How can bats navigate in total darkness amongst trees and branches, but still locate a tiny, fluttering insect with extraordinary acuity? All made possible through echolocation, an astonishing sensory mechanism…

Bats are a hugely successful order of mammals, and much of their success is down to echolocation. The generation of signals that bounce off surrounding objects and the processing of the returning echoes allows them to determine the direction, distance and features of these objects with extraordinary acuity. This astonishing sensory mechanism enabled bats to navigate in the dark and radiate into the virtually empty niche of nocturnal insect hunter. Although they are not the only animals to have evolved echolocation (it has independently arisen in toothed whales, a few ground-dwelling mammals and some birds), bats are certainly the masters of it. Fossil evidence suggests that they could echolocate even at an early stage of their evolution. It is not yet clear whether or not the oldest species known to date (Onychonycteris finneyi) was capable of echolocation. Like its ancestors, it may have relied on sight and smell, although a recent re-analysis of an osteological character potentially involved in echolocation in modern species has suggested that Onychonycteris had already evolved this capacity. In any event, echolocation in bats provides some remarkable examples of evolutionary convergence.

Evolutionary origins of echolocation

Laryngeal echolocation

Bats are the only animals with sophisticated laryngeal echolocation, that is where ultrasonic calls are produced in the larynx. The question how many times this mechanism has evolved within the group revolves around how one chooses to interpret bat systematics, and this has experienced considerable upheavals in recent years. Traditionally, bats were split into two presumably monophyletic suborders, identified on the basis of their morphology and echolocation characteristics. Microchiroptera (microbats) are mainly insectivorous, widely distributed and capable of laryngeal echolocation. On the other hand, Megachiroptera (megabats or Old World fruit bats) feed on fruit or nectar, are restricted to tropical forests in Africa and Indo-Australasia and do not echolocate using laryngeal sounds. Recently, however, numerous molecular studies based on large, diverse datasets have provided overwhelming evidence for microbat paraphyly, so that two different suborders are now widely accepted. The Yinpterochiroptera contain the non-echolocating Old World fruit bats (Pteropodidae) and the superfamily Rhinolophoidea, which comprises five echolocating insectivorous families (Rhinolophidae, Hipposideridae, Megadermatidae, Rhinopomatidae and Craseonycteridae). This grouping is particularly surprising, as horseshoe bats (Rhinolophidae) possess the most highly developed echolocation system of all animals. The remaining twelve echolocating bat families are united in the monophyletic suborder Yangochiroptera.

This has important consequences for the evolution of laryngeal echolocation. The traditional phylogeny would have suggested a single origin in the microbat lineage, but now two different scenarios are conceivable. Either laryngeal echolocation has arisen only once in the ancestor of bats and was then lost in the pteropodid lineage (perhaps related to changes in foraging habits and roosting locations) or it has evolved independently at least twice. There is support for both hypotheses. It has been argued that the large eyes of fruit bats do not reflect an ancestral state, implying that enhanced visual capacities evolved after the loss of echolocation in this group. The slightly enlarged cochlea of fruit bats and their use of broadband calls reminiscent of echolocation signals in communication could be considered further evidence for ancestral echolocation. However, echolocation is so advantageous that it is not easy to see why it should be lost. Furthermore, rhinopholoids are united by a skeletal character that is likely to be related to echolocation but absent in other bats, which has (controversially) been considered support for an independent origin of echolocation in this group.

Looking into the molecular basis of echolocation might help to disentangle the competing hypotheses, and two potentially relevant genes have been analysed in bats so far. Foxp2 encodes a transcription factor that is involved in vocalisation control in mammals and might thus be important for producing and interpreting echolocation calls. While conserved in most mammals studied, Foxp2 was very variable in echolocating bats, but as regards the evolutionary origin of echolocation, results are ambiguous. The “mammalian hearing gene” Prestin seems to be responsible for high-frequency sensitivity and selectivity in the auditory system of mammals and could have driven the auditory specialisations needed for echolocation. Sequence analysis of its coding region in echolocating and non-echolocating bats suggested that Prestin has undergone accelerated evolution in the echolocators, thus potentially providing molecular support for the convergent evolution of laryngeal echolocation. This conclusion has, however, been greeted with controversy and the debate is still ongoing. Intriguingly, another recent study on Prestin has revealed convergent amino acid changes in the in echolocating dolphins, most likely driven by natural selection.

Non-laryngeal echolocation

Interestingly, a different form of echolocation has evolved in fruit bats in the genus Rousettus. They produce short high-pitched sounds by clicking their tongue (similar to the echolocating swiftlets), most likely for orientation rather than for foraging. It was long assumed that this method is more primitive, but Egyptian fruit bats (Rousettus egyptiacus) performed equally well in obstacle negotiation tests as laryngeal echolocators. Brief clicks might have the advantage of reducing signal bandwidth, thus concentrating energy into those frequencies the bat hears best. As Rousettus is not basal in the pteropodid lineage, it is considered unlikely that tongue-clicking echolocation represents an ancestral stage, so this mechanism probably evolved secondarily.

There is some anecdotal evidence that Asian fruit bats (Eonycteris spelaea) roosting in Malaysian caves produced wing-clapping sounds in dark situations, but it remained unclear whether this is a basic echolocation system or just a consequence of slowed flight.

Convergence in call features

The echolocation calls of bats are extremely plastic and diverse, varying in frequency, duration and amplitude among and even within species. As the signals are used to perceive the environment and different types of environment impose different perceptual challenges, call design is shaped by the bat’s ecology rather than its evolutionary history. Thus, distantly related species foraging in similar habitats have independently evolved very similar calls, representing excellent examples of convergence.

Call type

Convergent echolocation calls are found in several yinptero- and yangochiropteran families. Probably the most striking case involves the so-called high duty cycle echolocation, where an emission of signals that consist of a constant frequency component and a broadband sweep allows for efficient detection and classification of targets. Remarkably, this highly sophisticated call type as well as the corresponding processing mechanisms has evolved independently in the Old World horseshoe bats and Parnell’s moustached bat (Pteronotus parnellii) of the New World. These bats have converged upon a very similar auditory physiology, including an acoustic fovea. Here, a substantial part of the cochlea is dedicated to a very narrow frequency range, which greatly increases sensitivity (analogous to the fovea of the eye, the electrical fovea in electric fish and the tactile equivalent in the star-nosed mole Condylura cristata). There are, however, differences between rhinolophids and P. parnellii as regards call emission (nasally in the former and orally in the latter), the functional mechanisms responsible for the fovea and the organisation of the auditory cortex.

Another example is a Neotropical insectivorous myotisid (Myotis nigricans) that has converged upon temperate pipistrellid bats. They occupy similar ecological niches and produce narrowband signals when hunting in open spaces, while switching to more broadband calls in edge habitats.

Whispering echolocation

At least six bat lineages have evolved so-called whispering echolocation, emitting signals of low intensity (although recent evidence has suggested that their calls might actually be much louder than previously thought). This could help to avoid being detected by prey (at least six orders of insects have evolved sensitivity to ultrasound). Furthermore, echolocation can be ineffective in cluttered environments and bats might rather listen to prey-generated sounds, especially when gleaning insects off surfaces. For example, species in the distantly related families Megadermatidae and Nycteridae produce faint broadband signals and have large ears. So similar is their morphology that they were classified as sister taxa in earlier phylogenetic analyses.

Nasal emission

Echolocation pulses can either be emitted through the mouth or through the nose, and oral and nasal emitters differ in their skull morphology. Nasal emission has evolved from the more ancestral oral emission several times. It occurs in the Old World horseshoe bats as well as in some yangochiropteran groups, including the Old World slit-faced bats (Nycteridae) and the New World leaf-nosed bats (Phyllostomatidae). In rhinolophids, the resonating chambers are formed by the nasal cavities, whereas they lie outside the bony nasal cavity in nycterids. Phyllostomid skulls lack resonating chambers. An advantage of nasal echolocation could be that it allows for simultaneous calling and chewing, which could be particularly useful when prey is large. However, it might limit the intensity of the call, and nasal emission and whispering echolocation often co-occur.

Jamming avoidance response

Central as echolocation is to the life of most bats, it also involves a perennial risk, especially in large groups, where many individuals call simultaneously. The resulting interference between signals leads to the possibility of jamming. Electric fish, which sense amplitude and phase changes of a self-generated electric field to determine location and features of surrounding objects, face the same problem, and both groups have independently evolved a jamming avoidance response (JAR). Many bats can actively adjust their signals to separate them acoustically from those of conspecifics, changing the timing of their calls or rapidly shifting their frequency. The type of frequency shift seen in JAR can vary, depending on the proximity of the other bat. Evidently JAR is important in avoiding mid-air collisions (although, interestingly some fish-catching bats will issue a warning “honk” if on a collision course). Not surprisingly, JAR is more useful in bats with a broad frequency range, while those with narrowband calls are at less risk.