Evolutionary Patterns in Sound Production across Fishes
Sound production by fishes has been recognized for millennia, but is typically regarded as comparatively rare and thus yet to be integrated into broader concepts of vertebrate evolution. We map the most comprehensive dataset of sound production yet assembled onto a family-level phylogeny of ray-finned fishes (Actinopterygii), a clade containing more than 34,000 extant species. Family-, rather than species-, level analyses allowed broad investigation of sound production mostly based on illustrations of acoustic recordings and morphological specializations (82%) strongly indicative of sound production along with qualitative descriptions (18%), and a conservative estimate of the distribution and ancestry of a character that is likely more widespread than currently known. Compilation of sonic-related morphological characters shows 60 families exhibiting muscles coupled to swim bladder vibration and 39 families that employ movement of skeletal parts against each other, i.e., stridulation. Eighteen of these families, mostly catfishes (13), include individual species exhibiting both mechanisms. The results show that families with soniferous species contain nearly two-thirds of actinopterygian species, including a clade originating circa 155 Ma, and that sound production has independently evolved approximately 33 times within Actinopterygii. Despite the uncertainties of presence-only data records and incomplete evidence of absence, under-sampling species, and assuming family-level conservation of sound production, sensitivity analyses show that these patterns of shared ancestry are robust. In aggregate, these findings offer a new perspective on the ancestry and convergent evolution of sound production among actinopterygians, a clade representing more than half of extant vertebrate species.

Soniferous behavior mapped onto phylogenetic tree of actinopterygian families. Tree shows three different lines of evidence for soniferous behavior used here and its phylogenetic distribution. Tree is pruned from species-level phylogeny of Rabosky et al. (2018) to family-level here.

Family-level phylogenetic tree of actinopterygians depicting evolution of soniferous behavior. Shown here are probabilities from ancestral-state reconstruction using stochastic character mapping. Probability is represented as a gradient, where blue indicates a high probability and red a low probability of soniferous behavior, and yellow is ∼50% probability. Tree is pruned from species-level phylogeny (Rabosky et al., 2018) to family-level here.

Probability of soniferous behavior being ancestral within major actinopterygian clades. (A) Otocephala and (B) Eupercaria. For phylogenetic trees showing the ancestral-state estimation and associated evolutionary probabilities of sound production being ancestral by stochastic character mapping, probability is represented as a gradient where blue indicates high and red is low probability of sound production; yellow is equivocal.

Family-level phylogenetic tree of actinopterygians as shown in Figure 1, but in this case mapping the distribution of three categories of soniferous mechanisms for 88 families: SBV, swim bladder vibration; STR, stridulation; non-SBV, non-swim bladder vibration (see Results section for details).

Sensitivity of ancestral-state reconstruction of soniferous fish clades to uncertainty of character states. (A) Box and whisker plot showing median and interquartile range of ancestral probabilities for different actinopterygian clades included in Table 1 related to increases or decreases in the number of soniferous fish families. (B) Variation in ancestral probabilities for each clade related to sampling uncertainty (the percentage of families within each clade with simulated uncertainty in character state).
Contributor Notes
Associate Editor: W. L. Smith.