How Some Insects Detect and Avoid Being Eaten by Bats: Tactics and Countertactics of Prey and Predator

Some insects have evolved audition and evasive behaviors in response to selective pressure from bats, and other insects were preadapted to detecting ultrasonic signals. Some bats have evolved in turn, improving the range or resolution of sonar signals and serendipitously making them less detectable by insects. In other words, there is a kind of evolutionary escalation going on between bats and insects. Our aim with this review is to present the complex interactions between echolocating bats and insects with bat-detecting ears and show how these interactions may be advantageous for predator or prey. To document our examples, we cite mostly newer studies and reviews in which the reader can find references to original works. Insects occupied all terrestrial habitats at least 300 million years ago, long before bats appeared in the Eocene, about 50 million years ago. Ears have appeared independently 19 times in the class Insecta. In the period before bats, ears and complex acoustical behaviors appeared independently in at least seven orders of insects (Hoy et al. 1989, Robert et al. 1992, Yager 1999). Antibat tactics, which must have appeared in insects since the Eocene, are now known in members of four orders: Lepidoptera (moths and nocturnal butterflies), Orthoptera (crickets), Dictyoptera (praying mantids), and Neuroptera (green lacewings), and possibly also in the Diptera (flies) and Coleoptera (beetles). Insect tympanal organs, or ears, consist basically of an external, thin membrane (the tympanum) and associated internal air sacs, or tracheae. The auditory (sensory) cells attach to the tympanum or to an internal membrane (Yager 1999). Tympanal organs of most modern tympanate insects respond to a wide band of frequencies extending well into the ultrasonic range (above 20 kHz), as was probably true for pre-Eocene tympanate insects as well. Tympanate insects are physically small animals that can produce high-frequency sounds more efficiently; hence, high frequencies are used by many insects for acoustical communication between conspecifics. Consequently, many sonorous insects were preadapted to the evolution of bats (Hoy 1992). According to one possible scenario, a vast larder of nocturnal, flying insects awaited exploitation, and a flying mammal, the microchiropteran bat, was one successful exploiter. Echolocation, or biosonar, was a prerequisite for success in darkness, and even the first nocturnal bats probably used it (see Hoy 1992). Most of the nearly 700 microchiropteran bat species eat insects that they detect using biosonar (Schnitzler and Kalko 2001). However, bat biosonar has two major disadvantages: attenuation and forewarning. The frequencies used by echolocating bats range generally from 20 kHz to 100 kHz, with some outliers using frequencies below 10 kHz or above 200 kHz. Higher frequencies improve resolution, but they attenuate at a greater rate (Surlykke 1988) and the detection distance is reduced accordingly. The source level is the sound pressure level (SPL relative to 20 µPa), in decibels (dB), measured 10 cm in front of the bat's mouth. A bat using a source level of 110 dB at 20 kHz could detect the echo from an object the size of a moth at more than 5 m. Detection would occur at no more than 2.4 m if the bat used 100 kHz (Surlykke 1988). From the insects' perspective, bats advertise their presence with the ultrasonic pulses used to stroboscopically probe the environment. Thus, insects are forewarned if they can hear ultrasound. This coincidentally exerts considerable selection pressure against those insects that either cannot hear or do not react (Miller 1982). Thus, the stage was set in the Eocene for an evolutionary escalation between bats and insects. Evasive behaviors in existing tympanate insects (presumably crickets, locusts, and mantids) probably appeared in response to selection pressure by bat predation (Hoy 1992). The same selection pressure generated new auditory and motor mechanisms in presumably earless insects (green lacewings and moths). Bats, too, could have developed countermeasures, for example, shifting signals out of the prey's hearing range (Fenton and Fullard 1981) or modifying hunting behaviors (Miller and Olesen 1979). Preexisting auditory systems in insects may have been preadaptively sensitive to bat echolocation. Tympanate insects that were normally diurnal may also have become active at night. Crickets, locusts, and mantids are considered here because they were probably some of the earliest insects with hearing and they are mostly active during the day, but often fly (migrate or disperse) at twilight and at night. However, all crepuscular and nocturnal insects are potential prey for bats. The most intensively studied insect auditory system is that of field crickets (Figure 1a, left). The majority of these studies concern intraspecific communication. The tibia of each foreleg contains an ear (Figure 1a, middle). For example the maximum sensitivity of the cricket Gryllus bimaculatus occurs at about 5 kHz as measured electrophysiologically from the auditory nerve. This is also the frequency of the calling song. However, the ear is sensitive to sound frequencies up to 100 kHz at least. Popov and Shuvalov (1977) first reported that dispersing crickets avoid being hunted by bats. Since then Ron Hoy at Cornell, Andrej Popov in St. Petersburg, and their colleagues have documented avoidance behavior in several species of crickets both behaviorally and neurophysiologically (see Hoy et al. 1989). Crickets in stationary flight steer away from the source of ultrasound (negative phonotaxis), with the most effective frequencies lying between about 10 kHz and 80 kHz (Figure 1a, right). An interneuron in Teleogryllus oceanicus (Int 1) initiates evasive behavior (Hoy et al. 1989), and its homologue in Gryllus bimaculatus (AN2; Popov et al. 1994) presumably does the same. The threshold for AN2 at 20 kHz is about 20 dB less than that of the behavior (Figure 1a, right), meaning that the neural response is more sensitive than the behavior. Some mole crickets hear ultrasound, in part with special neuronal pathways, and free-flying crickets show avoidance to batlike sounds (Mason et al. 1998). Bush crickets (Figure 1b, left), like field crickets, have their ears and associated acoustic tracheae in the tibia of the forelegs (Figure 1b, middle). Some species communicate entirely in the ultrasonic range. Many bush crickets can hear bats, but few seem to react to bat echolocation. However, the bush cricket, Neoconocephalus ensiger, shows an acoustic startle response during tethered flight in the laboratory (Faure and Hoy 2000). When the insects hear intense batlike sounds with frequencies from 15 kHz to at least 60 kHz (Figure 1b, right), they dive. However, they exhibit no directionality with respect to the sound source, even though bush crickets have directional hearing. A large prothoracic interneuron, the T-neuron, participates in mediating the behavior (Faure and Hoy 2000). The T-neuron is most sensitive to frequencies higher than those of the calling song (13 kHz peak frequency) for the species. The threshold is about 50 dB less than that of the behavior at 20 kHz (Figure 1b, right). Another primarily diurnal orthopteran, the locust Locusta (Figure left), has a of ears (Figure that are well studied and However, the of hearing in the of this locust of hearing may be to in response to batlike to our there are no of to bats. A locust in stationary flight with its and the both of which produce in the to the sound source 1989). behavior occurs at frequencies above 10 kHz (Figure right). et al. sensitive to high frequencies that from auditory and and are for shows the auditory sensitivity of one of these mantids (Figure are primarily but many at to be many species in the ears (Yager and Hoy The tympanal sensitive to ultrasound is in a between the two (Figure and of two tympanal A of cells is to each tympanal The ears are to as a (Hoy et al. 1989). mantids react at as as 10 m from a batlike signals at react by or with and The are all as would be from a The threshold for avoidance behavior occurs from 20 kHz to 80 kHz (Figure Yager 1999). Yager an interneuron that of the behavior (Figure right), in species The is most part of a neural for avoidance to avoidance behavior from to bat signals have that some can hear and react to bats, but at present is not to them on the of bat-detecting insects. The and crickets and bush crickets, and 1992, Robert et al. 1992). The calling of their nocturnal at 5 kHz as measured electrophysiologically from the This frequency is the peak of the but the hearing into the ultrasonic range. The hearing of is also to its which mostly in the ultrasonic range. can hear bats, but their to batlike signals or to bats in are the of are on by some bats, and ears were in The has ears on the first that are sensitive to sound between kHz and 60 kHz. signals in the behavior during stationary which the of ultrasonic a with some (see Yager et al. 2000). in may be used for intraspecific because they also produce sounds on the present we that hearing in these has been for detecting bats. In response to ultrasound, flying or the and hear at kHz. The ears are in the and have evolved independently from those in et al. for bat Ears have evolved independently in many of of and and for and that the ears are for hearing bat signals. evasive behavior to batlike ultrasound was documented in nocturnal of the that have ears on their and Fullard 2000). lacewings also have ears on their that are sensitive to ultrasound and as bat A of frequency sensitivity is to all these hearing and lacewings are In audition in and which is probably for hearing bats, shows frequency selection pressure has the the hearing or sensitivity to sound of moth hearing. Many species of hear between 20 and kHz, hear a range of ultrasonic frequencies and are more sensitive than their A example of this is the ear is to the biosonar and of the one bat species on the The same or a species of moth in on has a threshold of hearing and a This is probably an to the range of frequencies used by the many bat species there studies of their hearing and behavior The and in the have ears on the of the (Figure middle). the thin tympanal membrane is a large air which is from other air by and a A in 1988) are to the The two cells have (Figure and but they by about 20 dB in range is the range which neural occur to sound measured in dB from threshold to The and have a auditory range of about exhibit a of to bats and to batlike signals for to and 1992). from the source often and fly those to the source show and or the sounds the which the from to less evasive flight behavior. However, that the of and the as the rate of that the most sensitive could several but is to and sensitivity are in more intense than small and may be by bats as away as 10 m. However, from show that are also more sensitive and may detect bats as away as 100 m. Thus, there to be a of that can detect bats at about 10 times the distance which bats can detect (Surlykke et al. 1999). Many in the and of have been and Fullard For example, two from the in but they and to some motor the of the occurs the is and The in behavior is because the of are to like but are more to do In some produce ultrasonic sounds from organs they hear bat signals or are have on bats, as before it flying in response to ultrasonic 1979). The threshold for flight is about dB above that for and from pulses are at the rate of motor to the the rate and Fullard both and rate of the ultrasonic from and The and have ears on the first and 2001). In the (Figure left), the are to the and (Figure middle). The in and have a cells attach to each The range for hearing in the is about 50 dB because each a range than the A cells in a The frequency range of hearing to at least 100 kHz in all species studied (Figure and Some show flight and other to ultrasound et al. respond to bat and and shows a of in stationary the most of which is flight and 2000). which is the behavior with the occurs at to dB with respect to the threshold for the (Figure avoidance behavior to bat is that some flight to and 1998). The bat-detecting ear of some (Figure of that sound to the on either of the (Figure middle). have on the than from the or from the auditory in both of (Figure right). are most sensitive to frequencies between 20 and kHz with a threshold of dB sound pressure The range of the auditory in the is about 20 that in the ear of in tethered flight respond to ultrasound with in flight with and by The threshold is about dB or 20 dB above the threshold et al. and 1999). The ear of the (Figure is in a the of the in each (Figure middle). is a true tympanal which is mostly and contains a small is the tympanal of the cells respond to ultrasound (see for a lacewings in stationary flight show a of to before their flight (Miller (Figure occurs most to ultrasonic signals at of pulses were in the flight of A more of behavior from studies on flying lacewings and bats (Figure signals at possibly with insects to their and dive. The shows this response to signals (Figure between and The bat its rate as it a insect (Figure to before its the bat its rate to a maximum of about 200 signals in the for most bats (Figure the the insect its (Figure 5 presumably in response to the high rate signals. This response the and the bat in this which the insect its (Figure bat signals the in the same behavior (Miller and Olesen 1979). the with we that the selective of was or about the same as that for (see 1982). in the prothoracic respond to ultrasound, but their in behavior is (Miller can some and on we about avoidance at all the response for all the insects we find that the the frequency at which neural or have the between about 20 and 60 kHz with between about and dB The response range from at dB with respect to the the distance at which the insects respond to kHz sounds from m for bats using intense the insects have several of a to avoid even bats (Surlykke 1988). The is for bats, which of less than about dB and the response for flying insects are from on the hearing However, bats can detect a echo at about the insects in Many of the insects above are and normally do not bats. The of to be in avoidance behavior in preadapted insects as crickets, and mantids the This that these auditory have been into an as a of selection as in or flight have 20 dB higher than those for or in the same insects. Consequently, behaviors to may Insect avoidance behaviors and their mechanisms are for example the of The behaviors cannot be and the same may show or no response at all to the same this the of which a by making it for the predator to the prey The of it to but it to be present in the behavior and of some and and lacewings (Miller and Olesen 1979). the auditory and mechanisms in many in several orders of an of evolution in response to bat Some from the of predation by bats by in no bats are on all and on the majority of but some are of bats, for the in the and the in In moth hearing sensitivity to the frequencies and of the bat However, from the (Surlykke as well as those from Fullard 1994) have auditory sensitivity in from bats. have at the frequencies that are to the of with bats in or from also show evasive to ultrasound et al. 2000). The these have sensitive which is not may be to the of a if there is no selection pressure against In a that has may the a few as by in the This is to avoid bats. The that are active in the species and were in in a the and bat ears to bat sounds with as as their flying in the (Surlykke and A of hearing in a et al. active do not their even for intraspecific communication. their ultrasonic which that it is not to ears if they with as from the of with bats. The nocturnal moth from in to bat has with ultrasound. The in flight by the of the against the which is an on the by the (Figure The intense sounds dB at have peak kHz, which would the to bats. The are by bats. However, by being active at a of the with few bats, can ultrasound for intraspecific communication. it that a of hearing from a to a one has in (Surlykke and organs have developed independently in species of that are from bats in This that the ear the of intraspecific acoustic communication in and 1999). Some have to diurnal a in response to predation by bats. diurnal are by and like some many from the few are example is the which is and probably produce sounds that are but most of the is in the ultrasonic range their hearing is sounds to their and to (Surlykke and Fullard 1989). Some diurnal are and their hearing has no at For example diurnal in the from show of auditory from sensitivity in some species to reduced at ultrasonic in other species et al. the diurnal is to ultrasonic frequencies (Surlykke et al. 1998). Many organs on the of the (Figure 1999). the one or a of (Figure from about dB measured at 5 with maximum sound between and 80 kHz (Figure Fullard and and may they hear the biosonar signals of bats. for can be as as dB for signals and but higher for signals Fullard Some bats and insects the same and bats on insects. the in the as the bat The its in a which intense ultrasonic from a special of the membrane and do the are of could startle they could with the biosonar or they could bats of a as many in their or in special The of startle bats and a for the insect to However, a bat to the sounds and these and For bats, can as a can startle and bats but these bats also and Thus, must be used to have as startle signals. The ultrasonic of bat biosonar in two by et al. 1994) and by with range (Miller The long of at high Fullard et al. 1994) that of could as acoustical by the moth and a bat's their on the of the moth to its to the of the bat's signals. et al. this using bats to which from the were during the The bats by the of their but not the success rate of prey The that the by do not with the sonar because the rate is However, in the from with the of bats to range (Miller When by the a of about before an the bat's of range by as as A the with range et al. and with neural et al. 1998). In either a bat would its if from or if they with the bat's to the of as by The the are generally and are the that in response to bat sonar could the bat of a prey. a from studies in the laboratory and the In bats and to (Surlykke and and However, the same bats to moth with a studies that bats often their in the or the prey was the and 1992). not evasive to bats. more than of the and the few they The bats no of were more than of the moth has and associated with species. the that are acoustical bats that the and The and are not In are the may mostly startle bats. bats to the a is most in are for like which produce but intense in response to batlike signals (Surlykke and least some species of to the signals of bats at of about m (Surlykke and these the bat be to hear the an predator to avoid the prey. is for the bat to its behavior even if the during the or of the bat's and 1992). the at the bat's ear before the the bat's of range to the prey the bat to the using long of could this and selective for in to those by in flight behavior. The of many bats contains few because of their effective However, some bats prey on tympanate and how they do this is not known and we present some possible that may bat bats their presence to all potential prey with auditory bat and bats may acoustic to their would be an that frequencies of sonar signals out of the range tympanate insects are most sensitive and Some bats high which them less to most However, small as some (Surlykke and the (Figure right), and many 1988) are sensitive to high The attenuation of high frequencies the range of bat the of this may be of high frequencies were a against tympanate one would to find bats this in many do have sensitivity at high frequencies 1988). The bats that high frequencies are small which detect small prey. the of high frequencies may be a response to insect it may also be a to resolution the of bats to detect insects (Surlykke et al. A would be for bats to which would them less to and This is mostly by bats that insects from the and 1989, frequencies the range but the resolution of the biosonar, making large insects detectable A would be an for bats to the of their signals. is a of sound and is to the sound dB is that in sound pressure and the in sound in the same in dB The would be that the sound with the of the but the echo with the of the distance (Surlykke making the bat less detectable to the tympanate do not if bats this but we that all bats studied sounds if they signals at all 1989, and and For a bat, the of detection range is probably not an because sound the of from the (see and Kalko 2001). In field studies using to the distance and to the bat that at least some bats hunting insects in the intense signals with source up to dB the measured and 1999). bats that both and of for (Faure and 1994) and (Miller and high hunting flying insects. signals be less to prey. However, signals with less which are less to the bats, signals are from bats using and hunting to 1989, we that probably are not a against insect but an for the between from and those from prey. The most effective for a bat to to tympanate prey is to echolocating and detect prey using the insects produce bats may find them using hearing. bats as and have large (Faure and react with to as the sound of or of an insect on a Some bats the of biosonar the et al. However, generally calling before the and produce no 1989, and of the the of is to because tympanate insects probably detect echolocating bats long before the et However, the may (Miller the other or hearing signals. the sonar in the may not be a against tympanate it may be that the bat does not to the of a In the prey of in the bat to its biosonar nearly to the of is to if the acoustical of bats are to of the or to the hunting or are to insect In some bat signals are less to tympanate prey. The evolutionary between bats and their insect and to be Some as acoustical to be documented by laboratory and field studies in the field can some about bat and insect For example, bats hunting at (Miller and and 1994) to some about behavior and the of as or signals. modern systems that can and insect in Some be to An bat can at least some of its insect the of this insect those of their The of an behaviors be a response to the to the of the prey's behavior by the predator less and our for on the also the of Our has been by from the the and the Insects with audition for hearing bats. insects from the are in the with the of the ear as by the The of each ear is in the The for avoidance behaviors (negative for and for the and to be with the behaviors are in the The frequency range for bat biosonar signals is on the right). A of the ear of a cricket, Gryllus bimaculatus from et al. 1989). A of the ear of an from and A the ear of an Locusta from A of the of a from Yager and Hoy The are from and the neuronal are from Insects with audition for hearing bats. insects from the are in the with the of the ear by The of each ear is in the The for flight the and an interneuron are in the The frequency range for bat biosonar signals is on the right). A the of a moth A the first of a moth from the and A tympanal air the the in this The ear is like that of the for a from and as well as for a The of a with one in The sound are on the of the from et al. The of the of the of the for are in the The of a to a hunting The behavior was in a large flight using at The biosonar signals that the bat during the of hunting were The signals are and the are from and Olesen used for the of intraspecific communication sounds by the moth pulses with frequencies of about kHz are by the on the of the against the on the from and and The on the of the of the The a of each it The of the has a maximum at about 80 kHz. from and and from