Summary

I investigated the scaling of echolocation call parameters (frequency, duration and repetition rate) in bats in a functional context. Low-duty-cycle bats operate with search phase cycles of usually less than 20 %. They process echoes in the time domain and are therefore intolerant of pulse-echo overlap. High-duty-cycle (>30 %) species use Doppler shift compensation, and they separate pulse and echo in the frequency domain. Call frequency scales negatively with body mass in at least five bat families. Pulse duration scales positively with mass in low-duty-cycle quasi-constant-frequency (QCF) species because the large aerial-hawking species that emit these signals fly fast in open habitats. They therefore detect distant targets and experience pulse-echo overlap later than do smaller bats. Pulse duration also scales positively with mass in the Hipposideridae, which show at least partial Doppler shift compensation. Pulse repetition rate corresponds closely with wingbeat frequency in QCF bat species that fly relatively slowly. Larger, fast-flying species often skip pulses when detecting distant targets. There is probably a trade-off between call intensity and repetition rate because ‘whispering’ bats (and hipposiderids) produce several calls per predicted wingbeat and because batches of calls are emitted per wingbeat during terminal buzzes. Severe atmospheric attenuation at high frequencies limits the range of high-frequency calls. Low-duty-cycle bats that call at high frequencies must therefore use short pulses to avoid pulse-echo overlap. Rhinolophids escape this constraint by Doppler shift compensation and, importantly, can exploit advantages associated with the emission of both high-frequency and long-duration calls. Low frequencies are unsuited for the detection of small prey, and low repetition rates may limit prey detection rates. Echolocation parameters may therefore constrain maximum body size in aerial-hawking bats.