Auditory behavior of the cricket IV. Interaction of direction of tracking with perceived temporal pattern in split-song paradigms

1. Split songs — stimulus arrangements in which parts of cricket calling songs are delivered from different directions via two loudspeakers — have been used to place constraints on the recognition and direction-determining processes in the field cricketGryllus bimaculatus, walking on the Kramer spherical treadmill. Parts of a calling song, each ineffective when delivered alone, can combine to elicit recognition and tracking when interleaved from two speakers 135° apart in azimuth.

2. If two correct songs (40-ms syllable repetition interval) are delivered from the two speakers, but time-shifted by a half interval so that their composite (20-ms intervals) is incorrect, then the angle bisecting the speaker separation (where the two should appear equally loud) might be expected to be an ‘unfavored zone’; the animals do in fact tend to avoid that region. As in controls on ‘choice’ experiments with interleaving of complete chirps, if one speaker is increased in sound intensity it tends to be favored.

3. If two incorrect songs (80-ms syllable repetition interval) are delivered from the two speakers, again time-shifted by half a syllable interval so that the composite is correct (40-ms intervals), then angles near those bisecting the speaker separation might be expected to be favored, and most animals track in that way. Surprisingly, however, when one speaker is increased in sound intensity in this case most animals show a marked shift of tracking toward the less-loud speaker at test intensities near 60 and 70 dB. At 80 dB the effect is reversed.

4. About one-third of females track trill — continuous trains of syllables at the correct syllable rate but with no subdivision into chirps; these were tested with split trills. If alternate syllables of the trill were delivered from the two speakers, or alternating doublets, tracking tended to bisect the 135° angle between speakers. With alternating triplets or higher n-tuples, one or the other speaker was tracked quite accurately. This ability to lock on to one signal despite the presence of another (at 135°) was tested at smaller speaker separations (90°, 65°, 45°); the animals then usually fail to select one speaker.

5. The above results are difficult to explain via a putative ‘tracking comparator’ with ‘cardioid’ inputs from the two sides of the auditory system. In particular, the distinct behaviors with respect to intensity imbalance (2 and 3 above, for different ranges of syllable repetition interval of split songs) suggest that bilateral dynamic interactions in the CNS may be involved. Specific correlates of these effects can be sought in multiple-source or dual leg-phone experiments with electrophysiological monitoring of prothoracic and brain cells.