amphibians

The totally aquatic African pipid frog Xenopus borealis produces a range of acoustic signals underwater at night.

The repertoire of males in heightened reproductive condition consists of three call types. All of the calls are composed of the same impulsive, click-like components. The clicks have a rise-time of 0.5 msec, a duration of 2-5 msec, and most of their energy concentrated at 2600 Hz with a secondary peak at 1100 Hz.

The sound pressure levels average 109 dB SPL at 1 meter. The advertisement call is characterized by interclick intervals of 300-600 msec (depending on temperature) and very low coefficients of variation - 3%-10%). Phonotaxis experiments confirm that it is effective in attracting females. The approach call, produced when swimming toward or clasping another frog, has interclick intervals averaging 105 msec. Males show pronounced agonistic behavior accompanied by series of clicks with interclick intervals averaging 43 msec.

Unreceptive females sometimes produce a very weak release call when clasped, but are otherwise silent.
Comparison of this underwater acoustic communication system with terrestrial anuran systems shows surprisingly few differences, given the strong contrast in acoustic environments. The structure of the male's clicks, however, suggests that a major adaptation to underwater signalling may involve the sound production mechanism itself.

Calls produced by hybrids resulting from laboratory crosses of tetraploid Hyla versicolor females and either diploid Hyla chrysoscelis (type 1) or Hyla arborea (type II) males were induced through manipulation of environmental conditions. Type I hybrids produced trilled calls similar in note repetition to H. versicolor, but more similar in dominant frequency to H. chrysoscelis. Mean duration was shorter than in both parent calls. Type II hybrids produced calls which were longer in duration and lower in note repetition rate than H. versicolor, but shorter in duration and higher in note repetition rate than H. arborea. Dominant frequency of type II hybrids was lower than in H. arborea but not significantly different than in H. versicolor. Hybrid calls were not strictly intermediate, and may provide information regarding parental relationships.

In two-choice discrimination experiments females of Hyperolius marmoratus preferred the calls of lower frequency of the pair of stimuli. This preference was not shown in mating patterns observed in natural choruses, but is when females are phonotactically orienting in small choruses in an experimental enclosure. With an increase in chorus size, the mating pattern shifts from size-based, non-random (with some evidence of size-assortative) mating to random mating. This is the first time that frequency-based mate-choice by female anurans has been associated with chorus size, and hence with the sonic complexity of the acoustic environment.

The mating calls of the Iberian midwife toads, A. o. boscai and A. cisternasii show clear differences. We calls of A. o. boscai have a shorter duration (104.8 ms) and a lower fundamental frequency (1.33 kHz) than those of A. cisternasii (172.0 ms and 1.45 kHz), between 12° and 16°C. In both species signal duration was found to be influenced by temperature.

The reed frog Hyperolius tuberilinguis is a prolonged breeder with an advertisement call that varies in complexity from one to six click notes. Call complexity increases with chorus size, but calls containing more than three notes are rare. In playback experiments to males, subjects responded by increasing the complexity of their calls, without closely matching the stimulus and rarely exceeding the stimulus in complexity. Stimuli less complex than their own evoked a reduction in complexity. Call repetition rate remained unchanged in the responses. In two-choice phonotaxis experiments, females discriminated against one-note calls, and two- and three-note calls were the most attractive. Males thus adjust their calling in the presence of neighbours to a pattern most preferred by females. Calls of higher complexity may be more easily detected or located by females in the noisy environment of a chorus.

The acoustic properties of the environment influence sound propagation. Many previous studies examined whether various species of anurans, birds and mammals adjust usage and / or structure of their vocal signals to limit degradation during propagation in this environment (“acoustic adaptation hypothesis”). The present review examines how widespread such adaptations actually are across taxa. First, evidence or environment-related adjustments in usage of vocal signals is collected from studies in birds and other vertebrates (i.e., anurans and mammals). Second, a meta-analysis conducted by Boncoraglio & Saino (2007) on the influences of the environment on the acoustic structure of avian vocalisations is taken as a reference, and results from additional studies in birds are reviewed and compared to its conclusions. Finally, evidence from similar studies conducted in anurans and mammals is collected and discussed. Concerning the usage of vocal signals, evidence of environment-related adaptations in the few studies found was more widespread in anurans and mammals than in birds. Regarding structure of vocal signals, evidence from additional studies in birds did not completely confirm results of the meta-analysis of Boncoraglio & Saino (2007). Pooling all bird studies together presented minimum frequency, frequency modulations and frequency range as acoustic variables most often adjusted to the environment, in contrast to temporal features, repetition rate and maximum frequency. The few studies conducted in anurans and mammals did not allow the identification of specific acoustic variables that typically show environment-related variations. Overall, evidence for the acoustic adaptation hypothesis was not as widespread as expected across taxa. The different aspects of vocal behaviour adapted to environmental conditions varied according to the species and local habitats. Environment-related adjustments in structure of vocal signals seem to be constrained by call function in anurans and mammals. This effect was not examined in birds, but vocal learning does not appear to be a pre-requisite to environment-related adjustment in this group.

Phonemic restoration, a form of temporal induction, occurs when the human brain compensates for masked or missing portions of speech by filling in obscured or non-existent sounds. We tested for temporal induction and related abilities in females of the Gray Treefrog Hyla versicolor. The number of pulses in calls is used by females for assessment of males. Accordingly, an ability to “restore” or interpolate between masked or otherwise sonically degraded portions of calls could help females during mate choice in noisy choruses. In phonotaxis experiments, we employed unmodifed calls and those that had a centrally placed gap, a region overlapped by a portion of another call or filtered noise, or replaced with filtered noise. When offered call alternatives with equivalent numbers of clear pulses, we found that females discriminated against calls with gaps two or more times greater than the natural 25 ms interpulse interval. When a gap was replaced with a zone of call overlap or noise (so, again the call durations of the alternatives were unequal), females discriminated either in favour (overlap) of the modified stimuli or failed to discriminate (noise). However, when the unmodified and modified stimuli were the same duration, females discriminated against the latter. Normal calls were also chosen when paired against calls with multiple noise sections. Pulses formed from noise bursts were attractive, but less so than normal pulses. In single speaker tests, standardized rates of movement did not differ between calls containing noise segments of different duration. Our results therefore do not indicate that females of the Gray Treefrog employ a form of temporal induction that is fully restorative. However, the data indicate that acoustically anomalous sections of calls can retain attractive potential provided acoustic energy and pulses are present.