Acoustic stability in hyrax snorts: vocal tightrope-walkers or wrathful verbal assailants?

Vocal quality allows listeners to derive information on properties of the caller such as body size and state of arousal. Male rock hyraxes advertise themselves by singing. We found that the stability of a single song element can be used to predict both body size and social rank of the singer. We also showed that other hyraxes may consider this information when deciding how to interact with the singer. Abstract The source-filter theory proposes that information on caller properties is communicated through acoustic qualities, as physical state and performance ability are reflected in the voice. Vocal stability, manifested through harshness, is especially intriguing and has rarely been explored although harsh sounds are prevalent in nature. Male rock hyraxes (Procavia capensis) produce loud complex calls that we term songs. Only the calls of older, socially dominant males include a harsh sound termed snort. As snorts are the rarest element in songs, we hypothesized that high-quality snorts are difficult to produce, and that their quality consists in the ability to maintain smoothness throughout this low-pitched, harsh call. We quantified harshness by measuring periodicity deviations and expected to find a link between social parameters (residence, rank, and weight) and the ability to produce longer, smoother snorts. In addition, we presumed that if calls are used as vocal contests, conspecifics would avoid answering songs that exhibit a higher acoustic ability than their own songs. We found that in wild hyrax songs, snort harshness was associated with both weight and social rank, but in opposite directions. Heavier males produced smoother snorts and higher-ranked individuals produced harsher snorts, possibly indicating aggressiveness. Playback experiments showed that longer and harsher synthetic snorts, inserted into natural songs, reduced conspecific answer rates. Snorts may communicate complex information on hyrax weight and dominance by means of element length and harshness. Our present results provide a stimulating insight into the understanding of acoustics in mammalian vocal communication.