Michael D. Greenfield

Animal Choruses Emerge from Receiver Psychology (A Tale of Two Synchronies)

Univ. St. Étienne (ENES), France Univ. Kansas (Ecol & Evol Biol), USA

Labex CeMEB Mediterranean Center for Environment and Biodiversity What is an animal chorus ? (It’s about time)

Temporal adjustments in broadcasting at three levels of precision :

a b (an evening chorus) c d

individual e

12 18 24 6 12 h

------(collective singing * leader ------bouts) ------

0 5 10 15 min

90° phase angle (regular rhythm ------and precise phase ------relationships) 0 1 2 3 sec 0 5 10 Time (sec) Physalaemus pustulosus (Túngara frog; Anura: Leptodactylidae); 5 Male Chorus Physalaemus pustulosus (Túngara frog; Anura: Leptodactylidae); 5 Male Chorus

Individual D

-2 0 2 4 6 8 10 Time (sec)

Frogs have rules +

0 15 30 45 60

Time (sec) Magicicada cassini (Cicadidae); Periodical Cicada (17-year) Synchronous Chorus; Brood IV; June 1998; Douglas Co., Kansas Pteroptyx tener (Lampyridae);

Synchronous fireflies of the Indo-Malayan Region

Kumari Nallabumar 2002 Strogatz & Stewart 1993

Uca annulipes

(Crustacea: Ocypodidae);

Western Indo-Pacific;

Synchronized waving

Stefano Cannicci

Synchronized courtship in fiddler crabs; Backwell et al. 1998

Utetheisa ornatrix (Lepidoptera : Arctiidae)

Specialized rhythmic chorusing : potential adaptive features

* Retention of species-specific rhythm or call envelope ------* Evasion of predators ; detection of predators Time → * Maximization of collective signal intensity of local group

* Unmasking of sexually-selected signal characters ; ------ability to detect and evaluate rivals ------

(Greenfield 2005) Alternative to the adaptationist paradigm : (the null hypothesis)

Choruses might also arise simply as an ‘emergent property’ of local interactions between singing neighbors. That is, the overall chorus structure per se, even when very complex and seemingly specialized, is neutral in terms of preference by female receivers and the benefits that male signalers might accrue from generating it.

(Greenfield & Schul 2008 ; Greenfield 2015)

How might this work ?

3-step pathway :

Collective singing patterns in choruses simply emerge from the ‘receiver psychology’ of female perception and preference.

1) Females ignore male calls that follow a neighbor’s by a brief interval.

2) Males adjust call rhythm (phase) upon hearing a stimulus or neighbor.

3) When multiple males use equivalent adjustments an expansive chorus of synchrony and/or alternation may arise. Importantly, the display can be generated in the absence of any selection expressly favoring synchrony or alternation.

Loudspeaker 14 12 1 ------10 8 2 ------6 400 msec 4

Number Females 2 0 1 - 27.5 55.0 O + -- PulseSong Length (msec) 2

Neoconocephalus spiza (Orthoptera : Tettigoniidae) : female preference for song length

Phonotaxis 1 - Phase angle O + -- 2 1

- 24º

Neoconocephalus spiza : 12 female preference for 24º leading song overrides 1 song length 0 100 200 Time (msec) Phase Angle

-180° - - 1.0 -- --

0.8 -81° ------0.6 -54° - - 0.4 -- -- 0.2 -24° - -

Preference for long chirp long for Preference 0.0 -- -- -200 -150 -100 -50 0 50 100 150 200 o 0° - - Phase Angle ( ) -- -- Greenfield & Roizen 1993 24° - - 1 - -- --

O + 180° - - -- 2 -- --

400 msec

Neoconocephalus spiza - precedence effect in female phonotaxis Monte Carlo Simulation Response phase, Monte Carlo Simulation Response phase, {(T' - T) / T} • 360° InhibitoryPRC slope Resetting = s : {(T' - T) / T} • 360° InhibitoryPRC slope Resetting = s :

TT TT´ TT FocalFocal AnimalAnimal StimulusStimulus dd or Neighbor or Neighbor Time  StimulusStimulus phase, phase, Time  (d(d / /T) T) • •360360°° tt

rr Phase Response Curve (PRC)

Oscillator level

180° ' StimulusStimulus Phase Phase = = (d (d / /T) T) • • 360360°° Greenfield et al. 19971997 BasicBasic model model : : TT´ ´== (T (T + + ε )εResponse )+Response + (s (s • •d),d), Phase Phase where where = = ε {(T ε{(Tisis´´ - - a T)aT) stochastic/ stochastic /T T } } • • 360360°° elementelement 0°

FullFull model model : : Response Phase ' TT´ ´== (T (T + + ε )ε )+ + s{(d s{(d + + l l/ /v) v) - -(r(r - -t)}t)} + + (y (y --x)x) 0° 180° Stimulus Phase y (msec) _____ Neoconocephalus spiza ____ (Tettigoniidae) : ___

__ PRC intercepts adjusted by _

stimulus length Response Phase Response

Stimulus Phase Greenfield & Roizen 1993 Monte Carlo Simulation Response phase, Response phase, {(T' - T) / T} • 360° PRC slope = s Monte Carlo Simulation Response{(T' - T) / T} phase, • 360° PRC slope = s {(T ' - T) / T} • 360° PRC slope = s T T´ T Focal Animal T T´ T FocalStimulus Animal d or Neighbor Stimulus phase, Stimulus Time d Stimulus phase, (d / T) • 360° or Neighbor Stimulus(d / T) • 360 phase,° tTime  (d / T) • 360°

t r

Oscillator level

r level Oscillator Greenfield et al. 1997 Basic model : Basic model : Greenfield et al. 1997 BasicT´ model= (T + ε:) + (s • d), where ε is a stochastic element T´ = (T + ε) + (s • d), where ε is a stochastic element T´ = (T + ε) + (s • d), where ε is a stochastic element Full model : Full model : Full T model´ = (T +: ε) + s{(d + l / v) - (r - t)} + (y - x) T´ = (T + ε) + s{(d + l / v) - (r - t)} + (y - x) T´ = (T + ε) + s{(d + l / v) - (r - t)} + (y - x) Greenfield, Tourtellot & Snedden 1997 Monte Carlo Simulation

Michael Tourtellot

VisualBASIC® Maestro Monte Carlo Simulation

Spacing ; 2 males separated by 5 m ; 1 female

Calls of the 2 males ; mostly synchronized

Male song adjustment parameters Monte Carlo Simulation :

Alternation Synchrony

Greenfield et al. 1997 Male

A x

Neoconocephalus spiza (Tettigoniidae: Conocephalinae); Central America; (imperfect) synchrony x B

Time (sec) 3-step pathway :

Collective singing patterns in choruses simply emerge from the ‘receiver psychology’ of female perception and preference.

1) Females ignore male calls that follow + a neighbor’s by a brief interval. No experimental evidence for coevolution between male and female traits ? 2) Males adjust call rhythm (phase) + upon hearing a stimulus or neighbor.

Question 1 : Did male song adjustment (inhibitory resetting) coevolve with female response to relative call timing (leading vs. following calls) ? Tettigoniidae ; Bradyporinae

B Ephippiger diurnus

C Alternation + Synchrony individual

alt D D B alt syn 3 m C 5 s

Females ignore following calls ; Flightless ; negligible dispersal ; Male song adjustment (inhibitory resetting) ; narrow habitat preference ; Chorusing (alternation and synchrony) isolated, genetically differentiated populations

Ephippiger diurnus (Tettigoniidae)

Flightless ; negligible dispersal ; narrow habitat preference ; geographically isolated, genetically differentiated populations → 31 34 Application of comparative 65 09 11 phylogenetic methods possible 66

‘Comparative Method’ ; 17 E. diurnus populations ‘sampled’ ; Phylogenetic analysis indicates branching topology ; Mitochondrial (COI) data ; Best-fitting nucleotide substitution model

Clade 2

ʘ 7

Clade 1 Measurement of female

follower leader perceptual trait

Col de Chioula 0.5 Col de Chioula population

0.4 f for leader for * * * 0.3

0.2 f – maximum separation

leader

preference at which females ignore

† follower 0.1 follower delay 0.0

Index of Index 30 150 400 600 900 1200

time Leader – follower call separation (ms) Measurement of male song

t T peak adjustment trait rb basal r

Time →

T ′ T ′ level Oscillator Oscillator 1 T ′2 3 focal male

rb rb rb 1 2 3 stimulus d1 d3 m

m – minimum delay before male

resumes singing

10 600

ms bin ms

- Col de Chioula - Male # 37 Col de Chioula 500

8 (ms)

m Col de Chioula 400 6 population

Focal male does delay

300 Calls per 50 per Calls not call during 4 post-stimulus interval 200 2 100

Synthetic song 0 minimum ; m 0 stimulus at t = 0 0 200 400 600 800 1000 29 30 33 34 36 37 50 51 Post-stimulus call delay (ms) Male # Predictions :

If male song adjustments have coevolved with female perception, m should be ≥ f such that males do not broadcast ineffective following calls.

Moreover, if m ≈ f, males will maximize their call rate while at the same time

Populations: 1) Mireval, 2) Vias, 3) Peyriac de Mer, 4) Feuilla, 5) Col de Mantet, forgoing the broadcast of following calls. 6) Font Romeu, 7) Latour de Carol, 8) Col de Puymorens, 9) Hospitalet près l’Andorre, 10) Mérens-les-Vals, 11) Col de Chioula, 12) Carcanières, 13) Port de Lers, 14) Vilamòs, 15) Cigalère, 16) Le Lioran, 17) Pouzol. Col de Puymorens ; 8 Hospitalet près l’Andorre ; 9 Female precedence effect ; Maximum delay (msec) at which leading calls are preferred.

Male phase adjustment ; Minimum delay (msec) before singing after a stimulus or neighbor. A For the 17 populations sampled, m ≈ f .

B 2500

2000

1500 ρ = 0.87 1000

; minimum delay (ms) ; minimum delay

m 500 Populations: 1) Mireval, 2) Vias, 3) Peyriac de Mer, 4) Feuilla, 5) Col de Mantet, 6) Font Romeu, 7) Latour de Carol, 8) Col de Puymorens, 9) Hospitalet près l’Andorre, 10) Mérens-les-Vals, 11) Col de Chioula, 12) Carcanières, 13) Port de Lers, 14) Vilamòs, 15) Cigalère, 16) Le Lioran, 17) Pouzol. 0 0 500 1000 1500 2000 2500 f ; maximum separation (ms) Is the m – f correlation among the 17 populations an artifact of sampling ?

a) Remove ‘phylogenetic signal’ by application of phylogenetic independent contrasts (PIC)

‘working phylogeny’ of sampled E. diurnus populations – unrooted neighbor-joining (NJ) tree based on m – f correlation following microsatellite data : m, f syl

Carcanières 2.78 PIC correction 84 Col de Chioula 2.02 2000 88 Mérens-les-Vals 1.86 100 Hospitalet près l’Andorre 2.04 82 57 Vilamòs 1.42 1500 96 Cigalère 1.50 51 Port de Lers 1.15 89 Le Lioran 1.00 1000 93 Pouzol 1.02

Mireval 1.00

95 Vias 2.08 500 ρ = 0.76

Col de Mantet 4.28 Peyriac de Mer 5.28 97 0 56 Feuilla 5.85

Font Romeu 3.11

Col de Puymorens 3.28 (ms) contrasts positivized - delay minimum ; -500

100 m 50 Latour de Carol 3.64 0 500 1000 1500 2000 f ; maximum separation - positivized contrasts (ms) 600 ms 0.05 Is PIC justified for intra-specific comparison ?

Normally no, because gene flow resulting from inter-population migration will influence phenotypic vaues of a population. (Stone, Nee & Felsenstein 2011)

But, E. diurnus is a special case : Negligible dispersal ; Branching topology of population phylogeny ; Mitochondrial (COI) tree and microsatellite tree are very similar : ‘working phylogeny’ is probably the correct one

Nonetheless, we took a conservative approach : Identify genetically differentiated clusters of populations using a Bayesian clustering protocol (STRUCTURE) on microsatellite data

7 genetically differentiated clusters ; re-apply PIC on these 7 clusters K = 7 clusters

K 8

of of data [ Pr ] (X|K) 12

Probabiliity Number of clusters ; K

Population Unrooted NJ tree of 7 clusters m – f correlation following identified with STRUCTURE PIC correction

ρ = 0.94

m – f correlation is not a phylogenetic artifact Is the m – f correlation among the 17 populations an artifact of other correlations ?

b) Remove populations with long calls (≥ 3 syllables) ; confounding correlations between f and syl # and between m and syl #

ρ = 0.69

m – f correlation holds for the 10 populations having only 1-2 syllables per call ; no influence of syl # on correlation

Conclusion : Male song adjustment (m) has coevolved with female perception (f) Question 2 : Is receiver psychology (female perception of relative male call timing) important in the context of sexual selection within a chorus ? Kramer sphere (open loop) Leading calls Faster Higher syllable rhythm number

(Party, Brunel-Pons & Greenfield 2014)

Simple choice tests in 3 populations indicate female preferences for 1) leading calls, 2) faster rhythm, and 3) higher syllable number. (Party et al. 2014)

Priority of precedence : At a chorus females pay much more attention to call order than other signal characters. a. Merens-les-Vals

Paired t-tests ; p > 0.75 Question 3 : Natural Modified Do synchrony and alternation per se 2.0

have adaptive value ? 1.0

3 tests of synchrony, alternation, and 0.0 overlapping calls, as broadcast by chorusing Distance toward source (m) males, in the context of sexual selection : N1 N2 N3 M1 M2 M3 Chorus stimulus Natural chorus stimulus (Merens-les-Vals)

Modified chorus stimulus (Merens-les-Vals)

(Party, Streiff, Marin-Cudraz & Greenfield 2015) 10° phase angle Natural b) Peyriac de Mer Repeated-measures ANOVA ; p = 0.53 5 Natural Natural

4 Mod. Mod. Mod.

20° phase angle Modified 2

30° phase angle Modified 10 20 30 40 180

Chorus stimulus (phase angle, o )

2. Levels of female response to

natural (alternating and synchronous 40° phase angle Modified phase angles) and to modified chorus stimuli

(overlapping phase angles) ;

single stimulus tests on Kramer sphere ;

180° phase angle Natural Result : Differential responses to natural and modified chorus stimuli not observed (power = 0.82)

Merens-les-Vals Mireval Vilamòs Natural ; Synchrony 10° phase angle

Modified ; 30° Overlapping

Natural ; 180° Alternation

Result :

Overall preference for natural

chorus stimuli not observed Modified ; 40° Overlapping (Rayleigh and V-tests).

3. Female preference for natural (alternating or synchronous phase angles) vs modified chorus stimuli (overlapping phase angles) ; choice tests on Kramer sphere ;

Question 4 : Is chorusing under central control ?

90 86 dB 82 78 74 70 dB

78 dB

At Col de Chioula, most males pay attention to only one neighbor (90)

At Vilamós, most males pay attention to one or two neighbors (90, 86)

Question 5 : Will a population’s chorusing format impede or enhance gene flow following secondary contact between populations ? a

1 FR

2 FR

3 MER

4 FR 0 5 10 15 Time (s) a b c

14 12 12 12 10 10 10 8 8 8 6 6 6 4 4

Frequency 4 2 2 2 0 0 0 Min Maj Min Maj Min Maj Min Maj Min Maj FR MER CDM MIR LEU CDM

Same population male preferred Different population male preferred a b

12 350

10 300 250 8 200 6 150 4 Frequency 100 2 50

Mating Latency (min) Latency Mating 0 0 CC maleCC maleCC x CC male x femaleMERCC x FR femaleCCfemale female female x FRMER xmale FRFR male malemale x FRFR x femaleMER female MERfemale x MER male male x MER female CC maleCC maleCC x CC male x femaleMERCC x FR female CCfemale female female xFR MER xmaleFR FR male male male x FRFR x femaleMER female MERfemale x MER male male x MER female

Mated Unmated

c d 0.40 200

0.35 150

0.30 100

0.25 50

0.20 0

Spermatophore (proportion mlae weight) 0.15 CC maleCC maleCC x CC male x femaleMERCC x FR femaleCCfemale female female xFR MER xmaleFR FR male male male x FRFR x femaleMER female femaleMER x MER male male x MER female Number eggs deposited CC CCmale CCmale x CCmale x CCMERfemale x CCFRfemale female female FRx MER maleFRx FR male xmale FR x femaleMERMER female x MERmale malex MER female

Pairing Question 6 : What is true synchrony ? Does it exist ? Tettigoniidae ; Bradyporinae

B Sorapagus catalaunicus C individual Synchrony D

5 s

B Ephippiger diurnus C Alternation + Synchrony

alt D D B alt syn 3 m 5 s C Ephippiger diurnus Sorapagus catalaunicus

50 40 Synchrony - 2 male choruses 30 20 Percentage 10 0 1 2 3 4 Population

Ephippiger diurnus - Col de Chioula 100 80 Synchrony - 4 male choruses 60 40

Percentage 20 0 1 2 3 4 5 6 Ephippiger diurnus - Peyriac de Mer

100 80 60 40

Percentage 20 0 1 2 3 4 5 6 Ephippiger diurnus - Vilamòs

100 80 60 40

Percentage 20 0 1 2 3 4 5 6

Sorapagus catalaunicus - Font Romeu

100 80 60 0 40 2

Percentage 3 20 4 0 1 2 3 4 5 6 Chorus a Playback - precedence effect - Sorapagus catalaunicus

Right speaker

30 ms

Left speaker

1 s Precedence effect Playback - gap length

Playback - gap depth Gap Length

Gap Depth Relative call timing G

0 2.5 5

Temps (sec) Precedence effect :

Localization of the first of 2 (or more) sounds that are separated by a brief interval.

Left channel is the leader (by 7 ms). 0 50 Temps (msec) G

0 2.5 5

Temps (sec) Canal à la droite est le premier

0 50 Temps (msec) Le groupe ‘Montpellier’ Le groupe ‘Tours – St. Etienne’

Réjane Streiff (CBGP)

Virginie Party Mathieu Mahamoud-Issa

Thibaut Marin-Cudraz Yareli Esquer-Garrigos

And with thanks to : Flavia Barbosa, Guillaume Baudouin, Guy Bourdais, Odette Brunel-Pons, Marine Deluen, Séverine Devers, Marlène Goubault, Caroline Hébert, Aurelien Kerbrat, Justine Penin, Florian Plault, Darren Rebar, and Valery Terwilliger

A B 2500

2000

1500

1000

; minimum delay (ms) ; minimum delay

m 500

0 0 500 1000 1500 2000 2500 f ; maximum separation (ms)

1000

Populations: 1) Mireval, 2) Vias, 3) Peyriac de Mer, 4) Feuilla, 5) Col de Mantet, 6) Font Romeu, 7) Latour de Carol, 8) Col de Puymorens, 9) Hospitalet près 500 l’Andorre, 10) Mérens-les-Vals, 11) Col de Chioula, 12) Carcanières, 13) Port de Lers, 14) Vilamòs, 15) Cigalère, 16) Le Lioran, 17) Pouzol.

; minimum delay (ms) delay ; minimum

m 0 0 500 1000 f ; maximum separation (ms) m, f syl A Carcanières B 84 2.78 Col de Chioula 2.02 2000 88 Mérens-les-Vals 1.86 100 Hospitalet près l’Andorre 2.04 82 57 Vilamòs 1.42 1500 96 Cigalère 1.50 51 Port de Lers 1.15 89 Le Lioran 1000 93 1.00 Pouzol 1.02

Mireval 1.00

95 Vias 2.08 500

Col de Mantet 4.28 Peyriac de Mer 5.28 97 0 56 Feuilla 5.85

Font Romeu 3.11

; minimum delay - positivized contrasts (ms) (ms) contrasts positivized - delay minimum ; 100 Col de Puymorens 3.28 -500

m 50 Latour de Carol 3.64 0 500 1000 1500 2000 f ; maximum separation - positivized contrasts (ms) 600 ms 0.05

C D

Carcanières, Col de Chioula, Mérens-les-Vals, 2000 CarcaChiMerAndoHospitalet près l’Andorre 9595 LioranPouzLe Lioran, Pouzol 9696 9898 CigaLersVilaVilamòs, Cigalère, 1500 Port de Lers MireViasMireval, Vias ColmantetCol de Mantet

9898 PeyFeuPeyriac de Mer, Feuilla 1000 FontromPuyCarolFont Romeu, Latour de Carol, Col de Puymorens 500 0.050.05

; minimum delay - positivized contrasts (ms) contrasts - positivized delay ; minimum 0 500 1000 1500 2000

m f ; maximum separation - positivized contrasts (ms) A B C

10 600 0.5 Col de Chioula - Male # 37 Col de Chioula Col de Chioula 500 8 0.4 f m 400 6 0.3 300 4 0.2 200

2 0.1 100

0 0 0.0 0 200 400 600 800 1000 29 30 33 34 36 37 50 51 30 150 400 600 900 1200

Count 14 1200 1.0 Font Romeu - Male # 60 Font Romeu Font Romeu

12 ; minimum delay (ms) 1000 0.8

10 Index of Preference for Leader 800 m 0.6 f 8 600 0.4 6 400 0.2 4

2 200 0.0

0 0 -0.2 0 200 400 600 800 1000 40 44 45 50 52 55 60 65 30 150 400 600 900 1200

Post-stimulus Call Delay (ms) Male # Leader - Follower Call Separation (ms) A B 2500

2000

1500

1000

; minimum delay (ms) ; minimum delay

m 500

0 0 500 1000 1500 2000 2500 f ; maximum separation (ms)

1000

; minimum delay (ms) delay ; minimum

m 0 0 500 1000 f ; maximum separation (ms)

Hypothesis 1 : Inhibitory resetting mechanisms are favored by selection where psycho-acoustic precedence effects influence female receivers to prefer leading signals and to ignore following ones

T T´ T Focal Animal

Stimulus d Ligurotettix planum (Acrididae) or Neighbor Time 

Stimulus or neighbor at 0 sec t

Duration of precedence effect

Oscillator level

No signals by focal male during post-stimulus interval Call Delay (sec) Phase Angle

-180° - - 1.0 -- --

0.8 -81° ------0.6 -54° - - 0.4 -- -- 0.2 -24° - -

Preference for long chirp long for Preference 0.0 -- -- -200 -150 -100 -50 0 50 100 150 200 o 0° - - Phase Angle ( ) -- -- Greenfield & Roizen 1993 24° - - 1 - -- --

O + 180° - - -- 2 -- --

400 msec

Neoconocephalus spiza - precedence effect in female phonotaxis Hypothesis 1 : Inhibitory resetting mechanisms are favored by selection where psycho-acoustic precedence effects influence female receivers to prefer leading signals and to ignore following ones

T T´ T Focal Animal

Stimulus d Ligurotettix planum (Acrididae) or Neighbor Time 

Stimulus or neighbor at 0 sec t

Duration of precedence effect

Oscillator level

No signals by focal male during post-stimulus interval Call Delay (sec) Hypothesis 2 : Synchrony and alternation may arise as emergent properties of pairwise inhibitory-resetting interactions between neighboring signalers.

Imperfect Synchrony Alternation

Neoconocephalus spiza (Tettigoniidae) Ephippiger diurnus (Tettigoniidae)

T T´ T´

Rebound  T Rebound < T

Oscillator level ( Fast Rhythm ; T < 1 sec ) ( Slow Rhythm ; T > 1 sec ) A B C

10 600 0.5 Col de Chioula - Male # 37 Col de Chioula Col de Chioula 500 8 0.4 f m 400 * * 6 0.3 * 300 4 0.2 200

2 0.1 100

0 0 0.0 0 200 400 600 800 1000 29 30 33 34 36 37 50 51 30 150 400 600 900 1200

Count 14 1200 1.0 Font Romeu - Male # 60 Font Romeu Font Romeu

12 ; minimum delay (ms) 1000 0.8

10 Index of Preference for Leader * 800 m 0.6 f 8 600 0.4 6 * 400 0.2 4

2 200 0.0

0 0 -0.2 0 200 400 600 800 1000 40 44 45 50 52 55 60 65 30 150 400 600 900 1200

Post-stimulus Call Delay (ms) Male # Leader - Follower Call Separation (ms)