In many vertebrate species, evolution has favored morphological innovations that allow an animal to simultaneously speak with more than one voice. By independently controlling multiple sound-producing organs, frogs, songbirds and even chimpanzees can belt out ‘two-voiced’ calls that contain chaotic non-linear properties. Such features are thought to make vocalizations more noticeable for animals on the receiving end of a song or call. Among vertebrates, this ability has previously only been documented in tetrapods (limbed vertebrates); however, a new study by Aaron Rice, Bruce Land and Andrew Bass at Cornell University shows that the ability to produce acoustic chaos has also evolved in the largest group of vertebrates, the fishes.
Rice and colleagues chose to work on the three-spined toadfish (Batrachomoeus trispinosus), an intensively studied vocal fish. These whiskered, lumpy swimmers produce a range of ‘grunts’ and ‘hoots’ for communicating with rivals and mates in muddy-bottomed estuaries. Unique among fish, this species makes sounds with a swim bladder that is separated into two parts – each with its own independently innervated vocal muscle.
First the researchers housed multiple toadfish together and recorded grunts and hoots. Spectrograms of toadfish calls showed linear harmonic (bands of power at multiples of a dominant frequency) components. But about one-third of all calls also showed non-linearities. There were frequency bands at a consistent fraction of an original harmonic series (subharmonics), quick jumps in all frequency components, harmonics plus ‘noisy’ energy across many frequencies (deterministic chaos), splits in harmonic bands (bifurcation) and, in some cases, overlap of two harmonically independent ‘voices’ (biphonation). The team then used a set of analytical tools derived from chaos theory to show that these toadfish non-linearities are indeed very similar to those observed in tetrapods. These results show that toadfish can generate highly complex vocalizations comparable to those seen in frogs, songbirds and mammals.
Rice and colleagues then wanted to know whether toadfish generate chaotic calls by independently controlling the two parts of their swim bladders. To test this, the team recorded grunts in individual toadfish before and after transecting the vocal motor nerve leading to one side of their bladder. After surgery, the animals could only produce calls with linear components. These experiments suggest that independent neural control of a bilateral swim bladder is ultimately what gives toadfish their two voices.
The experiments in this paper are elegant and the writing is clear. But one frustrating thing about this study is that the reader never gets to actually hear a toadfish grunt and hoot. While it may not have been necessary from an analytical point of view, presenting audible examples of the chaotic calls themselves would have been a very simple way to add richness to this paper and get more people excited about the work.
Lack of audible grunting aside, the work of Rice and co-workers is important because it suggests that strong selection pressures favor innovations that enable non-linear signalling across all major lineages of vocal vertebrates. It is now apparent that this is a powerful force shaping the evolution of acoustic communication across vertebrate taxa. Finally, this work is beautiful because it shows that even something as simple as a fish grunt can contain a stunning level of complexity.
