ICI and click packet usage in relation to target range for presumed on-axis clicks for two trained bottlenose dolphins. (A,B) ICI of presumed on-axis clicks as a function of target range. ICIs from regular clicks (SAY: N=292; TRO: N=288) are shown as grey dots. For click packets found to contain a presumed on-axis click (SAY: N=141; TRO: N=67), the ICI for the first click in each packet (pre-interpacket interval, IPI) is shown as red upwards-pointing triangles, the interval between the last click in a packet and the next click (post-IPI) is shown as purple rightwards-pointing triangles, and within-packet ICIs are shown as blue downwards-pointing triangles. In A, five regular click ICIs, 22 pre-IPIs and nine post-IPIs exceed 800 ms, while in B, zero regular click ICIs, seven pre-IPIs and 17 post-IPIs exceed the 800 ms y-axis limit. The dashed line shows the TWTT. (C,D) Histograms showing the percentage of regular click ICIs above the TWTT in 10 m target range bins. (E,F) Histograms showing the percentage of combined pre-IPIs and post-IPIs above the TWTT. Empty bins in C–F are shaded in light grey.

Target range estimates based on the delay between echo return time and most recent click event as a function of time. This figure is based on the same trial as Movie 1. Regular clicks are shown as black dots and packet clicks are in blue. The inset in the upper right corner shows the sound pressure (black line, upper panel) received on the Dtag in the first 2 s of the trial with packet clicks highlighted in blue. Grey vertical dotted lines illustrate echo return times calculated using the dolphin's distance from the target at the time of each click event and the speed of sound in water. The inset also shows how the echo delay measured from the most recent click directly translates into the correct target range (black dots highlighted with black circles, lower panel) for the first three click events because ICI>TWTT. Returning echoes from click packets result in a series of underestimates of target range (black dots) except for the last packet echo that has a causal relationship with the last click in each packet and because IPI>TWTT. *Based on the results from Finneran et al. (2013). The lower dashed horizontal line indicates the approximate upper limit out to which a ∼20log₁₀(R) (where R is target range) receiving side automatic gain control (AGC) has been observed. The upper dashed horizontal line shows the upper limit where forward masking effects are known to occur following click production (depending on source level, SL).

SL of presumed on-axis clicks in relation to target range for two bottlenose dolphins. (A) SL for 291 regular clicks (grey dots) and 141 packet clicks (blue triangles) presumed on-axis for SAY. The dependence of all SLs (both regular and packet clicks) on target range is modelled as a linear regression: SL=19log₁₀(R)+170 dB re. 1 µPa (peak to peak, pp), r²=0.56, shown as a solid black line along with the 95% confidence interval (CI; grey shaded area). Inset: histograms showing the distributions of slopes and intercepts calculated for individual trials with five or more possible on-axis clicks (N=31) fitted to the model: SL=slope×log₁₀(R)+intercept. (B) SL for 288 regular clicks and 67 packet clicks presumed on-axis for TRO. The dependence of all SLs on target range is modelled as a linear regression: SL=17log₁₀(R)+171 dB re. 1 µPa (pp), r²=0.69, shown together with the 95% CI. Inset: histograms (N=21) as in A. In A and B, the SoundTrap clipping level (pp) and click detector threshold plus estimated transmission loss (TL; dotted lines) illustrate the dynamic recording range (for more information, see Fig. S1). Note that the linear regression data in this figure are included mainly for visualisation of the approximate degree of adjustment, but that LME analysis suggests significant differences between the range-dependent SL adjustment for regular clicks and packet clicks (see Table S4).