1. In Eptesicus the auditory cortex, as defined by electrical activity recorded from microelectrodes in response to tone bursts, FM sweeps, and combinations of FM sweeps, encompasses an average cortical surface area of 5.7 mm2. This area is large with respect to the total cortical surface area and reflects the importance of auditory processing to this species of bat. 2. The predominant pattern of organization in response to tone bursts observed in each cortex is tonotopic, with three discernible divisions revealed by our data. However, although cortical best-frequency (BF) maps from most of the individual bats are similar, no two maps are identical. The largest division contains an average of 84% of the auditory cortical surface area, with BF tonotopically mapped from high to low along the anteroposterior axis and is part of the primary auditory cortex. The medium division encompasses an average of 13% of the auditory cortical surface area, with highly variable BF organization across bats. The third region is the smallest, with an average of only 3% of auditory cortical surface area and is located at the anterolateral edge of the cortex. The region is marked by a reversal of the tonotopic axis and a restriction in the range of BFs as compared with the larger, tonotopically organized division. 3. A population of cortical neurons was found (n = 39) in which each neuron exhibited two BF threshold minima (BF1 and BF2) in response to tone bursts. These neurons thus have multipeaked frequency threshold tuning curves. In Eptesicus the majority of multipeaked frequency-tuned neurons (n = 27) have threshold minima at frequencies that correspond to a harmonic ratio of three-to-one. In contrast, the majority of multipeaked neurons in cats have threshold minima at frequencies in a ratio of three-to-two. A three-to-one harmonic ratio corresponds to the ''spectral notches'' produced by interference between overlapping echoes from multiple reflective surfaces in complex sonar targets. Behavioral experiments have demonstrated the ability of Eptesicus to use spectral interference notches for perceiving target shape, and this subpopulation of multipeaked frequency-tuned neurons may be involved in coding of spectral notches. 4. The auditory cortex contains delay-tuned neurons that encode target range (n = 99). Most delay-tuned neurons respond poorly to tones or individual FM sweeps and require combinations of FM sweeps. They are combination sensitive and delay tuned. The response of cortical delay-tuned neurons is phasic with an average of 1.1 +/- 0.6 (mean +/- SD) action potentials per pulse-echo pair. Thus delay-tuned neurons act as coincidence detectors without additional facilitation. 5. The distribution of best echo delays (BDs) of cortical delay-tuned neurons appears bimodal, with neurons in the first mode (n = 33) encompassing BDs from 2 to approximately 11 ms and the second mode (n = 66) from approximately 11 to 28 ms. The distribution of BDs in the first mode is similar to the range of BDs seen in other species of bats, but the distribution of BDs encompassed by the second mode appears unique and suggests a longer working range for echolocation in this species. As an ensemble, cortical delay-tuned neurons encode target ranges of 34-484 cm, corresponding to the behaviorally observed 5-m limit for the detection of insect-sized objects. 6. Delay-tuned neurons in the tonotopically organized area (n = 36) typically exhibit shorter BDs and longer echo facilitation latencies (EFLs) than delay-tuned neurons in the variable frequency area (n = 48). These differences are statistically significant and suggest that delay-tuned neurons in the variable and tonotopic regions constitute separate, multiple representations of target range information. 7. The distribution of BDs delay-tuned neurons is not topographically arrayed across the auditory cortex, at least not on the same large scale as tonotopic frequency organization. 8. Some (13 out of 25 tested) delay-tuned neurons systematically shifted their BDs in response to changes in the simulated pulse amplitude. These delay-tuned neurons are termed ''amplitude shift'' neurons. Of the 13 amplitude shift neurons, 11 exhibited a decrease in BD in response to decreasing pulse amplitudes, and 2 exhibited an increase in BD in response to decreasing pulse amplitudes. Amplitude shift cells shifted their BD by an average 11 +/- 8 ms for an average change of 13 +/- 8 dB SPL. The location of amplitude shift cells appears randomly distributed among delay-tuned neurons in the variable and tonotopic areas. 9. We develop a conceptual framework to understand the nontopographic organization of BDs by considering the overlapping multipeaked frequency-tuned and delay-tuned cortical fields in Eptesicus as parallel spectral and temporal components of a target range imaging process.
