Earlier studies of phonotaxis by female crickets describe this selective behavioural response as being important in the females' choices of conspecific males, leading to reproduction. In the present study, moderate (30+) to very large data sets of phonotactic behaviour by female Acheta domesticus L., Gryllus bimaculatus DeGeer, Gryllus pennsylvanicus Burmeister and Gryllus veletis Alexander demonstrate substantially greater plasticity in the behavioural choices, as made by females of each species, for the syllable periods (SP) of model calling songs (CS) than has been previously described. Phonotactic choices by each species range from the very selective (i.e. responding to only one or two SPs) to very unselective (i.e. responding to all SPs presented). Some females that do not respond to all SPs prefer a range that includes either the longest or shortest SP tested, which fall outside the range of SPs produced by conspecific males. Old female A. domesticus and G. pennsylvanicus are more likely to be unselective for SPs than are young females. Each species includes females that do not respond to a particular SP when responding to CSs with longer and shorter SPs. The results suggest that the plasticity of phonotactic behaviour collectively exhibited by the females of each species does not ensure that choices of a male's CS effectively focus the female's phonotactic responses on CSs that represent the conspecific male. The phonotactic behaviour collectively exhibited by females of each species does not readily fit any of the models for selective processing by central auditory neurones that have been proposed to underlie phonotactic choice.

In response to model calling songs (CSs), the phonotaxis of female Acheta domesticus ranges from being very selective to unselective. Within 15 min of nanoinjecting juvenile hormone III (JHIII) or picrotoxin (PTX) into the prothoracic ganglion, females become more selective for syllable period (SP) than in pre-tests. Controls for JHIII experiments, including nanoinjection of acetone into the prothoracic ganglion or nanoinjection of JHIII into the metathoracic ganglion, do not influence selectivity. Similarly, nanoinjection of saline into the prothoracic ganglion and nanoinjection of PTX outside of the prothoracic ganglion does not change the overall selectivity of the female's phonotaxis. These results indicate that circuits in the prothoracic ganglion modulate the SP-selectivity of phonotaxis.Photoinactivating both of the ON1 prothoracic auditory interneurones in old females that were previously unselective for SP also results in greater SP-selectivity during phonotaxis. Evidence suggesting that ON1 has this effect via its inhibitory input to L3 (another prothoracic auditory neurone) includes: photoinactivation of one ON1 neurone causes angular errors in the female's orientation to CSs at 85 dB (above the threshold of the L3), stimulation with 60 dB CSs (above the threshold of ON1 but below the threshold of L3) does not induce errors in angular orientation, inactivation of ON1 in old crickets results in greater angular errors (85 dB stimulus) than it does when ON1 is inactivated in young females, and photoinactivation of ON1 increases the firing rate of the L3 neurone. Fig. 4. Controls for picrotoxin (PTX) injection. (a) Tables showing positive phonotaxis (shaded boxes) in response to syllable periods (SPs)ranging from 30 to 90 ms (columns) for 26 females (each row is one female) before and after nanoinjection of 9.2 nL of saline into the ventral portion of the prothoracic ganglion. (b) Graphs showing the females in (a) grouped by the number of SPs they responded to in the pre-test: those that responded to only one to three of the SPs presented in the pre-test (left graph, n ϭ 5), those that responded to four or five SPs during the pre-tests (middle graph, n ϭ 18), and those that responded to six or seven of the SPs presented during the pre-test (right graph, n ϭ 6). (c) More controls for PTX injection. Tables showing positive phonotaxis (shaded boxes) in response to SPs ranging from 30 -90 ms (columns) for 12 females (each row is one female) before and after nanoinjection of 50 pg of PTX in 9.2 nL of saline into the haemolymph outside of the prothoracic ganglion. (d) Mean results from the females in (c). 330 G. Atkins et al.

Female crickets (Gryllus pennsylvanicus), caught in the field as nymphs, responded as adults in the laboratory with selective phonotaxis to model calling songs (CSs) that reproduced the dominant carrier frequencies and syllable periods (SPs) characteristic of the male's natural calling song. Extracellular recordings demonstrated two types of auditory interneurons in the female's cervical connectives that were very similar to the AN1 and AN2 neurons previously described in other gryllid species. The AN2 neuron responded to model CSs with a phasically encoded immediate response, and a more tonically encoded prolonged response. AN2's immediate response exhibited SP-dependent decreases (termed decrement) in its responses to sequential syllables of the CS that were greatest to CSs with the shortest SPs and diminished as SPs were lengthened, resulting in an SP-dependent habituation. Picrotoxin application transformed this SP-dependent habituation by AN2 to SP-selective responses in which the degree of decrement was greatest to SPs that were most phonotactically attractive. AN2's prolonged response was most sensitive to 5 kHz CSs and correlated with the carrier frequency tuning for the thresholds of phonotaxis by females. Thus, in females, AN2's immediate (in the presence of picrotoxin) and prolonged responses were selectively tuned to the SPs and carrier frequencies of the male's calls that were most attractive behaviorally. AN1's responses at threshold were also tuned to the dominant carrier frequencies of the male's CS.

The L3 auditory interneuron in female Acheta domesticus, produces two different responses to the male calling song: an immediate response and a prolonged response. The prolonged response exhibited spiking activity and a correlated prolonged depolarization, both of which are clearly seen in intracellular recordings. The morphology revealed by intracellular staining was clearly the L3 neuron. The amplitude of the prolonged depolarization associated with the prolonged response increased with increases in sound intensity, resulting in increased spiking rates. Both depolarization and sound presentation increased the spiking rate and the slope of pre-potentials (thus leading to spiking threshold more quickly). Injecting hyperpolarizing current had the expected opposite effect. The effects of positive current injection and sound presentation were additive, resulting in spiking rates that were approximately double the rates in response to sound alone. Short postsynaptic potentials (PSPs), whose duration ranged from 15-60 ms, which may lead to action potentials were also observed in all recordings and summated with the prolonged depolarization, increasing the probability of spiking.