
acoustic characteristics. For example, in the phyloge-           moquini).- WilsonBull. 96: 656-671.
                                                              Bergstrom,P. W. 1988.Breedingdisplaysand vocalizations       of
netic grouping we proposed based on the harsh-sound-             Wilson'sPlover.- WilsonBull. 100 (in press).
ing, pulsed RRCs, are a polygynous species (C. fuscicol-      Boswall, J. and Veprintsev,B. N. 1985. Keys to identifying
                                                                 Little Curlew.- Am. Birds39: 251-254.
lis), a polyandrous species (C. alba), and monogamous         Cramp,S. (ed.). 1983. Handbookof the Birdsof Europe, the
species (remaining species listed in Tab. 1; Pitelka et al.      Middle East and North Africa. The Birds of the Western
1974). RRC bandwidth is greatest for Limicola in this            Palearctic.Vol. 3, Wadersto Gulls. - OxfordUniv. Press,
group, and the remaining species differ little from one          Oxford.
another. Among these species with the ancestral non-          Davis, L. I. 1962. Acousticevidenceof relationship Capri-
                                                                 mulgus.TexasJ. Sci. 14: 72-106.
pulsed form of RRC are the polygynous C. ferruginea           Drury,W. H., Jr. 1961.The breedingbiologyof shorebirds       on
and C. melanotos; most of the other species are mono-            Bylot Island, Northwest Territories,Canada. - Auk 78:
gamous, and all have similar RRCs.                               176-217.
   The preceding contrasts were made for RRCs, a type         Ferrand,Y. 1983. A behavioralhypothesisderived from 5-
of sound used for long-distance transmission, so signall-
                                                                 year'sobservations rodingwoodcock.- In: Kalchreuter,
                                                                 H. (ed.). Second EuropeanWoodcockand Snipe Work-
ing distance was "held constant" for the comparisons.            shop. Int. Waterfowl    Res. Bur., Slimbridge, 68-82.
Fewer comparisons are possible for short transmission         Fiebig, J. and Jander, G. 1985. Der Steppenschlammlaufer,
distances. The latter permit a variety of sound charac-          Limnodromus                    als
                                                                                 semipalmatus, mongolischer      Brutvogel.-
                                                                 Ann. Orn. 9: 107-111.
teristics to be used because no effective degradation or      Glutz von Blotzheim, U. N., Bauer, K. M. and Bezzel, E.
attenuation occurs. Broad-band sounds are used by the            (eds). 1975. Handbuchder Vogel Mitteleuropas.Band 6.
polygynous Tryngites and Great Snipe Gallinago media             Charadriiformes Teil). - AkademischeVerlagsgesell-
                                                                 schaft, Wiesbaden.
(Loffredo and Borgia 1986), and graded acoustic sys-           - , Bauer, K. M. and Bezzel, E. (eds). 1977. Handbuchder
tems occur in the polyandrous jacanas and phalaropes             Vogel Mitteleuropas.    Band 7. Charadriiformes Teil). -
(Miller 1984), in keeping with the permissive role               AkademischeVerlagsgesellschaft,     Wiesbaden.
played by short distances in the evolution of sound           Gochfeld,M., Burger,    J. andJehl, J. R., Jr. 1984.The classifi-
characteristics.                                                 cation of the shorebirds the world.- In: Burger,J. and
                                                                 Olla, B. L. (eds). Shorebirds:  BreedingBehaviorand Pop-
   A relationship between mating system and acoustic             ulations.- PlenumPubl. Corp., New York, pp. 1-15.
characteristics in Scolopacidae may exist, but it is con-                               J.
                                                              Hayman,P., Marchant, andPrater,T. 1986.Shorebirds.          An
founded by phylogenetic relationships and demands of             Identification Guide to the Wadersof the World.- Croom
                                                                 Helm, London.
long-distance signalling. Furthermore, the nature and         Holmes, R. T. 1973. Social behaviourof the WesternSand-
intensity of competition for mates, and the kinds of             piper. - Ibis 115: 107-123.
information encoded in different sound classes must be                     A.
                                                              Kistchinskii, A. 1974.The biologyandbehaviorof Pectoral
crucial to the functioning of any such relationship; rele-       Sandpipers the tundras easternSiberia.- Byull. Mosk.
                                                                             in             of
vant detailed behavioral studies have simply not been            Obshch.Ispyt. Prir. (Otd.Biol.) 79: 73-88. (In Russian.)
                                                              Loffredo,C. A. and Borgia,G. 1986.Sexualselection,mating
done. We are pessimistic that any meaningful pattern             systems, and the evolution of avian acousticaldisplays.-
would emerge, in any case. It is hard to imagine sound           Am. Nat. 128: 773-794.
classes of scolopacids that are harsher-sounding or           Maclean, G. L. 1985. Roberts' Birds of SouthernAfrica. -
broader in bandwidth than Song of Aphriza or C. hi-              Trustees of the John Voelcker Bird Book Fund, Cape
mantopus, more complex than in C. bairdii, C. himanto-        Miller,E. H. 1979.Functions displayflightsby malesof the
pus, C. pusilla, or dowitchers Limnodromus, or more              Least Sandpiper,Calidrisminutilla(Vieill.), on Sable Is-
melodic than in C. minutilla - all monogamous species            land, Nova Scotia. - Can. J. Zool. 57: 876-893.
                                                               - 1983a.Structure displayflightsin the LeastSandpiper.
                                                                                   of                                       -
(Miller 1983a,b, Miller et al. 1983, 1984, 1987).                Condor85: 220-242.
   In our view, to explore the relationships of mating         - 1983b.The structureof aerial displaysin three species of
system to acoustic signalling, attention should be fo-           Calidridinae  (Scolopacidae).- Auk 100: 440-451.

266                                                                                           ORNIS SCANDINAVICA 19:4 (1988)
 - 1984. Communication in breeding shorebirds. - In: Burger,           Guide. Waders: Godwits, Phalaropes, Sandpipers. - Melo-
   J. and Olla, B. L. (eds). Shorebirds: Breeding Behavior             diya. All-Union Studio for Recorded Sound, Moscow.
    and Populations. - Plenum Publ. Corp., New York, pp.
 - 1985. Parental behavior in the Least Sandpiper. - Can. J.        Appendix: Recent literature on shorebird acoustics
    Zool. 63: 1593-1601.
 - 1986. Components of variation in nuptial calls of the Least      A review of published sound analyses (mainly sound spectro-
    Sandpiper (Calidris minutilla; Aves, Scolopacidae). - Syst.     grams) for shorebirds is in Miller (1984) (76 references). Addi-
    Zool. 35: 400-413.                                              tional references follow, and are indicated by the reference
 - and Baker, A. J. 1980. Displays of the Magellanic Oyster-        number in parentheses after each species. Taxa for which anal-
    catcher. - Wilson Bull. 92: 149-168.                            yses did not previously exist are denoted by an asterisk. My
 - , Gunn, W. W. H. and Harris, R. E. 1983. Geographic              earlier reference to Calidris minuta song in the paper by Tik-
    variation in the aerial song of the Short-billed Dowitcher      honov and Fokin (1981, their Fig. 1E) is incorrect; it should be
    (Aves, Scolopacidae). - Can. J. Zool. 61: 2191-2198.            C. temminckii (their figure legend is incorrect; P. Tomkovich in
 - , Gunn, W. W. H. and MacLean, S. F., Jr. 1987. Breeding          litt.). This is corrected as ref. 15 below.
    vocalizations of the Surfbird. - Condor 89: 406-412.
 - , Gunn, W. W. H., Myers, J. P. and Veprintsev, B. N.             Burhinidae
    1984. Species-distinctiveness of Long-billed Dowitcher            *Burhinus capensis (6), *B. vermiculatus (6)
    song (Aves: Scolopacidae). - Proc. Biol. Soc. Wash. 97:         Charadriidae
    804-811.                                                          Charadrius hiaticula (6), *Ch. mongolus (6), *Ch. pecuarius
Miskelly, C. M. 1987. The identy of the Hakawai. - Notornis           (6), *Ch. tricollaris (6), *Ch. wilsonia (2), Pluvialis apricaria
    34: 95-116.                                                       (12), P. dominica (6), P. squatarola (6), *Vanellus albiceps
Myers, J. P. 1979. Leks, sex, and Buff-breasted Sandpipers. -         (6), *V. armatus (6), *V. coronatus (6), *V. crassirostris (6),
    Am. Birds 33: 823-825.                                            *V. melanopterus (6), *V. senegallus (6), V. vanellus (12).
 - 1982. The promiscuous Pectoral Sandpiper. - Am. Birds            Glareolidae
    36: 119-122.                                                      *Cursorius rufus (6), Glareola nordmanni (6), *G. nuchalis
Nethersole-Thompson, D. and Nethersole-Thompson, M.                   (6), G. pratincola (6), Rhinoptilus cinctus (6).
    1986. Waders: Their Breeding, Haunts and Watchers. - T.         Haematopodidae
    and A. D. Poyser, Calton, U.K.                                    *Haematopus moquini (6), H. ostralegus (6, 12).
Parmelee, D. F., Stephens, H. A. and Schmidt, R. H. 1967.           Jacanidae
    The birds of southeastern Victoria Island and adjacent            *Actophilornis africanus (6)
    small islands. - Nat. Mus. Canada Bull. 222: 1-229.             Recurvirostridae
Pitelka, F. A. 1959. Numbers, breeding shedule, and territo-          Himantopus himantopus (6), H. mexicanus (13), Recurvi-
    riality in Pectoral Sandpipers of northern Alaska. - Condor       rostra americana (13), R. avosetta(6)
    61: 233-264.                                                    *Rostratulidae
 - , Holmes, R. T. and MacLean, S. F., Jr. 1974. Ecology and          *Rostratula benghalensis (6)
    evolution of social organization in arctic sandpipers. - Am.    Scolopacidae
    Zool. 14: 185-204.                                                *Aphriza virgata (9, 10), Arenaria interpres (6), Calidris alba
Sokal, R. R. and Rohlf, F. J. 1981. Biometry. The Principles          (6, 9), C. alpina (6, 9, 12), *C. bairdii (6, 9), C. canutus (6,
    and Practice of Statistics in Biological Research. 2nd rev.       9), C. ferruginea (6, 9), C. fuscicollis (9), C. himantopus (9),
    ed. - W. H. Freeman, San Francisco.                               C. melanotos (6, 9), C. minuta (6), C. minutilla (7, 8, 9), *C.
Sordahl, T. A. 1979. Vocalizations and behavior of the Willet.        ptilocnemis (9), *C. ruficollis (6, 9), *C. subminuta (6, 9), C.
    - Wilson Bull. 91: 551-574.                                       temminckii (6, 12, 15, 16), Coenocorypha aucklandica (11),
 - 1986. Evolutionary aspects of avian distraction display: var-      Gallinago gallinago (12, 17), G. media (6), *G. nigripennis
    iation in American Avocet and Black-necked Stilt antipred-        (6), Limicola falcinellus (9, 14), *Limnodromus semipalma-
    ator behavior. - In: Mitchell, R. W. and Thompson, N. S.          tus (5), Limosa limosa (6), Numenius arquata (6, 12), N.
    (eds). Deception: Perspectives on Human and Nonhuman              minutus (3), N. phaeopus (6, 12), Phalaropus lobatus (6),
    Deceit. State University of New York Press, Buffalo. pp.          Ph. tricolor (6), Scolopax rusticola (4), Tringa erythropus
    87-112.                                                           (6), T. glareola (6, 12), T. hypoleucos (6), T. melanoleuca
Svensson, B. W. 1987. Structure and vocalizations of display          (12), T. nebularia (6, 12), T. ochropus (6, 12), T. stagnatilis
    flights in the Broad-billed Sandpiper Limicola falcinellus. -     (6), T. totanus (6, 12), *Tryngitessubruficollis (6, 9), Xenus
    Ornis Scand. 18: 47-52.                                           cinereus (6)
Thielcke, G. A. 1976. Bird Sounds. - Univ. Michigan Press,
    Ann Arbor.                                                      References: 1 - Baker and Hockey (1984); 2 - Bergstrom
Tikhonov, A. V. and Fokin, S. Yu. 1981. Lek and calls of            (1988); 3 - Boswall and Veprintsev (1985); 4 - Ferrand (1983);
    shorebirds. - Ornitologia 16: 64-72. (In Russian.)              5 - Fiebig and Jander (1985); 6 - Maclean (1985); 7 - Miller
Tomkovich, P. S. and Fokin, S. Yu. 1984. Behavior of Tem-           (1985); 8 - Miller (1986); 9 - Miller et al. (the current paper);
    minck's Stint during the breeding period. - In: Sokolov, V.     10 - Miller et al. (1987); 11 - Miskelly (1987); 12 - Nethersole-
    E. (ed.). Communication and Ecology of Mammals and              Thompson and Nethersole-Thompson (1986); 13 - Sordahl
    Birds. Nauka, Moscow. pp. 208-226. (In Russian.)                (1986); 14 - Svensson (1987); 15 - Tikhonov and Fokin (1981);
Veprintsev, B. N. 1982. Birds of the Soviet Union. A Sound          16 - Tomkovich and Fokin (1984); 17 - Thielcke (1976).

18* ORNIS SCANDINAVICA 19:4 (1988)                                                                                                267

To top