Supporting information for

Diverse reproductive barriers in hybridising crickets suggests extensive variation in the evolution and maintenance of isolation

Supporting information A: Calling song analysis

Calling songs are produced by males to attract females. Previous studies (Thorson et al 1982; Simmons 1988; Doherty 1991) found that mate choice was influenced by the number of pulses per chirp, the chirp length and the time between chirps (interchirp length) (see Fig. 1). In this study we assessed female response to conspecific and heterospecific male calling song. We constructed a synthetic song for both species made out of song characteristics typical for each. We chose to use a species-specific synthetic song to control for variation between individuals and increase our power to detect differences between the species (Gray 2005, Jang and Gerhardt 2006). Estimates of the three variables (Table 1) were obtained from analyzing songs containing long sets of chirps from 26 G. bimaculatus and 10 G. campestris males, all recorded at 28  1 C. Background noise was removed using the noise removal filter in Audacity (freely available at Subsequently, the time between each pulse peak was determined using a custom JAVA program, which allows the distance between the beginning and end of pulses to be measured from .wav file waveforms. Parameter means for each individual’s song were used to calculate the species’ means. For each species we randomly selected a chirp consisting of the mean number of pulses of the focal species, and with an interpulse length matching that of the species mean. This chirp was repeated with the mean interchirp length to acquire the desired synthetic song.

Using a single standardised song has the benefit that it removes variance in mate choice due to the signal. However, if the particular chirp we selected is atypical, our findings may not be as general as we expect. We avoid this possibility because we measure many of the acoustic parameters of the song that are available to the animals and subsequently constructed an accurate single song for both species. In order to make a synthetic song, we rounded the mean number of pulses per song, which led to a greater divergence between our synthetic songs compared to the mean pulse number of the two species. This might lead to stronger preferences, but there is also the possibility that the rounded number mismatches the peak maximum of the absolute preference function of the female.

Table 1. The mean number of pulses per chirp, chirp length and interchirp length for the calling song of males of both species. The number of pulses per chirp for the synthetic song of G. bimaculatus and G. campestris was set to three and four, respectively. The mean chirp length was calculated for chirps with this species’ typical number. Standard deviations between brackets.

G. bimaculatusG. campestris

Pulses per chirp3.24 (0.54)3.71 (0.43)

Chirp length (ms)74.94 (9.93)90.84 (6.18)

Interchirp length (ms)278.19 (85.87)277.29 (52.97)

Figure 1. The waveform of the calling song of a G. campestris (A) and G. bimaculatus (B) male used for playback. Each discrete sound pulse is interspaced with an interpulse length making up one chirp of a discrete number of pulses and specific length (chirp length). Each chirp is separated from the next by a silence, the interchirp length.

Supporting information B: The trackball system

The phonotactic response of G. campestris, G. bimaculatus and hybrid females to the two synthetic songs was measured using a trackball system (D-sphere) (Wagner 1998; Green and Tregenza 2009). The focal female cricket is positioned on top of a very lightweight (approximately 3.4 g) hollow polystyrene sphere by means of a thin metal wire. At the bottom, the metal wire is attached to the cricket’s pronotum (by means of a modelling wax), and at the top to a metal stand placed above the centre of the sphere (Fig. 1A). Song was played from two speakers placed at a 70 angle from the animal’s length at a distance of 50 cm (Fig. 1B). The sphere rotates freely on a cushion of air pumped into a metal hemisphere in which it resides. The sphere’s movement is recorded by two optical sensors at the equator of the sphere and mounted at a 90 angle one to another on the front and side of the metal hemisphere. When the cricket responds phonotactically it rotates the sphere by walking on it. The trackball system has a number of advantages as compared to playback arena assays. The major advantage of the trackball set-up is that the female cannot move towards one of the two songs and therefore rapidly change the relative amplitude of the two. Additionally, by preventing the female from turning, we extend the point at which she makes a decision about which male to steer towards. If the female were allowed to turn, she would then experience a difference in the sound coming from the two speakers in favour of whichever speaker she had turned to face. Furthermore, multiple songs can be played consecutively, and the response recorded in quick succession. Both cricket species have been successfully used on this system (Schmitz et al 1982; Hedwig and Poulet 2005; Green and Tregenza 2009).

Figure 1. Schematic side view of the trackball system (A) and a top view of the trackball including the speakers used for playback (B).

General notes on statistical information

R takes the first level of a factor as it reference; hence parameters estimates with standard errors (s.e.) are only presented for the number of levels minus one. For example, ‘side’ has two levels, left and right and R thus only presents the estimates of the right side (relative to the left). In each level we give the levels of factors as used in the analyses. The estimates are presented from the summary of the model. P values from these models are not presented as these represent paired comparisons between the first factor level and subsequent ones, which we do separately. In the model selection columns, d.f. gives the degrees of freedom of the two models compared, l.r. is the likelihood ratio, ‘#’ indicates the order in which explanatory variables have been excluded (starting with 1).

Supporting information C: No-choice experiment

Linear-mixed model (lmm) results of long-range signal no-choice experiment.

Run (1st and 2nd), side (left, right), female identity (FemaleID; G. bimaculatus (‘B’), G. campestris(‘C’) and hybrid (‘H’)) and song type (conspecific and heterospecific).

lmmModel selection

Full modelEstimates.e.d.f.tΔ d.f.l.r.p#

Intercept7.112.462322.893

Run3.781.342322.827, 67.920.0053

Side-8.411.34232-6.267, 636.84< 0 .0013

FemaleID C-7.582.3281-3.277, 511.600.0033

FemaleID H-4.852.0781-2.343

Song type-1.791.34231-1.348, 71.810.1792

FemaleID C x Song type0.673.492290.1910, 82.530.2821

FemaleID H x Song type -4.303.08229-1.401

Contrasts

G. bimaculatus - G. campestris

Intercept6.833.001582.283

Run3.541.681582.106, 54.430.0353

Side-7.141.68158-4.246, 517.35< 0.0013

FemaleID C-7.562.6056-2.916, 58.040.0053

Song type-0.421.69157-0.257, 60.060.7992

FemaleID C x Song type0.573.661560.158, 70.020.8761

G. bimaculatus - hybrid

Intercept7.242.901842.493

Run4.331.641842.636, 76.910.0093

Side-10.311.64184-6.276, 736.22< 0.0013

FemaleID H-4.852.3262-2.096, 74.310.0383

Song type-2.371.63183-1.457, 62.140.1432

FemaleID H x Song type-4.573.32182-1.388, 71.940.1641

G. campestris - hybrid

Intercept7.242.901842.493

Run4.331.641842.636, 76.910.0093

Side-10.311.64184-6.276, 736.22< 0.0013

FemaleID H-4.852.3262-2.096, 74.310.0383

Song type-2.371.63183-1.457, 62.140.1432

FemaleID H x Song type-4.573.32182-1.388, 71.940.1641

Supporting information D: Two-choice experiment

Linear-mixed model (lmm) results of long-range signal two-choice experiment.

Run (1st and 2nd), and side conspecific (‘left’, ‘right’) and female identity (FemaleID; G. bimaculatus (‘B’), G. campestris(‘C’) and hybrid (‘H’)).

lmmModel selection

Full modelEstimates.e.d.f.tΔ d.f.l.r.p#

Intercept6.821.39834.902

Side conspecific-8.891.9674-4.533, 419.27< 0.0012

FemaleID C-2.912.7081-1.084, 61.490.4741

FemaleID H0.322.39810.131

Contrast

G. bimaculatus - G. campestris

Intercept6.691.78573.762

Side conspecific-9.452.4950-3.803, 413.52< 0.0012

FemaleID C-2.982.9256-1.024, 51.060.3011

G. bimaculatus – hybrid

Intercept8.961.72635.202

Side conspecific-11.892.3960-4.993, 422.61< 0.0012

FemaleID H0.332.57620.134, 50.020.8971

G. campestris - hybrid

Intercept2.331.01462.313

Side conspecific-3.532.0037-1.763, 43.120.0772

FemaleID H2.572.04441.261

Supporting information E: Short-range mate choice

Generalised linear-mixed model (glmm) results of short-range signal experiment as summarised in Table 1 in the main text. Explanatory variable is male/female species identity (MaleID/FemaleID; G. bimaculatus (‘B’), G. campestris(‘C’) and hybrid (‘H’)).Theparing types are female x male species.

SongglmmModel selection

Full modelEstimates.e.Zχ2Δ d.f.Pr(>χ2)#

Intercept1.610.413.891

MaleID C-1.480.54-2.7211.722, 40.0031

MaleIDH-1.900.68-2.791

Contrasts

G. bimaculatus x G. bimaculatus vs. G. bimaculatus x G. campestris

Intercept1.610.413.891

MaleID C-1.480.54-2.727.982, 30.0051

G. bimaculatus x G. bimaculatusvs. G. bimaculatus x hybrid

Intercept1.610.413.891

maleID H-1.900.68-2.798.122, 30.0051

G. bimaculatus x G. campestrisvs. G. bimaculatus x hybrid

Intercept< 0.0010.30< 0.0012

maleID H-0.410.65-0.640.412, 30.5211

MountingglmmModel selection

Full modelEstimates.e.Zχ2Δ d.f.Pr(>χ2)#

Intercept0.410.331.261

MaleID C-1.240.51-2.416.352, 40.0411

MaleIDH-0.960.63-1.531

Contrasts

G. bimaculatus x G. bimaculatusvs. G. bimaculatus x G. campestris

Intercept0.420.331.271

MaleID C-1.250.52-2.425.792, 30.0161

G. bimaculatus x G. bimaculatusvs. G. bimaculatus x hybrid

Intercept0.140.260.522

maleID H-0.900.60-1.482.272, 30.1321

G. bimaculatus x G. campestrisvs. G. bimaculatus x hybrid

Intercept-0.690.31-2.262

maleID H0.280.640.430.192, 30.6661

MatingglmmModel selection

Full modelEstimates.e.Zχ2Δ d.f.Pr(>χ2)#

Intercept-0.300.32-0.941

MaleID C-1.730.65-2.689.442, 40.0091

MaleID H-1.250.74-1.681

Contrasts

G. bimaculatus x G. bimaculatusvs. G. bimaculatus x G. campestris

Intercept-0.300.32-0.941

MaleID C-1.730.64-2.688.582, 30.0031

G. bimaculatus x G. bimaculatusvs. G. bimaculatus x hybrid

Intercept-0.570.27-2.072

maleID H-1.180.71-1.653.132, 30.0771

G. bimaculatus x G. campestrisvs. G. bimaculatus x hybrid

Intercept-12.2018.15-0.672

maleID H0.5136.530.010.012,30.9071

Full results of Table 1B (male G. bimaculatus)

SongglmmModel selection

Full modelEstimates.e.Zχ2Δ d.f.Pr(>χ2)#

Intercept1.120.264.222

FemaleID C-1.610.71-2.275.162, 40.0761

FemaleID H-0.570.63-0.901

Contrasts

G. bimaculatus x G. bimaculatusvs. G. campestris x G. bimaculatus

Intercept1.610.413.891

FemaleID C-1.610.71-2.275.092, 30.0241

G. bimaculatus x G. bimaculatusvs. hybrid x G. bimaculatus

Intercept1.390.314.472

FemaleID H-0.570.63-0.900.802, 30.3701

G. campestris x G. bimaculatus vs. hybrid x G. bimaculatus

Intercept0.650.361.832

FemaleID H1.040.751.391.972, 30.1611

MountingglmmModel selection

Full modelEstimates.e.Zχ2Δ d.f.Pr(>χ2)#

Intercept0.390.311.231

FemaleID C-18.953104.43-0.0118.522, 4< 0.0011

FemaleID H-0.140.51-0.281

Contrasts

G. bimaculatus x G. bimaculatusvs. G. campestris x G. bimaculatus

Intercept0.390.311.231

FemaleID C-18.953104.66-0.0117.872, 3< 0.0011

G. bimaculatus x G. bimaculatusvs. hybrid x G. bimaculatus

Intercept0.330.251.342

FemaleID H-0.140.51-0.280.082, 30.7781

G. campestris x G. bimaculatus vs. hybrid x G. bimaculatus

Intercept-19.575118.32< 0.011

FemaleID H19.815118.32< 0.0114.792, 3< 0.0011

MatingglmmModel selection

Full modelEstimates.e.Zχ2Δ d.f.Pr(>χ2)#

Intercept-0.290.31-0.921

FemaleID C-18.283104.44-0.0111.392, 40.0031

FemaleID H-0.470.53-0.881

Contrasts

G. bimaculatus x G. bimaculatusvs. G. campestris x G. bimaculatus

Intercept-0.460.25-1.821

FemaleID C-18.283104.68-0.0111.382, 3< 0.0011

G. bimaculatus x G. bimaculatusvs. hybrid x G. bimaculatus

Intercept-0.290.31-0.922

FemaleID H-0.470.53-0.880.792, 30.3751

G. campestris x G. bimaculatusvs. hybrid x G. bimaculatus

Intercept-11.4915.74-0.732

FemaleID H18.815118.33< 0.010.002, 311

Supporting information F: Number of eggs

Results of GLM analyses with quasipoisson error to assess differences in number of eggs laid between the three pairing types included: G. bimaculatusx G. bimaculatus (‘BB’), G. bimaculatusx G. campestris (‘BC’) and G. campestrisx G. campestris (‘CC’).

glmModel selection

Full modelEstimates.e.tΔ d.f.DevianceP(>|χ|2)#

Intercept5.280.1052.281

PairingBC-0.370.21-1.80103, 1054796.6< 0.0011

PairingCC0.700.135.611

Contrasts

G. bimaculatus x G. bimaculatusvs. G. bimaculatus x G. campestris

Intercept5.190.0956.902

PairingBC-0.380.22-1.7762, 63313.870.0671

G. bimaculatus x G. bimaculatusvs. G. campestris x G. campestris

Intercept5.290.1053.761

PairingCC0.700.125.7785, 863023.8< 0.0011

G. bimaculatus x G. campestris vs. G. campestris x G. campestris

Intercept4.900.1925.431

PairingCC1.090.215.2559, 603374.0< 0.0011

Supporting information G: Hatching success

Results of logistic regression to assess differences in hatching success between the three pairing types included in the analyses: G. bimaculatusx G. bimaculatus (‘BB’), G. bimaculatusx G. campestris (‘BC’) and G. campestrisx G. campestris (‘CC’).

glmModel selection

Full modelEstimates.e.ZΔ d.f.DevianceP(>|χ|2)#

Intercept-1.340.02-54.401

PairingBC-0.260.07-4.0079, 81342.4< 0.0011

PairingCC0.440.0315.391

Contrasts

G. bimaculatus x G. bimaculatusvs. G. bimaculatus x G. campestris

Intercept-1.340.02-54.401

PairingBC-0.260.07-4.0043, 4416.6< 0.0011

G. bimaculatus x G. bimaculatusvs. G. campestris x G. campestris

Intercept-1.340.02-54.401

PairingCC0.440.0315.3971, 72246.3< 0.0011

G. bimaculatus x G. campestris vs. G. campestris x G. campestris

Intercept-1.600.06-26.171

PairingCC0.700.0611.1944, 45142.6< 0.0011

Supporting information H: Reproductive barriers in Gryllidae

The ‘Other’ category in Figure 7 in the main text contains different mechanisms affecting gene flow. This barrier is present in both types of mixed-species pairing, unless stated otherwise. Only the female species of the mixed-species pairing is mentioned. If the full cross is mentioned (indicated by an ‘x’), the female species is mentioned first. (Example: sp A – sp B means both mixed-species pairing types of the two species, sp A x sp B indicates the mixed-species paring female sp A with a male sp B)

a: Differences in spermathecal duct length between the parental species reduces gene flow.

Allonemobiussp. nov. – A. socius and G. fultoni- G. vernalis.

b: Differences in cuticular hydrocarbons between the parental species.

Laupalacerasina – L.kohalensis, L. paranigra– L. kohalensis and G.campestris – G. bimaculatus.

c: Differences in seminal fluid between the two parental species.

G.firmus x G. pennsylvanicus.

d: Aggression and chemical cues differences between the parental species.

G. integer x G. lineaticeps(certain) and G. lineaticeps x G. integer (uncertain).

References

Doherty JA (1991) Song recognition and localization in the phonotaxis behavior of the field cricket, Gryllusbimaculatus (Orthoptera, Gryllidae). J. Comp. Physiol. A 168:213-222.

Gray DA(2005) Does courtship behavior contribute to species-level reproductive isolation in field crickets? Behav. Ecol. 16:201-206

Green K, TregenzaT (2009) The influence of male ejaculates on female mate search behaviour, oviposition and longevity in crickets. Anim. Behav. 77:887-892.

Hedwig B, PouletJEA(2005) Mechanisms underlying phonotactic steering in the cricket Gryllusbimaculatus revealed with a fast trackball system. J. Exp. Biol. 208:915-927.

Jang Y, Gerhardt HC (2006a) Divergence in female calling song discrimination between sympatric and allopatric populations of the southern wood cricket Gryllusfultoni (Orthoptera :Gryllidae). Behav. Ecol. Sociobiol. 60:150-158

Schmitz B, Scharstein H, WendlerG (1982)Phonotaxis in Grylluscampestris L. (Ortoptera, Gryllidae). J. Comp. Physiol. A 148:431-444.

Simmons LW(1988) The calling song of the field cricket, Gryllusbimaculatus (De Geer): constraints on transmission and its role in intermale competition and female choice. Anim. Behav. 36:380-394.

Thorson J, Weber T,HuberF(1982) Auditory behavior of the cricket. II. Simplicity of calling-song recognition in Gryllus and anomalous phonotaxis at abnormal carrier frequencies. J. Comp. Physiol. A 146:361-378.

Wagner WE(1998) Measuring female mating preferences. Anim. Behav. 55:1029-1042.

1